AVENDUS SPARK RESEARCH SENSEX

Closing
83,443
% YTD
6.79
% 1 YR
4.35
08 July 2025 NIFTY 25,461 7.68 4.69

WHAT’S INSIDE SECTORAL INDICES

Closing % YTD % 1 YR
▪ Avendus Spark Thematic on Data Center Ecosystem – A Multi-Year Growth Proxy on India’s Data Explosion and Localization Wave BSE Meta l 31,664 9.6 (5.4)
BSE IT 38,068 (11.9) (1.1)
− Rising penetration of mobile/internet & data-heavy services is set to drive global data generation to 210 ZB by 2025 and >600 ZB by
BSE Cons Dur 61,047 (5.3) 5.5
2030 (25 p.c.+ CAGR)
BSE Power 6,869 (1.4) (14.7)
− India accounts for 20 p.c. of global mobile data consumption but only for ~3 p.c. of global data centre capacity BSE Ca p Goods 72,001 6.2 (4.8)

− See a multi-year data centre demand super cycle with likely growth of 25 p.c.+ p.a. led by India’s lower cost, improved domestic BSE TECK 18,756 (3.7) 6.1

capabilities & policy push BSE Rea l ty 7,545 (8.4) (13.2)

BSE Ba nk 63,576 10.1 5.8
− Top five DC players unlisted: Sify Infinite Spaces, CtrlS and Nxtra may list in the medium term; Anant Raj & Techno Electric are BSE Auto 53,539 3.6 (6.8)
emerging players BSE PSU 19,914 5.5 (10.5)
− Electrical and cooling solution providers are also key proxies to benefit from the ongoing data centre capex cycle BSE Hea l th 44,877 (0.9) 17.3
BSE Oi l 28,445 9.1 (6.6)
▪ Exchange and Currency Performance BSE FMCG 20,478 (1.4) (2.1)
▪ Avendus Spark Focus Stocks
NIFTY OUTPERFORMERS
Price % YTD % 1 YR
Di vi 'S La bora tori es Ltd 6,889 12.9 54.3
Bha rti Ai rtel Ltd 2,034 28.1 41.7
Ba ja j Fi na nce Ltd 925 35.7 30.4
Ba ja j Fi ns erv Ltd 2,006 28.0 27.8
Hdfc Life Ins ura nce Co Ltd 789 27.9 27.1

NIFTY UNDERPERFORMERS
Price % YTD % 1 YR
Indus i nd Ba nk Ltd 854 (11.0) (40.6)
Ta ta Motors Ltd 689 (6.9) (31.3)
Coa l Indi a Ltd 384 0.0 (22.1)
Hero Motocorp Ltd 4,307 3.5 (21.7)
Oi l & Na tura l Ga s Corp Ltd 242 0.9 (19.3)

find AVENDUS SPARK RESEARCH on Bloomberg [RESP SPAK <go>] | FACTSET | REFINITIV EIKON
This file was  Analystwebsite
generated for vishnu.agarwal@wipro.com from Avendus Spark Research Certification and17,
on July Required Disclosures begin on last two pages – Click here | Page 1
2025 06:30:53
 This file was generated for v i s h n u . a g a r w a l @ w i p r o . c o m from Avendus Spark Research website on July 17, 2025 06:30:54

DATA CENTERS
A Multi-Year Growth Proxy on India’s Data Explosion and Localization Wave
“Data is the new oil, and India does not need to import it. We have it in super-abundance.” — Mukesh Ambani, Chairman-RIL, 2017
When Mr. Mukesh Ambani made this now-iconic statement, few could have foreseen how swiftly that vision would materialize. India’s per capita mobile THEMATIC REPORT
data consumption has surged to among the highest globally, with its share of global mobile data consumption rising from just 5% a decade back to 8 July 2025
nearly 20% now. However, what hasn’t kept pace is the commensurate expansion in domestic data center infrastructure required to store and process
this data. India’s data center capacity has grown at a 20%+ CAGR since 2019, reaching approximately 1 GW with a tight vacancy rate of ~4%, yet it still Market data
accounts for only ~3% of global capacity. This imbalance is now correcting as hyperscalers increasingly localize data processing to be closer to end- BSE SENSEX 83,443
users—driven by low-latency needs, improved domestic capabilities (especially in networking and reliable power), and regulatory tailwinds including
data localization mandates. As a result, data center demand is entering a multi-year super cycle with likely growth of 25%+ annually, fueled by rising NIFTY 25,461
data consumption (particularly video and AI workloads), expanding cloud adoption, increased digital participation, low capex/opex cost and strong Performance (%)
policy support. However, supply (planned data center capacities) is also scaling up rapidly along with entry of newer players, which may exert pressure
on pricing and margins. Sustaining return metrics in this high growth landscape will require strategic differentiation through operational scale, energy CMP 3M 12M
efficiency, a balanced hyperscaler-enterprise mix, and edge-readiness by data center players. Another way to capitalize on this structural growth is by
Sensex 83,443 14 4
adopting a “shovel-in-a-gold-rush” approach—investing in MEP players (electrical and HVAC solution providers) who are poised to be critical
beneficiaries of the ongoing data center capex cycle. Anant Raj 544 31 6
Key thoughts covered in this note: TEECL 1,579 70 7
Data Centers, Core Infrastructure Bet in the Global Digital Supercycle: Global data generation is doubling every 3–4 years, driven by rising mobile/internet
penetration and widespread adoption of data-heavy services like OTT streaming, e-commerce, gaming, and digital payments—all significantly boosting per- E2E 2,486 34 47
user consumption. The integration of AI into applications is set to accelerate this further. More data was created over 2023–25 than in all prior years
combined, with global volumes projected to reach 210 ZB by 2025 and over 600 ZB by 2030 (25%+ CAGR). This surge necessitates large-scale investments in
hyperscale data centers, as AI workloads require ~3-5x more compute than traditional ones—75% of new DC capacity is expected to be AI-ready. India is
outpacing global trends, enabled by ultra-cheap mobile data. Its share of global mobile data consumption has jumped from 5% to ~20% over the past
decade, with among the highest per-user usage (~25 GB per month) globally. Yet, smartphone and 5G adoption remain under 50%, offering ample
headroom. On the enterprise side, India’s public cloud market is growing at 20%+ CAGR, led by rising adoption of flexible, pay-as-you-go models.
Simultaneously, India is emerging as a major AI market, now second only to the United States (ex-China) in terms of adoption. The government’s proactive
push—such as the India AI mission aimed at boosting GPU availability through public-private partnerships—is expected to ease some of the supply-side
constraints. Furthermore, the Indian government’s broader digitization drive, especially across public sector undertakings (PSUs) to improve governance,
uptime, and service delivery, would be key enabler for data center demand.
India’s Data Center Momentum led by Upgraded capabilities, Lower capex/opex cost, and Policy Tailwinds: India consumes ~20% of global mobile data RESEARCH ANALYST
but accounts for just ~3% of global data center capacity—highlighting the need for data localisation. Hyperscalers are scaling up in India, attracted by strong
Gaurav Nagori
data growth, better fiber connectivity, reliable power, and lower capex/opex versus regional peers. With Singapore limiting new DCs due to land and power gaurav.n@avendusspark.com
constraints, global players are increasingly turning to India, Malaysia, and Thailand. India’s appeal is reflected in a 4x rise in colocation capacity leased by +91 44 4344 0072
hyperscalers over the past five years. Policy support is also strengthening the case: RBI and SEBI data localization norms are already in force, and the DPDP
Act could further accelerate demand. Incentives like infrastructure status and state-level exemptions on land and electricity costs enhance viability. India’s
third-party data center demand is expected to grow at 25–30% CAGR, or 250–300 MW annually—driven by AI, Gaming and streaming-led data growth,
broader cloud adoption, and a supportive regulatory environment—solidifying India’s position as a competitive data center hub in the region.
(Cont’d) Page 2
find AVENDUS SPARK RESEARCH on Bloomberg [RESP SPAK <go>] | FACTSET | REFINITIV EIKON
This file was generated for vishnu.agarwal@wipro.com from Avendus Spark Research website on July 17, 2025 06:30:53
 DATA CENTERS
A Multi-Year Growth Proxy on India’s Data Explosion and Localization Wave
Demand Surge Drives Supply announcements and new players entry; Pricing/Margin Pressure from Hyperscaler a key monitorable: India’s data center capacity stands at ~1 GW, with a
tight vacancy of just 4%, reflecting strong underlying demand. The market remains geographically concentrated, with Mumbai and Chennai accounting for ~65% of capacity. Large-format
data centers (>20 MW) now make up 56% of total capacity across top cities, up from 42% in 2020—signaling a shift toward hyperscale-ready infrastructure. Hyperscalers now drive ~54%
of demand (vs. 43% in 2021), with BFSI and tech adding another ~30%. Annual DC investments have risen to $1–1.5 bn and are expected to grow to $2–3 bn, supported by aggressive
hyperscaler capex and colocation build-outs. Planned capacity additions would aid in India’s data center capacity doubling over next three years and trebling to 3 GW by 2030E.
The market is currently consolidated, with NTT and STT GDC accounting for ~50% of capacity and the top five players together holding ~80%. However, this share may decline to 55–60%
by 2027E as new players—supported by PE or global partnerships—enter the fray. Hyperscalers are negotiating hard on pricing, with colocation rentals down ~20% in the last few years.
Hyperscalers have announced “own and operate” capacities and we believe this is in line with the global strategy of captive data center models for internal workloads, while preferring
third-party colocation for cloud business. As a result, rental trends—particularly for hyperscaler-led demand—remain a key variable for returns in the medium term.
Positioning for the Data Center Supercycle: Assessing the business model- Colocation versus Services (Cloud/Managed): Data centers offer a unique blend of growth and yield. Third-
party colocation provides steady annuity-like returns with 10–15% pre-tax yield, backed by long-term contracts with hyperscalers and enterprises. Edge data centers in Tier-2/3 cities offer
higher IRRs driven by latency-sensitive workloads like CDN, gaming, and 5G. In contrast, cloud infrastructure, while capex-intensive, offers scalable RoEs (20%+) as utilization improves.
Managed services provide an asset-light, high-margin, annuity-like revenue stream by layering value-added offerings on top of existing infrastructure. Both cloud and managed services
face scalability constraints due to intense competition from hyperscalers and IT companies. Our analysis of six leading DC players (mostly colocation-focused) shows revenue CAGR of 20–
30% over the last three years, though returns remain depressed due to ongoing investments. Players with a balanced colocation-services portfolio, diversified client mix (hyperscalers for
scale/cash flows, enterprises for margin), strong leasing track record, and prudent capacity expansion supported by internal accruals and conservative leverage would be best placed to
leverage this demand tailwind.
The top five DC players with strong execution and scale remain unlisted, though Sify Infinite Spaces, CtrlS and Nxtra are expected to list in the medium term. Among listed names, Anant
Raj (Rs.185bn Mcap) and Techno Electric (Rs.183bn Mcap) are emerging players, aiming to scale from negligible capacity to 200–300 MW over the next 3–4 years. Anant Raj, an NCR-
based developer, has carved a niche with PSUs, benefiting from government digitization and outsourcing. Its JV with Orange is aimed at expanding its cloud offerings. Techno Electric, a
leading EPC player in power infra, is developing its maiden DC in Chennai with a strong focus on energy efficiency (low PUE). It has plans to commence similar large scale data centers in
Kolkata and Noida and also announced JV with Railtel to develop 100MW of edge DC. E2E Networks (Rs.50bn Mcap) is the only listed pure-play cloud provider, offering GPU-based
infrastructure tailored for AI/ML workloads.
Indirect Proxies: Selling Shovels in the Digital Gold Rush - Given the pricing pressure from hyperscalers in colocation and the capital-intensive nature of cloud infrastructure, one may
consider indirect beneficiaries of India’s multi-year data center boom. Electrical systems contribute ~30–35% of total DC capex, with major spends per MW on diesel gensets, UPS,
batteries, switchgear, and power cables. Key listed beneficiaries include Cummins, ABB, Schneider Electric, Siemens, and Polycab. Cooling infrastructure (HVAC) forms ~25% of capex,
primarily driven by investments in chillers, heat exchangers, and cooling towers. While global players like Vertiv and Stulz dominate, Indian industrial cooling firms such as Blue Star are
also well-positioned. Other enablers include data center information management (DCIM) software from Schneider and Siemens, optical fiber providers like HFCL and Sterlite
Technologies, and EPC contractors such as S&W and KEC International, who support physical connectivity and buildouts.

find AVENDUS SPARK RESEARCH on Bloomberg [RESP SPAK <go>] | FACTSET | REFINITIV EIKON
This file was generated for vishnu.agarwal@wipro.com from Avendus Spark Research website on July 17, 2025 06:30:53 Page 3
Data Centers in India: A Play on Growth in Data Consumption and Localization

This file was generated for vishnu.agarwal@wipro.com from Avendus Spark Research website on July 17, 2025 06:30:53
DATA CENTER INDUSTRY THEMATIC

Growth drivers for Indian data center ecosystem: Data growth (OTT, Video Streaming, AI) + Data accessibility (Cloud) + Data localization (Regulatory thrust)

Exploding Datacenter Growth in India

▪ Growth in India’s data center industry is driven by a
confluence of rapid data growth, hyperscalers scaling-up
operations in India and regulatory push for data
localization. Data Growth Data Accessibility Data Localisation
▪ India’s explosive data growth stems from surging internet
penetration and the rapid rollout of 5G networks.
Low Latency
Simultaneously, the uptick in digital activities—OTT Traffic and New data Growth Accessing on Cloud
streaming, e-commerce, gaming, digital payments, AI/ML versus Devices ▪ Hyperscalers preferring
and heightened cloud adoption—has driven the need for ▪ Mobile subscriber/ internet DC locally
low-latency, edge-ready infrastructure. users' growth ▪ Consumers preferring storing
data on Cloud than end points ▪ New applications requiring low
▪ Incrementally, Hyperscalers have been rapidly scaling up ▪ 5G upgrade from 4G latency for better adoption
the presence in India locally, considering strong network ▪ Enterprise Cloud adoption
▪ Government thrust on
connectivity, uninterrupted power supply and low cost of digitization of data records ▪ Captive DC to third-party DC
capex/opex. Policy Thrust on
▪ Increased DC capacities and capabilities demonstrated to Localisation
operate resulting in enterprises looking to outsource to
▪ DC draft bill 2020, RBI, DPDPA
these domestic third-party data centers. India benefits Change in Nature
2023
from lower land and power costs, improving infrastructure of Data Usage
and grid enhancements that make it attractive as a regional ▪ Achieving low latency
▪ New applications require more
datacenter hub for the world. data usage
▪ A significant catalyst has been data localization mandates—
▪ Social media, Streaming, Emerging as
such as the Draft Data Centre Policy 2020, the Digital
gaming, Ecommerce, payments Datahub for World
Personal Data Protection Act 2023, and draft DPDP rules—
that require domestic storage of personal data. ▪ Adoption of AI (3x-5x more ▪ Lower cost of land and
Government’s impetus on digitization across functions is computing needs) uninterrupted power supply
resulting in DC capacity demand from PSUs.
▪ Improving capabilities

Source: Industry, Avendus Spark

This file was generated for vishnu.agarwal@wipro.com from Avendus Spark Research website on July 17, 2025 06:30:53 Page 5
DATA CENTER INDUSTRY THEMATIC
Worldwide Data Explosion – Worldwide data created is doubling every 3-4 years led by converging trends in digital behavior (internet
adoption) and technological advancement (Cloud/AI)

Worldwide Data created, captured or replicated in Zettabytes What is going to drive data explosion and DC growth?
435 ZB ▪ Global data creation is doubling every 3-4 years, driven by AI,
More data was made during CY23-25
cloud adoption, and digitalization. More data was created in
than in all of preceding years
2023-25 than in all previous years combined, with volumes
expected to reach 210 ZB in 2025 and 612 ZB by 2030.
420 ZB ▪ Growth over next 5 years is expected to be majorly led by AI
& Machine-Generated Data, IoT devices adoption, higher
usage of video & immersive content and Cloud/SaaS.
2 5 7 9 13 16 18 26 33 41 64 84 101 120 140 175 600+ ▪ Only 10% of this data created is stored whilst the rest is
transient. Of the stored data: (I) 40-50% is in data
2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2030
centers/cloud (II) 10-15% in edge infrastructure (III) 35-50%
at endpoints.
▪ Major shift since 2010, when endpoints held 80% of stored
data. DC/cloud will account for 50% and 70-80% by 2030E,
First gen AI Open AI reflecting a structural move toward centralized, scalable
model launched launched- GPT-1 ChatGPT launched publicly infrastructure with enterprise workloads shifting to the
and reached 1mn users cloud. High-frequency sync, logging, and backup increase
within 5 days data volume.

Note- 1ZB= 1000bn GB. Source: IDC Global datasphere, Avendus Spark

Data created is not data stored; Only 10% of data created is stored in DC/Edge/Devices Share of DC/Cloud for data storage/computing is increasing significantly owing to change in
nature of consumption
Shift in stored data from Endpoints (local systems) to Datacenters and Edge
Data
Stored
40% Endpoint

Transient Data
175ZB 80%
Edge
Data Created
(cache, streaming
logs etc) 50%
90% 10% DC/Cloud
15%

2010 2025

Source: Seagate, IDC Global datasphere, Avendus Spark Source: Seagate, IDC Global datasphere, Avendus Spark
This file was generated for vishnu.agarwal@wipro.com from Avendus Spark Research website on July 17, 2025 06:30:53 Page 6
DATA CENTER INDUSTRY THEMATIC
India Mobile data consumption - India’s share of global mobile data consumption has surged from 5% a decade ago to around 20% today,
where it has remained stable over the past five years
India & Global mobile data traffic have grown at the same India & North East Asia mainly China contribute to ~50% of India’s mobile data contribution to the global level has been
rate over the last 5 years global mobile data consumption stable for the last 6 years
Global Mobile Data Traffic (EB/Month) India Mobile Data Consumption as % of
Mobile Data Consumption Across Regions (%) Increased
Global Mobile Data Consumption
smartphone adoption
NA, 7%
30% LATAM, 6%
CAGR India, 20%
122 Western 21% 20% 21%
20% 20% 20%
Europe, 8% Launch 17%
16%
Central & 13%
South East Eastern
32 24 Asia & Europe, 6%
2 7 5%
Oceania, 14% MEA, 9% 3% 4%
2% Stable as % of total mobile
2014 2019 2024 data consumption
North East
Global (EB/Month) India (EB/Month) Asia, 31% 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024

Note- 1EB= 1billion GB. Source: Ericsson, Avendus Spark Source: Ericsson, Avendus Spark Source: Ericsson, Avendus Spark

Global mobile traffic increased due to increased contribution from video streaming which rose Video streaming has the highest data intensity among most common activities at 20mb-
from sub 50% in 2019 to 74% currently; This is estimated to increase to ~82% by 2030 150mb/minute
Mobile traffic by application category - Global Activity MB per minute
Video
Audio Streaming (Minutes) 0.01 - 0.02
Social Networking
Message/multimedia attached (Message) 0.1 - 0.5
52% Audio
66% Emails & Email/1MB attached (Message) 0.1 - 1.1
74% 82% Web Browsing
Social Media (Minutes) 0.4 - 1
13% Software Update
10% File Sharing Web Browsing (Minutes) 1-5
8%
6%
Other App downloads/updates (Download) 3 - 30

2016 2020 2024 2030 Video Streaming (Minutes) 20-150

Source: Ericsson, Avendus Spark Source: Ericsson, Avendus Spark

This file was generated for vishnu.agarwal@wipro.com from Avendus Spark Research website on July 17, 2025 06:30:53 Page 7
 DATA CENTER INDUSTRY THEMATIC
Drivers for Mobile Data Consumption growth in India – Scope for improvement in Smartphone and 5G Adoption in India; Usage Intensity
which is defined by data consumption per user per month is already one of the highest globally
India (CAGR / ▪ Average wireless data usage used to be <1GB per month in India on a per subscriber
USA China EU India Growth Change)
(2030) basis but with the introduction of Jio, the data consumption increased to ~2GB by
Mar-17.
Internet User ▪ Another inflection point for data consumption in the country was with the
317 1,142 486 836 982 3% widespread adoption of smartphones as consumption per month grew from 2GB to
(Mn. People)
9GB YoY.
▪ Since FY19, data usage/wireless subscriber/month has been growing at ~2GB/month
Penetration (%) 93% 80% 90% 58% 65% 7%  with the number reaching 25-30GB/month in Mar-25.
▪ We expect consumption growth to stabilize/witness moderate growth going forward
with incremental growth coming from increased 5G adoption & internet usage
Smartphones becoming data intensive.
289 1,105 471 550 696
693 4%
(Mn. People) ▪ Hence expect data consumption to increase by ~5-10 EB/month from Mar-25 versus
Mar-2030.

Penetration (%) 85% 77% 92% 38% 46% 8%  Avg Wireless Data Usage/Wireless data subscriber/Month

Mobile data per month Consumption Widespread

22 25 19 25-30 30-35
30-35% 5-6% 2GB YoY increase every year
per smartphone (GB) below adoption of
since Mar-19
<1GB/month Smartphones 30

Mobile data per month 24

8 38 10 24 30-35 9%
(EB)* 20
17
16
12
5G Penetration 71% 55% 30% 29% 50% 21%  11
9

2
0 1
Cost of Data/GB 5.62
5.62 0.41
0.41 0.79
0.79 0.17 CAGR  Change
Mar-16 Mar-17 Mar-18 Mar-19 Mar-20 Mar-21 Mar-22 Mar-23 Mar-24 Mar-25 Mar-30

* Mobile data per month used for USA is NA, for China it is North East Asia, For EU it is West Europe & for India it is India/Nepal/Bhutan
Source: GSMA, Ericsson, Kantar, CNNIC, IAMAI, Avendus Spark Estimates Source: TRAI, Avendus Spark Estimate
This file was generated for vishnu.agarwal@wipro.com from Avendus Spark Research website on July 17, 2025 06:30:53 Page 8
DATA CENTER INDUSTRY THEMATIC

Cloud adoption- India’s cloud industry is expected to grow at 20%+ CAGR over the next 5 years led by companies increasingly preferring
subscription-based pay-as-you-go models, reducing upfront CapEx; India contribution to global cloud industry is meagre at ~1%
Global cloud industry stands at ~$920bn, it is expected to grow at 15% (CAGR) until FY30 from Key Cloud Growth Drivers in India
FY25, expect catch-up of the India’s cloud industry share in the global landscape
Digital ▪ Shift to Cloud infrastructure as a foundation is especially prominent among
India Public Cloud Industry ($Bn.) Global Public Cloud Industry ($Bn.) & Transformation SMEs, which are leveraging cloud solutions to modernize their IT systems,
India Cloud Contribution (%) Initiatives enhance operational efficiency, and deliver better customer experiences.
22 1.17%
20% Growing
CAGR 0.83% ▪ The integration of AI and data analytics into cloud services is accelerating
Demand for
demand, This shift is driving the need for scalable, high-performance cloud
0.59% Advanced
26% 1,846 infrastructure to support advanced digital capabilities.
8 Technologies
CAGR
2 922
▪ Govt. policies focused on strengthening the country’s digital infrastructure.
412 Government
The National Digital Communications Policy (NDCP) outlines a strategic vision
Support
FY20 FY25 FY30 FY20 FY25 FY30 to migrate a substantial portion of government services to the cloud by 2025.

Source: Nexdigm, Avendus Spark Source: Nexdigm, Avendus Spark

Global data center share is shifting from on-premise to hyperscale either in their own DCs or According to a study done by EY, 80% of Indian companies have already adopted cloud
in COLO capacities leased by DCs services partially; ~65% transitioning even their applications to cloud
Global share of DC capacity across captive and colocation % of companies adopting cloud in India
10% 15% 90%
20%
10% Hyperscale: owned 80%
15%
20% 20% 67%
Hyperscale: leased
20%
20%
Colocation: non-hyperscale

60%
50% Enterprise on-premise
40%

Indian companies which partially India companies transitioning India companies attributing AI
2017 2022 2027 migrated to cloud applications to cloud adoption possible due to cloud

Source: Nexdigm, Avendus Spark Source: EY, Avendus Spark

This file was generated for vishnu.agarwal@wipro.com from Avendus Spark Research website on July 17, 2025 06:30:53 Page 9
DATA CENTER INDUSTRY THEMATIC

India Cloud Segments- India’s cloud market segment can be divided into Public, Private & Hybrid cloud with Public cloud taking the
lion-share (60%); In the Indian cloud segment ‘Infrastructure as a Service’ has the highest contribution of 55%
Key Drivers for Enterprise Shift to Colocation Public cloud contributes 60% of India’s cloud market followed by hybrid cloud
Lower CapEx &
▪ Avoid high upfront costs (land, build, equipment)
Pay-as-you-go
India Cloud
Speed & Scalability ▪ Deploy faster; scale capacity as needed Market
Segmentation

Operational Simplicity ▪ Reduce need for 24×7 specialized staff Public Cloud Hybrid Cloud

Interconnection &
60% 25% Private
▪ Easy hybrid/multi-cloud setups, IX access, low latency Cloud
Cloud Access
15%
Security & Compliance ▪ Certified infrastructure with updated standards

Source: Industry, Avendus Spark Source: Nexdigm, Avendus Spark

On a market segment level, IaaS has the highest contribution with 55% versus 15% in
FY20; Largely led by increase in PaaS ▪ Platform as a Service (PaaS): Offers a full cloud-based development and deployment environment,
Market Segmentation of India Cloud Industry enabling rapid app creation with operational savings of up to 50%. Rising demand from startups and AI-
driven applications (e.g., AIPaaS) is accelerating adoption. Examples: AWS Elastic Beanstalk, Azure,
70%
Heroku.
55% ▪ Software as a Service (SaaS): Delivers internet-based software, removing the need for on-prem
hosting. Examples: Office 365, Salesforce.
FY20 ▪ Infrastructure as a Service (IaaS): Provides virtualized compute, storage, and network infrastructure on
demand. Ideal for SMEs and scalable workloads, IaaS supports hybrid workforce needs and removes
22% FY24 the capex burden of owning hardware. Examples: AWS, Google Compute Engine, Azure, DigitalOcean.
16%
12% 13% ▪ Other Segments: Growing adoption seen in Security as a Service (SECaaS), Business Process as a
8% 5% Service (BpaaS), Desktop as a Service (DaaS), and Disaster Recovery as a Service (DRaaS). BpaaS is
gaining traction in customer management and HR.
IaaS SaaS PaaS Others

Source: Nexdigm, Avendus Spark Source: Nexdigm, Avendus Spark

This file was generated for vishnu.agarwal@wipro.com from Avendus Spark Research website on July 17, 2025 06:30:53 Page 10
DATA CENTER INDUSTRY THEMATIC

AI Adoption- Significant booster to DC capacity requirement globally; 75% of incremental capacity globally would be AI-enabled; AI
workloads consume 3–5x more compute hence require 2-3x more power
The global AI market is expected to rise from $110-130bn currently to $320-380bn i.e., growth Global data center capacity is expected to grow 70% from 2024 to 2027, with AI data centers
of 43% largely led by increased adoption of AI/LLMs growing by 3x over the same period; AI workloads consume 3–5x more compute hence require
2–3x more power
Global AI Market ($Bn.)
Global AI & Non-AI DC Split (%)
320-380
42GW 70GW 100GW

43% CAGR
21%
40%
52% AI Data Centers

110-130 Non-AI Data Centers

79%
60%
48%

Current 2027 2024 2027 2030

Source: JLL, Avendus Spark Source: BCG, JLL, Avendus Spark

India AI industry is expected to grow by 35% over the next 3 years versus global growth of 43% India has the second highest number (8%) of ChatGPT users after USA (19%)
India AI Market ($Bn.) United States
17-22 19%

35% CAGR
India
Countries 8%
with highest
7-9 ChatGPT Users
Brazil
Others
5%
61% Canada
4%
UK
3%
Current 2027

Source: JLL, Avendus Spark Source: Reuters, Avendus Spark

This file was generated for vishnu.agarwal@wipro.com from Avendus Spark Research website on July 17, 2025 06:30:53 Page 11
DATA CENTER INDUSTRY THEMATIC

AI Adoption – India is at the forefront of AI adoption after USA largely led by various government initiatives such as IndiaAI
mission; Sourcing issues, Chip export control could weigh on capacity expansion in India
India’s AI capacity growth will largely be led by Hyperscalers, AI data center capacity is Initiatives leading to increased AI adoption in India
expected to grow at 80% (2024 to 2027) & a further ~40% from 2027 to 2030 ▪ Launched in March 2024 with a Rs. 104bn (~$1.3 bn) budget, aims to build national AI
India AI & Non-AI DC Split (MW) The IndiaAI infrastructure with a key focus on deploying 18,000-29,000 GPUs via public-private
Mission partnerships. Over 34,000 GPUs have been empaneled and 17,000+ are already
installed across facilities run by CtrlS, Yotta, and NTT.
3,000 MW
▪ Indian BFSI, IT services, healthcare, and retail firms are moving beyond POCs to
Enterprise AI production-scale GenAI rollouts (chatbots, KYC automation). Notably, HDFC Bank, TCS,
619 Adoption and Infosys have initiated internal AI model deployment initiatives requiring dedicated
1,825 MW compute.
1,030 MW AI Data Centers
242 ▪ India has over 2,000+ AI-focused startups, many of which rely on colocation or cloud-
Non-AI Data Centers
AI startups hosted GPU clusters due to cost barriers of owning infrastructure. Yotta, AWS India,
41 2,381 and ESDS are offering “GPU-as-a-service” models targeted at this startup ecosystem
1,583
989 ▪ RBI and SEBI are encouraging the use of AI for fraud detection, anomaly detection, and
compliance monitoring. Fintechs (e.g., Razorpay, Pine Labs, Cred) are deploying AI to
Fintech
monitor real-time transaction patterns -necessitating low-latency, always-on GPU
2024 2027 2030 compute in India-hosted zones due to data localization.

Source: JLL, Avendus Spark Source: Industry, Avendus Spark

Commentary on slowdown of capacity expansion by Hyperscalers

USA AI Diffusion/Exports control on GPUs & its impact on India- GPU availability is the key for ▪ AWS has paused some global data center lease talks, focusing on post-2026 capacity tied to
AI workloads.
AI enabled DC in India
▪ The move reflects routine capacity planning, not a strategic retreat, as AWS tightens pre-
▪ U.S. imposed strict export controls on advanced AI chips (especially GPUs like NVIDIA A100/H100, AMD lease windows and prioritizes efficient, AI-ready deployments.
MI250/MI300) to China and certain countries. The controls aim to limit the diffusion of cutting-edge AI
capabilities for military or strategic uses.
▪ Microsoft canceled hundreds of MW of leases with major colocation operators and paused
▪ The policy is aimed primarily at China but affects global AI infrastructure indirectly due to supply builds in Ohio, Wisconsin, and the UK.
concentration and license restrictions. ▪ The pullback is linked to revised OpenAI workload forecasts and a shift toward optimized,
self-built infrastructure.
▪ global supply bottlenecks, pricing inflation, and licensing complexities are creating delays and cost
escalations for India’s AI data center projects.
Hyperscalers remain AI-invested but are rebalancing colocation vs. captive strategy.
▪ Lead times for H100-class chips in India are now 6–12 months, even for non-restricted buyers. Colocation providers face pressure to offer high-density, liquid-cooled, power-secure
Implication
capacity. Near-term caution reflects capacity recalibration but not long-term demand
softness.
Source: Industry, Avendus Spark Source: Industry, Avendus Spark
This file was generated for vishnu.agarwal@wipro.com from Avendus Spark Research website on July 17, 2025 06:30:53 Page 12
DATA CENTER INDUSTRY THEMATIC

Data Localisation- A Key priority for Government - India holds only 3% of global data center capacity but drives 21% of global mobile
data consumption; Indian Government wants to expedite shift of India data stored outside the country to domestic DCs
India has 3% share of the global operational DC capacity with USA leading the pack followed by China Majority of the Indian data is stored outside its territory due to higher usage of
hyperscaler services; Although 20% of global data demand is from India, only 3%
is stored in India Data centers
Mobile data Mobile data Data Center
Data Center
Others  8,635
CY2024 consumption consumption Capacity
Capacity (GW)
 21% (EB/Month) contribution % contribution %

 1,228 Americas 15
15 12%
12% 20
20 48%
48%
UK
 3%
EMEA 28
28 23%
23% 99 23%
23%
Canada  1,258
+ LATAM
 3%
 4,231  1,500
APAC ex India 54
54 44%
44% 11
11 26%
26%
Europe Japan
 10%  4%
India 25
25 21%
21% 11 3%
3%
4,500 
China
 16,363 11%  Source: Cushman & Wakefield, Ericsson, Avendus Spark
USA
 39%
India’s high Petabyte/MW ratio indicates a mismatch between its large data
 1,561 generation and insufficient colocation capacity, especially compared to peers
APAC
 4% like China
1,030  India Country-wise Data consumption versus Datacenter capacity (Petabyte/MW) 19
3% 
16

Australia  1,300
5
3
 3% 0.1 0.3 0.3 0.4 1 1

USA

S. Korea
France
Australia

UK

Indonesia
China

India
Japan
Singapore
 Installed capacity in 2024 (MW)
 Country wise capacity contribution %

Source: Cushman & Wakefield, Avendus Spark Source: Cushman & Wakefield, Avendus Spark; Note: 1mn GB = 1PB
This file was generated for vishnu.agarwal@wipro.com from Avendus Spark Research website on July 17, 2025 06:30:53 Page 13
DATA CENTER INDUSTRY THEMATIC

Central Government Policies on Data localisation: RBI and SEBI data localization mandates have been key catalysts for domestic
data storage demand, with the DPDP Act expected to further accelerate this trend

Data localization mandate for the SEBI mandate to all its regulated The Digital Personal Data
payments sector made by the RBI, 2018 entities to store data in India, 2023 Protection Act (DPDP), 2023

Mandate: RBI’s 2018 directive required all payment data to be Mandate: SEBI required all regulated entities—stock exchanges, Objective: The Digital Personal Data Protection Act (DPDPA), 2023
stored exclusively in India. brokers, mutual funds, depositories, KYC agencies, etc.—to store governs the processing of digital personal data to protect
Cross-border Clause: Copies for foreign transactions allowed only and process data within India. individual privacy while enabling lawful use by organizations.
after local storage. Timeline: Entities were given a 12-month window to comply. Key Provisions: Establishes clear obligations for data fiduciaries on
Enforcement: Compliance made a licensing and operational collection, storage, and processing. Grants individuals rights to
Access Flexibility: SEBI does not restrict data access from abroad,
condition for PSOs. consent, access, correction, and erasure.
but mandates domestic storage.
Penalties: Non-compliance led to onboarding bans on Mastercard Data Localization: While it does not explicitly mandate
and Amex (2021). Objective: To ensure SEBI’s unrestricted access, including search localization, it empowers the government to restrict cross-border
Impact: Triggered a surge in demand for domestic data center and seizure rights, even when entities adopt cloud services. data transfers to specific countries. This creates implicit incentives
capacity, as firms scrambled to localize data. for domestic data storage and processing.

Draft Data Center Cloud adoption in Public

IndiaAI Mission
Policy 2020 Services in Sectors

Objective: India’s Data Centre Policy 2020 supports the vision of a Drivers: Legacy IT constraints and evolving privacy regulations Mission: The IndiaAI Mission, supported by a ₹219 bn MeitY
$1 trillion digital economy. are pushing government departments toward secure, scalable budget for FY25 (₹5 bn allocated for AI), targets building a $1
cloud platforms. trillion AI economy by 2035.
Investment Impact: Aims to attract $4.9 billion+ in data center
Policy Push: The Government is building a strong cloud ecosystem
investments through key enablers. Infrastructure Push: Plans include deployment of 10,000+ GPUs,
via the GI Cloud (Meghraj) initiative to enhance e-governance and
Key Enablers: Infrastructure status for data centers; Single- launch of AI-as-a-Service, and creation of a centralized AI-ready
optimize ICT spend.
window clearances and faster approvals; Dedicated data center datasets platform.
Adoption: As of October 2024, over 1,895 departments and
economic zones Impact: Expected to drive substantial demand for high-
28,000+ virtual machines are using Meghraj.
Operational Benefits: Enhances ease of doing business, ensures Procurement Framework: To simplify adoption, MeitY has performance, scalable, and compliant cloud and data center
reliable power and connectivity, and encourages domestic empaneled 22 Cloud Service Providers on GeM, covering public, infrastructure across India.
manufacturing of IT and power equipment. private, and community cloud models.
Source: The Gazette Of India, MeitY, Industry, Avendus Spark
This file was generated for vishnu.agarwal@wipro.com from Avendus Spark Research website on July 17, 2025 06:30:53 Page 14
DATA CENTER INDUSTRY THEMATIC

State-level incentives- Subsidized land banks and electricity duty exemptions- are emerging as key enablers for accelerating data center capacity expansion across India

India has 3% share of the global operational DC capacity with USA leading the pack followed by China
Development Electricity Infra- Development Electricity Infra-
Stamp duty Capital Power Green Ease of Stamp duty Capital Power Green Ease of
/ FSI related duty structure Tax Benefits / FSI related duty structure Tax Benefits
exemption Subsidy Subsidy Incentives approvals exemption Subsidy Subsidy Incentives approvals
incentives exemption Support incentives exemption Support
West

North
Maharashtra ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ Haryana ✓ ✓ ✓ ✓ ✓ ✓ ✓

Tamil Nadu ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓

Central
Uttar
Pradesh
✓ ✓ ✓ ✓ ✓ ✓
South

Telangana ✓ ✓
West

East
✓ ✓ ✓ ✓ ✓ ✓
Karnataka ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ Bengal

State Tamil Nadu Uttar Pradesh Telangana West Bengal Maharashtra

Policy Launch Year 2021 2021 2016 2021 2023

100% exemption on electricity duty;
Lifetime electricity duty exemption; Rs.
100% subsidy on electricity duty; Dual Dual Power Grid availability; Electricity at cost of generation; Dual
1/unit power subsidy for 5 years
Power Grid availability; Exemption of Exemption of 50% on wheeling Power Grid availability; Renewable Waiver on electricity duty; Dual Power
Power & Energy 50% on wheeling charges; charges; 100% exemption on energy under open access; Subsidised Grid availability
(outside Zone I); Open-access power
allowed; Support for captive
Concessional open access charges transmission charges; Open access fuel prices
renewable energy
availability

Classified as Essential Services; Separate Infrastructure Category; Classified as Essential Services; Access Classified as Infrastructure; MIDC to
Uninterrupted water supply; Special
Infrastructure building regulations
Uninterrupted water supply; Special Access to high-speed internet and to high-speed internet and continuous develop parks; Relaxed FSI/building
building norms continuous water supply water supply; Special building norms rules; Residential property tax rates

25-50% land subsidy for first 3 DC 100% exemption on stamp duty;

parks; 100% stamp duty exemption on Rs.1/unit tariff subsidy (Zone II);
50-100% stamp duty concession; 25-
first land transaction, 50% on second; Land at subsidized cost by CCITI; Up to Lifetime electricity duty exemption;
Fiscal Incentives 50% land subsidy for first 3 DC parks;
60% interest reimbursement; 7% 50% building fee rebate by CCITI
100% exemption on stamp duty
Residential property tax; Extra FSI;
50% land subsidy in select districts
capital subsidy up to Rs. 200mn Green parks get additional land/infra
(excluding land/building) benefits
Source: Cushman & Wakefield, CBRE, Avendus Spark Research
This file was generated for vishnu.agarwal@wipro.com from Avendus Spark Research website on July 17, 2025 06:30:53 Page 15
DATA CENTER INDUSTRY THEMATIC

Significant growth in Hyperscaler Colocation in India- 4x jump in colocation capacities leased out to Hyperscaler in last five years

Hyperscaler colocation capacities have seen significant growth in last 3-4 years

2016 2017–2019 2020 Onward 2022

AWS enters India via NTT (Netmagic) Slow ramp-up, minimal hyperscale leasing Strategic shift to large colocation deals Present – Explosion in colocation demand

▪ Amazon Web Services (AWS) launched its ▪ Other hyperscalers like Microsoft (Azure) ▪ Equinix's acquisition of GPX in 2020 ▪ Hyperscalers started taking 100+ MW
first Indian region in Mumbai in June 2016. and Google Cloud also entered the Indian enabled interconnection at scale. capacities, particularly in Mumbai,
market, but initial deployments were Hyderabad, and Chennai.
▪ Instead of building its own facility initially, ▪ Hyperscalers began pre-committing large
limited.
AWS leased colocation space from NTT- MW blocks to colocation providers due to ▪ AI/GenAI workloads, data sovereignty
Netmagic. ▪ Infrastructure was either owned or leased demand spikes, speed-to-market, and land mandates, and latency needs accelerated
modestly from players like STT GDC, CtrlS, constraints. this shift.
▪ This marked the first significant hyperscaler
and Netmagic.
colocation engagement in India, laying the ▪ Providers like STT GDC, Nxtra, Yotta,
foundation for others. AdaniConneX, and Web Werks gained from
these hyperscale deployments.

Number of players in Hyperscale DC development increased from 5 in 2019 to 15 in 2024 DC capacity under colocation in India: 4x jump in colocation capacities leased to
owing to burgeoning demand hyperscalers in last five year
Number of players in Hyperscale DC Development Hyperscaler Colocation Capacities in MW
15
528
Industry Colo Capacity in MW
Hyperscaler Contribution to Overall Colo Capacities
8
231
5
115

43% 54%
35%
2019 2021 2024 2019 2021 2024

Source: JLL, Avendus Spark Source: JLL, Avendus Spark

This file was generated for vishnu.agarwal@wipro.com from Avendus Spark Research website on July 17, 2025 06:30:53 Page 16
DATA CENTER INDUSTRY THEMATIC

Power and Land Constraints in Singapore’s aided in hyperscalers shifting DC capacities to India

Hyperscaler Locations under Colocation in India Power Constraints & Their Impact on Singapore's DC Market & Benefits conferred to DCs in India

▪ Land scarcity and power limitations—particularly availability of high-voltage power for large-scale
campuses—have long been challenges in Singapore.
▪ In 2020, the Singapore government imposed a moratorium on new data center construction due to:
Background ̶ Excessive power consumption (~7% of total electricity use)
̶ High water usage (for cooling)
̶ Sustainability concerns

▪ Land scarcity and power limitations—particularly availability of high-voltage power for large-scale
Delhi
campuses—have long been challenges in Singapore.
Moratorium & ▪ The moratorium was lifted in July 2022, but new guidelines were introduced:
Lifting (2022) ̶ New builds must be energy-efficient (PUE ≤ 1.3)
̶ Projects undergo stringent environmental and land-use vetting
̶ Only ~60MW of new capacity/year is permitted under RFPs (Request for Proposals)
Kolkata

Mumbai ▪ Limited new capacity: DC operators face caps on power allocation, slowing hyperscale expansion.
Pune ▪ Regional spillover: Hyperscalers are shifting capacity to nearby countries like Malaysia (Johor), Indonesia
Hyderabad
Implications (Batam), and India (Mumbai/Noida).
▪ Sustainability mandates: Operators are investing in green DCs, liquid cooling, and renewable energy
integration.

Chennai

Example ▪ Equinix, STT GDC, and Digital Realty have had to prioritize ultra-high-efficiency builds.
Players ▪ Amazon, Microsoft, and Google have adjusted regional deployment strategies by expanding in India,
Affected Malaysia, and Thailand
Current Future Local Zones

Source:, Databyte, Avendus Spark Source: Avendus Spark

This file was generated for vishnu.agarwal@wipro.com from Avendus Spark Research website on July 17, 2025 06:30:53 Page 17
 DATA CENTER INDUSTRY THEMATIC

India as a hub for Global DCs: Low capex/electricity cost, uninterrupted power supply and improved capabilities could result in India
becoming preferred destination for DCs
Data center development cost ($/W) is one of the lowest in India at $7/W with only China’s India’s electricity cost is lower by 20% versus USA. Among APAC nations, ASEAN electricity
cost being lower cost is cheaper versus India

Capex Cost for DC's Across Countries ($/W) Electricity Cost (USD Cents/kWh)
14 45

11 11
10 10 10
26
6 7 21
16
11 13
7 9

China India South Korea Australia USA Indonesia UK Japan Indonesia China South Korea India USA Japan Australia UK

Source: DCCI, Avendwus Spark Source: GlobalPetrolPrices, Avendus Spark

India has 20 sub-sea cable stations and will see Power generation capacity is expected to increase by 10% CAGR Transmission Infrastructure Bottlenecks for Data Centers in India
5 more sub-sea cables over CY25-27E from FY25 to FY30E ▪ Grid Connectivity Delays: Availability of high-capacity transmission lines
is often a constraint, especially in urban hubs like Mumbai, Chennai, and
Major Submarine Cables in India India Power Generation Capacity (GW) Noida.
▪ Lack of Substation Proximity: Many upcoming data center parks require
Pipeline 820 dedicated 220 kV/400 kV substations. However, land availability and
3 ~10%
regulatory hurdles slow down commissioning.
Existing
▪ Limited Redundancy: In several metros, dual-feed power redundancy is
unavailable or cost-prohibitive, which is required for Tier III/IV data
476 centers.
2 371 ▪ State-Level Variations: While MH & TN have invested in upgrading grid
13 268 capacity, other regions lag in evacuation infrastructure.
▪ Green Energy Wheeling Constraints: Developers looking to use
7 renewable power via open access often face transmission corridor
bottlenecks, curtailment risks, or regulatory caps on wheeling quantum.
▪ Developers are now lobbying for dedicated transmission corridors and
Mumbai Chennai FY15 FY20 FY25 FY30 priority grid approvals for data center parks, especially in Mumbai,
Chennai, and Hyderabad.
Source: Submarine Cable Map, Avendus Spark Source: NPP, Avendus Spark
This file was generated for vishnu.agarwal@wipro.com from Avendus Spark Research website on July 17, 2025 06:30:53 Page 18
Assessing Supply Cycle and Competitive Landscape

This file was generated for vishnu.agarwal@wipro.com from Avendus Spark Research website on July 17, 2025 06:30:53
DATA CENTER INDUSTRY THEMATIC

India Colocation Datacenter Supply - Current data center capacity stands at ~1 GW with a vacancy rate of 4%; Expected to triple to
3 GW by 2030E, driven by hyperscaler demand, rapid cloud adoption, and AI-led workloads
Since 2019, India’s data center IT capacity has grown by 24%; we expect this momentum to continue with a ~20% CAGR over the next three years on a high base of 1 GW, reaching 3–3.5 GW by 2030
Third party (Colocation) DataCenter Capacity (MW)

Stage 1 Stage 2 Stage 3

2016 Onwards
(Pre 2007) (2007-12) (2012-16)

Captive DC Rising data needs drove Global cloud Rising data needs led to hyperscale DC construction, with occupiers Area Occupied in msf
growth corporates to outsource players adopting hybrid and multi-cloud models.
peaked post storage to colocation DCs. entered India
dot-com and Data usage soared due to falling prices, post-demonetisation digital
boom in India entered the 3G era in promoted payments, and Reliance Jio’s low-cost 4G launch in 2016.
2005-06. 2008 with MTNL and BSNL cloud
launching 3G mobile and awareness. AI: Generative AI has opened a world of possibilities, resulting in a
Broadband data services. significant surge in compute usage across multiple applications;
policy laid Widespread expected to create a second stream of data center demand 3,000-3,500
the adoption of comparable to the cloud
groundwork smartphones
for India’s and the 2012 795MW+ /
internet launch of 4G 21% CAGR
expansion. services
drove a surge 1,825
in high-speed 24% 1,570
data usage. CAGR
1,279
1,030
854
722
565
449
346
122 207
2 50

2000 2007 2010 2014 2019 2020 2021 2022 2023 2024 2025 2026 2027 2030

0 1 2 3 5 7 9 10 11 13 16 19 22 37

Source: CARE, JLL, Avendus Spark Estimates

This file was generated for vishnu.agarwal@wipro.com from Avendus Spark Research website on July 17, 2025 06:30:53 Page 20
DATA CENTER INDUSTRY THEMATIC

Demand from AI and Cloud to support 25%+ CAGR Growth for Data centers in the next decade outstripping the likely supply

Market Overview
▪ India has a large consumer market and proficient talent pool
▪ Many Cloud Players are setting up Hyperscale facilities via either Captive or Co-location data centres
▪ Supply shortfall of at least 1,500 MW over the next 10 years

Additional Demand and Supply mismatch in next 10 years

7,000 MW

6,000 MW
Demand / Supply Gap
1,542 MW
5,000 MW

4,000 MW

3,000 MW 6,043 MW 2,628 MW Land Based Potential Supply

2,000 MW

1,000 MW 1,873 MW
Under Construction Supply till 2028
0 MW
Demand Potential till 2033 Supply Potential

Total Investment Required in next 10 years: 6,043 MW x US$ 6 mn = ~US$ 36 bn

Source: JLL, Avendus Spark

This file was generated for vishnu.agarwal@wipro.com from Avendus Spark Research website on July 17, 2025 06:30:53 Page 21
DATA CENTER INDUSTRY THEMATIC

Mumbai and Chennai account for ~65% of India DC capacity; More than 50% of DC stock has capacity of more than 20MW

Highest capacity in Mumbai & Chennai largely due to sea landing cables and availability of Chennai has the highest vacancy of 9% whilst other cities have vacancy rates <5%
uninterrupted power Vacancy Across Cities
Others; 14; 1%
Kolkata; 8; 1% Pune 1%
Bengaluru; 81; 8%
Mumbai; 536; 52% Bengaluru 1%

Mumbai 2%
DC Capacity
Delhi; 114; 11%
(MW) & India
Hyderabad 2%
Share (%)
All India 4%
Hyderabad; 55; 5%
Kolkata 4%
Pune; 109; 11% Chennai 9%
Chennai; 113; 11%
Source: JLL, Avendus Spark Source: JLL, Avendus Spark

>20MW i.e., large-sized DCs accounted for 56% of the overall capacity across top cities versus More than 5MW IT load (Hyperscalers) has tariff of $70-80/KW/month versus $110-
42% in 2020 120/KW/month for less than 250KW requirement (Enterprises)
Size-wise split of DC capacity across top cities in India Tariff per month in Rs. Mumbai Chennai Pune Hyderabad Bengaluru Kolkata
7% 5% <4MW
<250 Kw 9,750 8,625 7,688 8,625 8,250 9,000

39% 4-20MW 250 kW-1 MW 8,500 8,250 7,238 7,800 7,988 7,875
51%
1-5 MW 7,500 7,500 6,750 7,050 7,500 -
21-50MW
41%
32% 5+ MW 7,000 6,563 6,188 6,750 6,638 -
>50MW
10% 15% Average Pricing 8,188 7,734 6,966 7,556 7,594 8,438
2020 Current Note: The above-mentioned numbers are based on INR/KW/Month basis available market data on likely achievable rates. The
above pricing assumes standard racks between 5kVA-6.5kVA without any customization.
Source: Colliers, Avendus Spark Source: JLL, Avendus Spark
This file was generated for vishnu.agarwal@wipro.com from Avendus Spark Research website on July 17, 2025 06:30:53 Page 22
DATA CENTER INDUSTRY THEMATIC

Hyperscalers now account for ~54% of India’s data center demand, up from 43% in 2021. BFSI and Tech combined contribute ~30%,
underscoring the growing dominance of hyperscale workloads
Hyperscale, BFSI & Technology accounted for 84% of demand in 2024 with the least demand Hyperscale witnessed the highest demand mix increase to 54% from 43% in 2021
coming from Entertainment, Healthcare & Energy
BFSI DC Capacity Demand
18%

Others
19% 19% 16%
34%
11% 11% 12% Technology
Hyperscale User Demand Technology
54% 2024 8% 20% 20% 18%
12%
15% Banking & Financial Services
Telecom; 4%
Retail & E-Commerce; 4% 50% 50% 54%
43%
Hyperscale
Other; 4%
Entertainment & Media; 2%
Healthcare; 1% 2021 2022 2023 2024
Energy; 1%
Source: JLL, Avendus Spark Source: JLL, Avendus Spark

Annual investments in DC of around $1-1.5bn over last two years are expected to increase to NTT & STT are the largest players accounting for ~50% of India’s capacity. Top five players
USD2-3bn annually over next five years account for ~75% of the industry
PDG, Colt,
Investments in DC (USD Bn) Top 5 Players Capacity Share Digital Connexion,
AdaniConnex,
20-25
75%
Web Werks, Equinix,
ESDS, Pi,
28% NextGen etc
24%
20%
1.2
0.8
0.6 10%
0.4 0.4 9% 9%
0.1

2019 2020 2021 2022 2023 2024 2025-30 (Agg.)

NTT STT SIFY CTRLS Nxtra Others

Source: CBRE, Avendus Spark Source: Company, Avendus Spark

This file was generated for vishnu.agarwal@wipro.com from Avendus Spark Research website on July 17, 2025 06:30:53 Page 23
DATA CENTER INDUSTRY THEMATIC

Supply announcements by new players to result in capacity share of top six players declining to 60% by 2027E versus 85% currently

Upcoming DC capacities would be outside Mumbai and Chennai, with these cities annual Strong visibility on leasing from Hyperscalers for upcoming supply
contribution declining from 65% in 2024 to 50% in 2027 Metric Value
Contribution to capacity addition during the year
Total upcoming DC capacity (2025–27) ~1,000–1,200 MW
249 291 255
Delhi + Others
MW MW MW
Leased/Pre-committed by Hyperscalers ~600–700 MW
Kolkata
% of upcoming supply 50–60%
Bengaluru
23% Source: JLL, Avendus Spark
18% Hyderabad

Pune New entrants in Industry announcing DC capacities in last few years

21%
Year of
49% 51% Chennai Company Key Locations Target Capacity
announcement
30%
Mumbai Digital Connexion (Brookfield, Digital
Mumbai, Chennai 250 MW 2023
Realty & RIL)
2025 2026 2027
Bridge Data Centres Mumbai, Navi Mumbai 100 MW+ 2023
Source: JLL, Avendus Spark
Digital Edge Navi Mumbai 216 MW 2023

Supply announcements by new players to result in capacity share of top six players declining NexTGEN (Next-Gen Infra) Hyderabad, Bengaluru NA 2023
to 60% by 2027E versus 85% currently
Mantra Data Centers Tier-2 cities Edge-focused 2023
India Datacenter capacity share
Trimax Data Centres Mumbai Small-scale colocation 2023

15% Techno Digital Infra Chennai, Pune, Noida 250 MW 2024

40%
CapitaLand DC Mumbai, Hyderabad 150 MW 2024
Rest of industry
Lumina CloudInfra Pan-India (Mumbai, Chennai) NA 2024
85% Top six players BDx Data Centers Mumbai (acquired from Nxtra)* NA 2024
60%
L&T Cloudfiniti Pan-India (starting with Mumbai) NA 2024

ECS Data Centers Mumbai, Chennai NA 2024

2024 2027E
Colt DCS-RMZ JV Mumbai, Chennai 250 MW 2024
Source: CBRE, Avendus Spark Source: Company, Avendus Spark * Expansion in South India
This file was generated for vishnu.agarwal@wipro.com from Avendus Spark Research website on July 17, 2025 06:30:53 Page 24
DATA CENTER INDUSTRY THEMATIC

Visible Rental pressure: Colocation rentals are down 20% since 2019; Hyperscalers rentals are 20-30% lower compared to enterprises

Pricing in Pune declined by 5% given the sharp rise in hyperscale mix from 51% to 69%
Rentals have declined the most in the <250kW; pricing appreciated in the 1-5+ MW range
CAGR
2021 2022 2023 2024 All India Rentals (Rs./kWh/Month)
(2021-24) Hyperscale Mix
<250 kW 10,022 9,750 9,750 9,750 -1%
43%
Mumbai (Rs.)

250 kW-1 MW 7,980 8,438 8,438 8,500 2%

50% 50% 54%
1-5 MW 6,867 7,313 7,313 7,500 3% 7,915
7,592 7,592 7,683
5+ MW 6,495 6,938 6,938 7,000 3%

Average 7,841 8,109 8,109 8,188 1%

<250 kW 10,950 8,625 8,625 9,750 -4%

Chennai (Rs.)

250 kW-1 MW 8,908 8,250 8,250 8,438 -2%

1-5 MW 7,238 7,500 7,500 7,313 0% 2021 2022 2023 2024

5+ MW 6,124 6,563 6,563 6,938 4% Source: JLL, Avendus Spark

Average 8,305 7,734 7,734 8,109 -1%
Pricing declined by 4% YoY in 2022 given the Colocation rentals for both Hyperscalers and
<250 kW 10,764 7,688 7,688 7,688 -11% increasing mix of hyperscaler Enterprises are down 20% in last five years
250 kW-1 MW 8,723 7,238 7,238 7,238 -6% Rentals (Rs.) 2021 2022 2023 2024 Change Avg. Colo Rentals ($/kW/Month)
Pune (Rs.)

1-5 MW 6,867 6,750 6,750 6,750 -1% 120

<250 kW 10,356 8,588 8,588 8,813 -5% 100
5+ MW 5,753 6,188 6,188 6,188 2% 90
Average 8,027 6,966 6,966 6,966 -5% 70
250 kW-1 MW 8,314 7,943 7,943 7,993 -1%

<250 kW 10,022 8,250 8,250 8,250 -6%

1-5 MW 6,867 7,223 7,223 7,223 2%
Bengaluru (Rs.)

250 kW-1 MW 7,980 7,988 7,988 7,988 0%

1-5 MW 6,867 7,500 7,500 7,500 3% 5+ MW 6,124 6,615 6,615 6,703 3%

Hyperscale Retail
5+ MW 6,495 6,638 6,638 6,638 1%
India Colo 2019 Current
7,915 7,592 7,592 7,683 -1%
Average 7,841 7,594 7,594 7,594 -1% Rentals
Source: JLL, Avendus Spark Source: JLL, Avendus Spark Source: JLL, Avendus Spark
This file was generated for vishnu.agarwal@wipro.com from Avendus Spark Research website on July 17, 2025 06:30:53 Page 25
DATA CENTER INDUSTRY THEMATIC
Risk to Colocation DC players with planned Hyperscalers captive Datacenters: Negotiation pressure from hyperscalers expected to
increase.; Hyperscalers are expected to prioritize captive data centers for internal workloads over public cloud expansion
Land bank buying by hyperscalers in last few years Impact of Hyperscaler “Own and Operate” model on Third-Party Colocation
(Colo) Providers
▪ Acquired a 54-acre parcel in 2022 in Thane for Rs.19bn for
hyperscale campus ✓ Hyperscalers typically lease multi-MW blocks (10–50 MW) — they are the largest revenue contributors for
colocation providers.
▪ In 2023, Amazon Web Services (AWS) announced it plans to
✓ With higher “own and build” capacity, anchor leases could be shorter tenor, leading to lower rental visibility for
Amazon Web invest US $12.7 billion into cloud infrastructure in India by 2030
colo players.
Services (AWS) to meet growing customer demand for cloud services in India.
✓ Hyperscalers that still lease DC space negotiate aggressively on pricing. With growing self-sufficiency, they have
Mumbai, ▪ Purchased 38-acre land in Palava (near Mumbai) for a future more leverage, pushing colo providers to offer- Lower per-kW rentals and Custom build-outs with low margins
hyperscale campus, and planning another US $2 billion to build ✓ Hyperscaler-owned DCs often come with custom power architecture, GPU cooling, internal orchestration, etc.
Hyderabad, Palava
more data centers in Telangana. Colocation facilities, even Tier III+, may lack software-defined integration or AI readiness.
✓ Players like Nxtra, STT GDC, and Yotta expanded aggressively assuming steady hyperscale demand. If captive
adoption rises quickly, India may face a supply glut.

▪ Announced three additional captive DC campuses in Hyderabad Our discussions with industry players indicate that third-party colocation
(~100 MW each), expanding to six sites over 10–15 years. players would still have some offsets
Microsoft Azure ▪ Microsoft to invest US $3 billion to expand Azure cloud & AI
✓ Even with owned DCs, hyperscalers use colocation for faster market entry or edge deployments.
Pune, Mumbai, capacity in India.
✓ Hyperscalers are increasingly building captive data centers in India primarily to host their own application
Chennai, 6x in ▪ 48 acres near Hyderabad (Ranga Reddy district. workloads (e.g., search, e-commerce, video, AI inference), rather than to expand general-purpose cloud
Hyderabad (IaaS/PaaS) availability for external customers.
✓ The surge in AI workloads (GPU, storage, networking) means even self-built DCs may need supplemental colo.
✓ Colo providers with better local approvals or fiber networks may still win regional workloads.

Few steps taken by colocation DC players to offset likely impact of “Own and
Operate” capacities by hyperscalers
▪ In talks to acquire 22.5-acre land in Juinagar, Navi Mumbai for
its first custom-built captive DC, estimated capacity ~200 MW ✓ Move up the value chain by offering managed services, GPU-as-a-service, AI-ready pods.
Google
✓ Partner with hyperscalers as satellite zones or edge facilities.
Navi Mumbai ✓ Target high-growth verticals such as Government cloud, fintech, OTT, gaming.
✓ Focus on Tier-2 cities where hyperscalers will not deploy directly.

Source: Industry, Avendus Spark Source: Industry, Avendus Spark

This file was generated for vishnu.agarwal@wipro.com from Avendus Spark Research website on July 17, 2025 06:30:53 Page 26
DATA CENTER INDUSTRY THEMATIC

Established and New entrants in colocation data center - Current and Upcoming capacity expansion plans

Current Upcoming
Key Shareholder Long-term Capacity Announcements
Capacity (MW) Capacity (MW)

▪ NTT DATA will invest $1.2 billion to build a 400 MW AI data center with 25,000 GPUs in Hyderabad, with Neysa and
NTT NTT Data Inc (Japan) 290 410
the Telangana government.

Nxtra Airtel (76%), Carlyle (24%) 88 312 ▪ Set to invest Rs 50bn over the next three years.

ST Telemedia (74%), Tata Comm ▪ Announced USD 3.2bn (Rs.260bn) for expansion over the next 5-6 years. Immediate capex of approximately
Established

STT GDC 205 346

(26%) Rs.57.5bn over FY25 to FY27.

▪ A group company is constructing DC VI and DC VII at an estimated cost of Rs. 2.4bn, with completion expected by Q2
CtrlS Pioneer Group, Sculptor (29%) 90 910 FY2026. It aims to double its current operational capacity of ~90 MW over the medium term. The company has
149 MW under contract and is finalizing an additional 50 MW.

▪ The company aims to scale to 407+MW by 2025, with campuses expandable to 970+MW. Key sites: Mumbai,
Sify Promoter (84%), ADRs (16%) 111 568
Hyderabad, Noida, Chennai, Bengaluru.

Yotta Hiranandani Group 34 946 ▪ Aims to deliver a total of 1030 MW of data center capacity across key regions by 2027.

▪ The company is investing ~$1 billion in Chennai and Mumbai, with a 150 MW Mumbai campus. Two phases (~46
Princeton DG Warburg Pincus, OTPP, Mubadala 25 200
MW) are near completion and fully pre-leased; revenue starts H1 FY26.
New Entrants - PE Backed

Blackstone RE & Tactical ▪ The company plans to develop ~600 MW of data center capacity across India, starting with Mumbai and Chennai,
Lumina - 500
Opportunities Fund followed by Pune, Delhi/NCR, and Hyderabad.

▪ The company may invest an additional ~Rs. 70 bn over the next 3-5 years. It has secured a construction loan for
CapitaLand CapitaLand - 244
Phase 1 of CapitaLand DC Navi Mumbai 1.

▪ Anticipates investing further capital to expand its footprint of wholesale and custom-build hyperscale solutions to
Bridge DC Bain Capital - 20
more than 300 MW over the next two years across Asia, including India.

ESR SWFs, Pension Funds - 35 ▪ Additional facility with a 19.2 MW IT load capacity is projected to be ready for service by May 2025.

Source: Industry, Avendus Spark

This file was generated for vishnu.agarwal@wipro.com from Avendus Spark Research website on July 17, 2025 06:30:53 Page 27
DATA CENTER INDUSTRY THEMATIC

Established and New entrants in colocation data center - Current and Upcoming capacity expansion plans

Current Upcoming
Key Shareholder Long-term Capacity Announcements
Capacity (MW) Capacity (MW)
▪ Colt DCS has formed a $1.7 billion JV with RMZ to expand in India. Initial investments focus on scaling existing sites in
Colt Fidelity Investments 22 183
New Entrants - Global Players

Navi Mumbai and Ambattur, Chennai, with plans for a third location.
Digital RIL, Digital Realty, Brookfield ▪ Its first data center campus in Chennai is designed to cater to a 100 MW IT load, with its initial phase already
20 120
Connexion Infrastructure boasting a 20 MW capacity. Mumbai DC will have 40MW of capacity
ACX AEL (50%), Edge Connex (50%) 17 2,116 ▪ Aims to deliver DC Capacity of 1 GW by 2030. Planning to invest ~US $5 Bn over the next 5 years
▪ The company continues to invest in Mumbai and Chennai IBX sites and is evaluating additional metros across India
Equinix Equinix Pacific LLC 10 8,225 cabinets
for further growth
▪ Plans for hyperscale deployments in Hyderabad, Bengaluru, Chennai, and Delhi NCR. Yondr is reportedly negotiating
Everyondr Everstone, Yondr - 60
an exit from the JV, allowing Everstone to assume full ownership.
Digital Edge (50%), NIIF (45%), ▪ Investing Rs. 48bn in a 300+ MW Navi Mumbai campus over 7-10 years. BOM1 (18.32 MW) and BOM2 (~48 MW) are
Digital Edge - 300
AGP (5%) under construction and fully pre-leased.
Iron Mountain (54%), Promoters ▪ Web Werks (with Iron Mountain) intended to increase capacity to 200 MW by 2026, backed by ~$1 billion
Web Werks 11 189
(46%) investment and $150 million equity from Iron Mountain.
ESDS Promoter (55%), GEF Capital (41%) 7 30 ▪ The company aims to scale from 1 GW to 10 GW of capacity over eight years to meet rising AI compute demand.
EVP & MLC PE (28%), Angel
Pi 8 - ▪ Pi Datacenters is planning to invest Rs 2bn (~$24m) in developing a new data center in Hyderabad, India.
Investors (24%)
▪ NxtGen had planned a ~$175 M investment to build 10 new data centers and 236 edge facilities across Chennai,
NxtGen Intel Capital, IFC, Iron Mountain 2,000 Racks -
Hyderabad, and Visakhapatnam
Small-Scale

Promoters (63%), Futuristic Capital

Cyfuture 5 1,000 Racks ▪ Building a new Chennai site (500 racks Phase I), and expanding to Mumbai and additional Bangalore capacity.
(21%)
▪ L&T is investing approximately Rs. 36bn (~$415 M) to build three new data centers in Bengaluru, Panvel, and
L&T Cloud Larsen & Toubro 30 120
Mahape, aiming to boost capacity from 30 MW to 150 MW by 2027.
▪ Launching with a US $1 billion investment to build a network totaling 250 MW of combined hyperscale and edge
Techno
Promoters - 250 data center capacity across India. Partnered with RailTel to deploy 102 edge data centers across 23 states, targeting
Electric
low-latency services in Tier II/III cities
▪ The company has committed $2.1 bn (Rs. 180bn) capex for its DC vertical aiming to achieve operational capacity of
Anant Raj Promoters 6 300+
~310MW by 2031.
Nexus (16%), Z47 (16%), NTTVC (15%),
Neysa ▪ Partnering with NTT DATA & Telangana government to build a 400 MW AI data center campus in Hyderabad with an
Anchorage (4%), Founders (43%), Others NA NA
Networks (6%) investment of ~$1.2bn.
Source: Industry, Avendus Spark
This file was generated for vishnu.agarwal@wipro.com from Avendus Spark Research website on July 17, 2025 06:30:53 Page 28
DATA CENTER INDUSTRY THEMATIC

Global DC Overview- Market size of $75bn and likely to grow at 15% CAGR led by high AI workload; Equinix is market leader with 12% market share

DC capacity across regions grew at 15% CAGR over last five years; APAC capacity growth Over the past 4-5 years, global colocation market grew ~10%; Expect revenue growth to
would be highest over next five years increase to 15% CAGR led by higher colo requirement for AI workload

Global DC Snapshot 2019 2024 2027

CAGR CAGR Global Colo Market Size (USD Billion) 146
(2019-2024) (2024-2027)
15%
Global Installed Capacity (GW) 20 42 65 16% 15% CAGR
~10%
- Americas 9 21 29 17% 12% CAGR
73

51
- APAC 5 12 23 18% 24%

- EMEA 5 10 13 14% 11%

Global Vacancy 12% 8% 2019 2024 2030

Source: DC Byte, Cushman & Wakefield, JLL, Avendus Spark Estimates Source: Precedence Research, Structure Research, Avendus Spark

DC capacity to see significant shift with AI workload taking up 35% of capacity versus 15% Top 5 players account for ~35% of the global market with Equinix having the
currently highest share of 12%
Shift in DC Workloads Composition Equinix, 12%

Enterprise Digital Realty, 9%

20%
31%
Major Global
15% 35% AI Workload NTT, 6%
Players & Their
Market Share Digital Bridge, 3%
54% 45% Cloud (AI & Non-AI) Others, 67% Cyrus One, 3%

Current 2030

Source: McKinsey, Avendus Spark Source: NTT, Avendus Spark

This file was generated for vishnu.agarwal@wipro.com from Avendus Spark Research website on July 17, 2025 06:30:53 Page 29
 DATA CENTER INDUSTRY THEMATIC

Stargate is OpenAI’s $500 billion AI super-data center initiative to build the world’s largest compute infrastructure backed by SoftBank, Oracle, and MGX

▪ Stargate traces its origins to late 2022, when OpenAI and Microsoft began conceptualizing a $100 billion-class
supercomputing infrastructure to support frontier AI model development.
▪ As compute demands rapidly intensified, the idea matured into a sovereign-scale initiative aimed at securing long-
term AI capacity beyond traditional cloud dependency.
▪ On January 21, 2025, the project was formally launched as Stargate LLC during a White House briefing led by
President Trump, with an initial $100 billion committed - part of a broader $500 billion plan over four years.
▪ The venture is backed by major private players including OpenAI, SoftBank, Oracle, and MGX, with both OpenAI and
SoftBank contributing approximately $19 billion each and holding ~40% equity.
▪ Masayoshi Son was named Chairman, signaling SoftBank’s financial leadership, while OpenAI assumed operational
control. Key partners like OpenAI, Oracle, NVIDIA, Microsoft, Arm, and MGX are co-developing the compute
backbone, drawing on long-standing hardware partnerships.

Source: OpenAI, Stargate Source: Industry, Avendus Spark

Key Players Stargate 1st Project: Abilene, Texas

▪ OpenAI is a lead partner with operational responsibility for Stargate. They are the Campus Size ▪ 1,200 acres
OpenAI primary beneficiary, requiring this massive compute power to train and operate Developer ▪ Crusoe (U.S.-based data center start-up)
their increasingly complex AI models.
Construction Timeline ▪ Start: June 2024Completion: Mid-2026
▪ Primarily taking on the financial responsibility for the project. SoftBank's CEO, Power Capacity ▪ 1.2 GW
SoftBank Group
Masayoshi Son, is the chairman of the Stargate venture (Stargate LLC).
Building Count ▪ 8 interconnected data center buildings
▪ Oracle is providing its expertise in cloud computing, hardware, and data center
Oracle Compute Deployment ▪ Up to 400,000 GPUs, including 50,000 NVIDIA Blackwell chips
management.
Rack Power Density ▪ Up to 130 kW per rack (vs. 2-4 kW for traditional CPU racks)
▪ An investment firm backed by Abu Dhabi, MGX is contributing sovereign wealth
MGX Cooling System ▪ Closed-loop liquid cooling (1 million gallons total water use)
financing to the project.
▪ The company doesn't have any financial stake; have a right of first refusal on Deployment Milestone ▪ Targeting fastest 100 MW+ data center deployment to date
Microsoft
excess compute.
Stargate 2nd Project: UAE
Nvidia ▪ GPU Supplier
Total Planned Capacity ▪ 1,000MW
Crusoe ▪ Data center construction start-up First Phase ▪ The first phase will of 200MW will be operational by 2026
Source: OpenAI, Avendus Spark Source: Bloomberg, Avendus Spark
This file was generated for vishnu.agarwal@wipro.com from Avendus Spark Research website on July 17, 2025 06:30:53 Page 30
Positioning for the Data Center Supercycle: Direct (Colocation/Cloud) vs. Enabler

This file was generated for vishnu.agarwal@wipro.com from Avendus Spark Research website on July 17, 2025 06:30:53
DATA CENTER INDUSTRY THEMATIC

Business Models in Data Centers: Colocation, Cloud, and Managed Services

1. Colocation is a stable, yield-driven segment backed by long-term contracts with 3. Managed services provide an asset-light, high-margin, annuity-like revenue stream
Various business models enterprises and hyperscalers. by layering value-added offerings on top of existing infrastructure.
to play Data Center
2. Cloud infrastructure is capital intensive but offers scalable RoI as utilization improves 4. Edge data centers present a high-yield, high-IRR opportunity in Tier-2/3 cities, driven
growth story
and services deepen. by latency-sensitive workloads like CDN, gaming, and 5G.

Enterprises (Retail) Hyperscaler (Wholesale) Traditional IT services offering to enterprises (PaaS)

Rental for rack space, power consumption, Rented by Hyperscalers, cloud Outsourced IT operations and support functions offered on top of core
1 and cooling; typically rented by Small to providers, large enterprises having 3 colocation or cloud infrastructure. These services reduce the customer’s
mid-size businesses, startups, SaaS firms, requirement of >250KW-1MW; Long need to manage their own hardware, networks, and data security, while
Banks - Short tenor of 3-5 years tenor of 5-10 years; ensuring 24×7 uptime, compliance, and operational efficiency

2 3
1 Cloud 4
Managed
Colocation Services Edge DC
Services

Offered by DCs to SMEs (IaaS) VMs and services at edge nodes for latency-sensitive workloads
Offer customers access to virtualized Small-scale facility (100KW to 2MW) located closer to end users or devices to
2 computing resources — like VMs, storage, and 4 reduce latency and improve performance. Primary function such as localized
network — using flexible pricing models processing, caching, and connectivity needs — typically within metro cities,
(Subscription and Pay as you go based model) industrial zones, or Tier 2/3 regions. Pricing based on rack per month or power
used per month etc
Source: Industry, Avendus Spark
This file was generated for vishnu.agarwal@wipro.com from Avendus Spark Research website on July 17, 2025 06:30:53 Page 32
DATA CENTER INDUSTRY THEMATIC

Business Model Comparison- Colocation offers yield and can be scaled-up relatively faster whereas Cloud/managed services offer
growth and RoEs albeit with scalability constraints owing to hyper-competition
Colocation Cloud Infra IaaS/PaaS Managed Services

Offer virtualized compute/storage with

Rent out rack space, power, cooling Build and sell bare-metal orchestration and billing (IaaS) or End-to-end IT mgmt (monitoring, backup,
Description
(bare shell) compute/storage/networking layers Prebuilt environments (DB, analytics, OS, security, compliance)
container mgmt, APIs, DevOps)

Elastic billing (per vCPU, RAM, storage),

Revenue Model Fixed rentals (Rs/MW/month) Usage-based (VM/hr, TB/month) Subscription/AMC/project-based
Subscription for PaaS

Build scalable digital infra, asset Compete with public cloud, private
Strategic Value Infra monetization, REIT yield play Compete with IT services companies
monetization cloud offerings

Capital intensity High Very High NA NA

Capex per MW Rs.500-600mn Rs.800mn-1bn NA NA

Max Annual Revenue per

Rs.80-100mn Rs.500-800mn 20-30% higher than Cloud services 30-40% higher than Cloud services
MW

EBITDA Margins % 60-70% (ex electricity passthrough) 30-40% 40-50% 25-30%

Payback period 10-12 years 4-5 years NA NA

RoE metrics Stable, low-risk yield (10-15%) Moderate (20-25%) Moderate (20–25%) Highest (25-30%)

Technological Obsolescence in server,

Margin pressure from Hyperscalers,
High capex, Leasing risk, Rental pressure High capex and underutilisation, Hyper-competition, In-house operations,
Key risk Scale disadvantage, Fast changing tech
due to oversupply In-house DC; Strict Security Human Resource intensive
and integration complexity
requirements

Source: Industry, Avendus Spark

This file was generated for vishnu.agarwal@wipro.com from Avendus Spark Research website on July 17, 2025 06:30:53 Page 33
 DATA CENTER INDUSTRY THEMATIC

Capex cost to set up 1MW of Datacenter and Cloud - Rs.1.5bn for Non-AI colocation & Cloud and Rs.2.5bn for AI cloud colocation & cloud

Capex cost per MW to establish for different type of colocation data centers Break-up of Capex cost per MW for colocation data centers
Data Center Type Build Cost (per MW) Details Land and Component Details
IT
Civil Infra,
Infrastructure ▪ Land price varies heavily by
Tier III Colocation Rs.0.5-0.7bn ▪ Includes IT infra, redundancy, cooling, and security. 20%
(Racks), 15% Land and Civil Infra city (e.g., Navi Mumbai >
Chennai > Hyderabad).
▪ Large scale facilities, often for cloud providers. Higher Electrical & Power ▪ UPS, transformers, DG sets,
capex cost due to higher redundancy, high density racks Networking
Infra power distribution.
Hyperscalers Rs.0.6-0.7bn requiring more cooling and power, Purpose-built for & Security,
cloud-scale workloads, multi-layer physical/logical 15% ▪ CRAC units, chillers,
Cooling Systems
controls and also invest in grid infra precision cooling.
▪ Structured cabling,
15-20% higher than Networking & Security
AI Hyperscalers ▪ No chillers and PAHUs required as it uses liquid cooling surveillance, fire systems.
non-AI Cooling
Systems, Electrical & ▪ Depends on whether it is
Power IT Infrastructure
▪ Containerized or pre-fab structures, lower redundancy 20% white space-only or fully
Edge/Mini Data Center Rs.0.4-0.5bn Infra, 30% (Racks)
and air cooled managed.

Source: Industry, Avendus Spark Source: Industry, Avendus Spark

Capex cost to set up server/cloud cost in a data center AI enabled data center cost is 50-60% higher than non-AI enabled data center
Cloud Capex Non-AI Cloud AI (GPU) Greenfield Capex cost for 1MW DC in Rs.mn
Average rack density in kW 5-10 25-40
Total racks required in 1MW DC 150 30 Rs.2-2.5bn
Power per server 0.3-0.5kW 2.5-5.0kW
Server per rack 15-20 3-5
Rs.1.3-1.5bn Cloud
CPU servers 3,500-4,000 300-400 1500
Cost of 1 server in Rs.mn 0.3-0.4 3-4
Dual Intel Xeon / AMD EPYC NVIDIA A100/H100/L40 800 Colocation
Server specs RAM of 128GB-256GB DDR RAM of 512GB-1TB DDR
Storage 2-4TB SSD Storage 4-8TB SSD
700
Total server cost in Rs.mn 700-800 1,200-1,400 500
Networking + Infra cost 50-80 50-80
Non-AI AI
Overall cost to set up cloud servers in Rs.mn 800-900 1,300-1,500

Source: Industry, Avendus Spark Source: Industry, Avendus Spark

This file was generated for vishnu.agarwal@wipro.com from Avendus Spark Research website on July 17, 2025 06:30:53 Page 34
DATA CENTER INDUSTRY THEMATIC

Operational cost and Margins – EBITDA margins of 60-70% in Colocation assuming electricity pass through and 40-50% in Cloud business

P&L for Colocation DataCenter P&L for Cloud DataCenter

% of Revenue Remarks % of Revenue Remarks

Revenue 100% Revenue 100% ▪ 4-5x of colocation revenues

Cost 30-40%
Cost 50-60%
▪ Electricity costs for servers, cooling, UPS, and
Electricity cost Pass through
lighting, Backup power
Colocation rentals 10% ▪ Cloud players pay to colocation players
Cooling cost 10% ▪ Consumables, Servicing, AMC

▪ Internet Service Provider (ISP) costs, cross Power cost 12% ▪ PUE of 1.5x, Pass through from colocation player
Network & Bandwidth 7%
connect fees, fibre optics
▪ AMC charges for hardware, HVAC, UPS, and
Maintenance & AMC 5% generators, spare parts and hardware Network cost 11% ▪ Mostly pass through
replacement cost

Workforce & Staffing 5% ▪ Salaries and training cost

Software stack cost 10% ▪ Licensing

Security (physical and cyber) 3% ▪ Cameras, Anti-virus, Patches

Employee cost 7% ▪ Salaries and training cost
Cloud & Software Costs (For
3% ▪ Licensing feed or any other automation software
cloud services model)
Misc.(customer ops, infra
10%
▪ Compliance certifications, Disaster Recovery & srvices)
Additional Cost Factors 2%
Backup, upgrades

EBITDA 60-70% EBITDA 40-50%

Source: Industry, Avendus Spark Source: Industry, Avendus Spark

This file was generated for vishnu.agarwal@wipro.com from Avendus Spark Research website on July 17, 2025 06:30:53 Page 35
DATA CENTER INDUSTRY THEMATIC
Return metrics and IRR – Pre-tax yield on investment of 10-12% in Colocation with long gestation period; High RoEs in Cloud Services owing to high asset
turnover (5x-6x of colo) – increasing rack density augurs well for data center return metrics

Business Model Colocation Cloud Services Remarks

Capex required for Greenfield Colocation and Cloud DC; Brownfield capex cost is much lower; Incentives such as capital
AI-led rack densification boosts data center ROI: Higher
subsidy and electricity improve return metrics further densities (30–100 kW vs. 4–8 kW) pack more compute
Capacity 1MW 1MW
per sq. ft., enabling premium rentals, better space
utilization, and higher EBITDA per MW—all without
▪ Cloud capex is higher by 50-60% versus
Capex in Rs.mn 500 800 proportional capex escalation.
Colocation capex
Debt funding 60% 60%
▪ Debt to Equity ratio of 60:40; Low cost of debt
Debt in Rs.mn 300 480
available
Equity in Rs.mn 200 320 AI load requires higher rack
Return metrics potential for Colocation and Cloud density (30-100kW)
▪ Colocation revenue with pricing of Rs.7500
Potential revenue at 90% occupancy /KW/month; Cloud revenue is 5-6x of
81 542
in Rs.mn Colocation revenues; Rental growth of 3-4%
annually
▪ EBITDA margins in both models assuming Higher rack density
EBITDA margins 75% 45%
electricity pass through in colocation implies higher
EBITDA in Rs.mn 61 244
power per sq. ft of
space
▪ Depreciation for colocation at 15-20 years
Depreciation 25 120 (Lower MEP, higher for building/structure)
Servers for Cloud services in 5-6 years

EBIT in Rs.mn 35 124

Interest 27 43 ▪ Interest rate of 9-10%
PBT 8 81
PAT 6 61
Pre-tax Yield on Investment (EBITDA ▪ Colocation is Yield play with pre tax yield of 10-
12% 29%
to Capex) 12% Higher power capacity leading to higher
Return on Equity % 3% ~20% ▪ RoE of 20% in cloud business. rentals/sft & better return metrics
Source: Industry, Avendus Spark
This file was generated for vishnu.agarwal@wipro.com from Avendus Spark Research website on July 17, 2025 06:30:53 Page 36
DATA CENTER INDUSTRY THEMATIC

Overview of Top six Indian Colocation DC players – Revenue CAGR of 20%+ for all top players; Depressed metrics owing to investment phase

2024 Financials NTT STT Sify Infiniti CtrlS Nxtra Yotta (Nidar Infra)

Colocation (Cloud in Colocation (Captive for

Revenue source Colocation Colocation Colocation and Cloud Colocation and Cloud
subsidiary) Airtel)

Capacity- Current 290 205 107 90 88 84

Upcoming/Planned 410 346 568 910 312 946

Revenue in Rs.bn 27 20 11 13 18 (~70% captive) 5

3Y CAGR 31% 19% 26% 20% 18% NA

EBITDA in Rs.bn/EBITDA
10/36% 9/44% 5/41% 6/47% 7/38% 0.9/21%
margins

3Y CAGR 41% 22% 23% 12% 19% NA

PAT in Rs.bn/PAT margins 0.4/1% 1.1/5% 0.6/5% 2.1/16% 2.3/13% (4.5)/NM

3Y CAGR -9% 24% -10% -4% 9% NA

RoE % 2% 4% 4% 16% 9% 29%

Net Debt Rs.bn 60 20 13 15 8 13

Net debt to EBITDA (x) /Net

6.2/3.0 2.3/0.6 2.9/0.7 2.4/1.0 1.1/0.3 13/2.3
debt to Equity (x)
Source: Company, Industry, Avendus Spark
This file was generated for vishnu.agarwal@wipro.com from Avendus Spark Research website on July 17, 2025 06:30:53 Page 37
DATA CENTER INDUSTRY THEMATIC

Datacenter Architecture

The below picture depicts the key components housed in a data center
Monitoring
How DCs work? Control stations for data centers and building security systems serve as central
commands in the data center. All important information is collected through a DCIM
A data Center is a specialized (Data Center Infrastructure Management) system and displayed on screens

computer system that combines high-

Server Rooms
power hardware and availability in a
Data rooms or server rooms have standard servers
controlled environment for the and storage units. The racks are usually located in an
purpose of storing, safeguarding and enclosed space to enable optimal cooling
processing data on a large scale Heat exchangers
On the roof of the data center,
heat exchangers release excess
heat from turbo-cooling units

Power management and

SERVER Meet-Me-Rooms
Connectivity is key for data centers. All these
backup systems. While the connections come together in Meet-Me-Rooms or
generators go through a brief MMRs, so connections can be made easily and
start-up phase, the batteries relatively inexpensively. There is always at least one
provide power allowing the MMR in a data center
system to operate without
any interruption

Fire extinction
All server rooms where data is stored
CRAH/PAHU are equipped with smoke detection
Maintains proper systems that monitor the room 24/7. In
temperature and humidity for the event of a fire, traditional
optimal equipment operation Support Personnel extinguishing systems, such as water or
Professionals in-charge of foam, can cause more damage to a data
operation, maintenance and center than fire itself
troubleshooting at the data
center
Physical security
Interconnection Measures and systems to protect
Enables communication between data centers against unauthorized
different networks and systems inside access and physical threats
and outside the data center
Source: ACS, Avendus Spark
This file was generated for vishnu.agarwal@wipro.com from Avendus Spark Research website on July 17, 2025 06:30:53 Page 38
DATA CENTER INDUSTRY THEMATIC

Selling Shovels in Gold rush: Cooling solution (HVAC) & electric equipment account for 50-60% of colocation data center capex cost

In the Full-stack data center with cloud services servers account for 70% & rest contributed by construction cost (Land, EPC, Electrical Equipment, Cooling Solutions and Other Supporting Infra

65% 1,000 35% 500

Overall Cost Contribution %

Cost in Rs.mn per MW Servers Construction Cost

12% 70 8% 80 10% 50

Land EPC Network Connectivity

Designing, Engineering and Construction System integrators, ISPs, Cabling

25% 125 30% 150 15% 75

Cooling Solution Electrical Equipment Other Infra

Precision Air Handling Unit (PAHU), Chillers/Heat Genset, UPS, Batteries, Switchgears, Transformers, PDUs
exchangers, Cooling towers, Dielectric fluid (liquid cooling) (Power distribution unit)

3% 15 4% 20 4% 20 3% 15 2% 5

DC management infrastructure
Floor tiles Security and Cybersecurity Gas separation units Racks PDU
system
Raised floor, and power panels for CCTVs, Digital systems for cyber Software to monitor power and
Any fire extinguishing activities Host servers, storage etc
cables routing attacks cooling

Source: Industry, Avendus Spark

This file was generated for vishnu.agarwal@wipro.com from Avendus Spark Research website on July 17, 2025 06:30:53 Page 39
DATA CENTER INDUSTRY THEMATIC
DC Ecosystem beneficiaries: As per our channel checks, the following companies are supplying critical components to data centers in India, with the
majority being MNCs
Potential incremental demand (Annually)
Category Functions Companies supplying to DC ecosystem
from Datacenter capex in Rs.bn
▪ Targus, Schneider, AECOM, Jacobs, T&T, Sterling & Wilson, L&T, IDAI, Burns & McDonnell, Panchshil,
Designing and Engineering
Bearys, Planckdot, Prasa
EPC 20-25
Construction ▪ L&T, Sterling & Wilson, KEC, JMC, SW, CCCL

IT equipment Networking ▪ Dlink, Comm scope, Panduit, Rosenberger, R&M, Cisco, Juniper 13-15

Computer room air handlers (CRAHs)/PAHUs ▪ Vertiv, Stulz, Johnson controls, Schneider, Bluebox
Thermal/
Cooling Chillers/Heat exchangers ▪ Climaveneta, Carrier, Johnson Control, York, Vertiv, Daikin, KRN, Microcool, Aircon 30-35
equipment
Cooling towers ▪ Paharpur, Kelvion, Danfoss

Backup diesel Generators ▪ Cummins, Caterpillar, Jackson, Rolls Royce, Leroy-Somer, Sterling generators, MPU

UPS ▪ Schneider, Vertiv, Delta, Eaton, ABB

Electricals Switchgears ▪ ABB, Schneider, Siemens 35-40

PDUs and busy ways ▪ Schneider, Vertiv, Eaton, Datsons

Industrial cables ▪ Polycab, KEI, RR Kabel, Finolex

Racks ▪ APC Schneider Rack, Rittal, HP, Comm scope

Optical fiber ▪ CommScope, Panduit, Anixter, Chatsworth, Sterlite technologies, HFCL

Raised floor ▪ Unitiles, Stulz, Alpha

Misc. Dielectric fluid ▪ 3M, Shell, Lubrizol, Castrol 20-25

Gas separation unit ▪ Johnson control- Ansul (System integrators), Novac for gas

DC management infrastructure system ▪ Schneider EcoStruxture platform, Honeywell, Siemens

Security camera and Cybersecurity ▪ Siemens, JCI, Honeywell

Source: Industry, Avendus Spark

This file was generated for vishnu.agarwal@wipro.com from Avendus Spark Research website on July 17, 2025 06:30:53 Page 40
DATA CENTER INDUSTRY THEMATIC

DC Ecosystem beneficiaries: Part - I

Category Product Remarks from Recent Company Filings

Switchgear & Power ▪ Company won a significant order one-off order from a data center in Q1 CY24
ABB India Electrical Solutions
Distribution ▪ Data center demand started picking up in 16-17 but wasn't very significant; Currently substantial part of the portfolio with a high growth rate

Cables ▪ Company has supplied cables which connect power from substations to DCs; supplied to hyperscalers in India like AWS, Microsoft, Google and Adani
Apar Industries Electrical Solutions Dielectric Fluid (Pre-test ▪ Following successful execution for Microsoft India, company is now part of global vendor list & could supply to Microsoft DCs in USA
phase) ▪ Company has liquid coolant product but doesn't know how successful the product will be as they haven't been able to get any test sites

▪ Company won order for design & turnkey build execution project for EDC in NCR
AurionPro Design &
Design & Civil Works ▪ Won order for design consultancy in Mumbai (85MW) & Chennai (20MW)
Solutions Engineering
▪ Management expects to grow at 30% in the space

▪ Focused on networking, structured cabling, wireless & private LTE, typically 10–15% of DC capex.
Black Box System Integrator Connectivity Solutions ▪ Data centers contribute 15–20% of revenues, targeted to reach 25–30% in 3–4 years.
▪ Secured Rs. 2.5bn order from a major hyperscaler (3 U.S. sites).

▪ Company claims to have strong order finalization from data centers.

Blue Star Electrical Solutions Chillers ▪ Currently supplies chillers with plans to diversify product range.
▪ It is in discussion with global players to potentially enter the liquid cooling segment as adoption picks up in India

▪ Castrol Ltd (Parent) is developing coolants for immersive and direct-to-chip cooling
Castrol India Cooling Solutions Dielectric Fluid ▪ Castrol (India) is conducting proof-of-product concept pilots & assessing commercial viability of liquid cooling for Indian customers
▪ More visibility in CY25 with DC cooling market expected to grow ~20% CAGR

▪ Data centers are contributing to the order book but management finds it hard to provide any visibility on the DC contribution for 3-4 quarters in the future.
Cummins India Electrical Solutions Gensets ▪ The parent has announced a $200M CapEx across three regions, with some factories being repurposed for large data center orders, possibly shifting certain
product lines to China and India.

▪ Commissioned 220 kV GIS for a Mumbai data center (Veritas) and secured a new order for 220 kV GIS from a leading EPC player for another Mumbai DC
GE Vernova T&D project.
Electrical Solutions Gas Insulated Switchgear
India ▪ Data center orders are growing; they remain a small share of overall T&D business.
▪ Management expects larger data centers (400–765 kV) in future to drive higher-value opportunities as demand scales toward gigawatt-scale capacities.

Networking ▪ Capex of Rs. 1.38bn underway to expand capacity for next-gen data center cables.
HFCL Fiber Optic Cables
Solutions ▪ Expanding into passive connectivity solutions within data centers (Cabling, enclosures, connectors, fiber panels, and related infrastructure)

Source: Company, Avendus Spark

This file was generated for vishnu.agarwal@wipro.com from Avendus Spark Research website on July 17, 2025 06:30:53 Page 41
DATA CENTER INDUSTRY THEMATIC

DC Ecosystem beneficiaries: Part - II

Category Product Remarks from Recent Company Filings

▪ 1 in 3 Indian data centers powered by company’s integrated solutions.

Hitachi Energy India Electrical Solutions T&D ▪ Data center orders dipped YoY but pipeline remains strong; targeting Rs.20bn annual orders over 3-4 years.
▪ Signed SLA with the world’s largest data center operator; executing projects for Google (Thailand) and other hyperscalers.

KEC International EPC EPC ▪ Currently executing 5 data center projects across India, with ~100MW total capacity under development.

▪ Company provides fire pumps & cooling system pumps.

Kirloskar Brothers Cooling Solutions HVAC Pumps & Chillers ▪ Executed cooling project for Meta in Singapore
▪ Company is developing containerized plug-and-play cooling systems for DCs

Kirloskar Oil Engines Electrical Solutions Gensets ▪ DCs are the fastest-growing segment in the high horsepower engine category, driving strong domestic and international demand.

▪ Designs, manufactures, and integrates AI- and HPC-optimized servers, private cloud/HCI solutions, data center servers, and high-performance storage
Netweb System integrator (high-end
System Integrator ▪ Offers the Skylus.ai platform (private cloud & orchestration), supporting full AI lifecycles - development, training, inference - catering to CSPs, hyperscalers,
Technologies India computing)
enterprises, and government

▪ Data center segment now contributes around 2% of quarterly revenues

Pitti Engineering Electrical Solutions Laminations ▪ Electrical steel laminations, rotors, stator assemblies, and precision machined components for backup diesel generator (DG) sets used in data centers
▪ Serves as a tier-1 supplier to Cummins India, supplying components for high-reliability gensets essential for 24×7 uptime in data center power systems
▪ Supplies fire-retardant copper power and control cables, essential for 24/7 uninterrupted operations in data center
▪ These cables connect critical systems like generators, UPS units, PDUs, providing reliable and safe power delivery
Polycab Electrical Solutions Cable
▪ Offers Cat-6 and Cat-6A copper Ethernet cables, designed for high-bandwidth and gigabit networking within data halls
▪ Provides a range of fiber optic cables (6 to 576 core), suited for server-to-server and inter-building connectivity
▪ RailTel is transitioning from a telecom infrastructure provider into a hybrid data center & edge DC enabler. It combines: (I) Core and regional Tier-III DCs
Networking with managed and cloud operations (II) An expansive fiber-fed edge network for low-latency deployments
RailTel Operator
Solutions ▪ RailTel operates two Tier-III data centers in Gurugram and Secunderabad, each featuring ~100 racks
These support colocation, managed hosting, and its “RailCloud” services for government and enterprise clients.

Digitisation (Network & ▪ Schneider derives 15% of its revenue from the DC segment
Schneider DCIM
Distribution) ▪ Schneider APAC witnessed strong growth driven by strong double-digit growth in the data center & infrastructure.

▪ Aiming 25% revenue from data centers

Networking Fibre and copper cabling
Sterlite ▪ Recently announced the launch of a new generation of Data Centre solutions, ranging from cabling to end-to-end connectivity offerings designed to power
Solutions solutions
the relentless demands of AI-driven data centre infrastructure.
Source: Company, Avendus Spark
This file was generated for vishnu.agarwal@wipro.com from Avendus Spark Research website on July 17, 2025 06:30:53 Page 42
DATA CENTER INDUSTRY THEMATIC

Global Valuations- DC (Colocation and Cloud) companies trade at 20-22x EV/EBITDA on 1-year forward basis

Equinix- Largest DC company trades at 20x EV/EBITDA (1-year forward) Keppel DC is Singapore listed DC company trades at 22x EV/EBITDA (1-year forward)

Equinix Keppel DC REIT

30 30

25 25 22
20
20 20
20
15 15

10 10
Jul-15 Jul-16 Jul-17 Jul-18 Jul-19 Jul-20 Jul-21 Jul-22 Jul-23 Jul-24 Jul-15 Jul-16 Jul-17 Jul-18 Jul-19 Jul-20 Jul-21 Jul-22 Jul-23 Jul-24

1Y forward EV/EBITDA Last 10Y average 1Y forward EV/EBITDA Last 10Y average

Source: Bloomberg, Avendus Spark Source: Bloomberg, Avendus Spark

Global companies in Colocation and Cloud space- Median EV/EBITDA valuation of 15-20x
Sales in Rs.mn Sales EBITDA in Rs.mn EBITDA Net debt P/E(x) EV/EBITDA (x)
Mkt Cap EV
Company Business description CY24-26E CY24-26E to EBITDA
(US$bn) (US$bn) CY24 CY25E CY26E CY24 CY25E CY26E CY24 CY25E CY26E CY24 CY25E CY26E
CAGR% CAGR% (CY24)
Equinix Inc Colocation 77 93 8,748 9,221 9,944 7% 3,566 4,500 4,956 18% 4.3 94.5 54.1 50.9 26.1 20.7 18.8
Digital Realty Trust Data center REIT 59 76 5,592 5,883 6,545 8% 2,934 3,171 3,509 9% 5.5 124.8 114.3 99.8 25.9 24.0 21.7
Akamai Tech CDN + Edge Cloud 12 15 3,987 4,129 4,352 4% 1,662 1,714 1,823 5% 1.2 11.9 12.3 11.6 9.0 8.7 8.2
CoreWeave GPU-native cloud provider 79 90 1,915 5,050 11,530 145% 1,219 3,336 8,490 164% 6.3 NM NM NM 73.8 26.9 10.6
SMB-focused cloud
DigitalOcean Holdings 3 4 776 882 1,006 14% 314 346 389 11% 3.2 15.0 13.6 13.2 12.8 11.7 10.4
infrastructure
Nutanix Hybrid cloud infrastructure (HCI) 21 20 2,138 2,527 2,916 17% 404 595 690 31% -1.3 56.3 40.5 36.6 50.0 34.0 29.3
Fastly Edge cloud, security 1 1 542 590 630 8% 28 43 56 41% 1.5 NM NM NM 39.2 25.7 19.8
Iron Mountain Storage/Info management 30 47 6,166 6,791 7,406 10% 2,217 2,510 2,758 12% 5.9 91.4 74.1 52.0 21.3 18.8 17.1
Keppel DC REIT Global DC REIT in Singapore 4 5 229 318 332 20% 173 239 253 21% 5.2 33.2 21.9 20.6 29.8 21.5 20.3
largest private data center
Vnet Group 2 5 1,129 1,302 1,560 18% 317 397 506 26% 4.6 NM NM NM 15.9 12.7 9.9
service provider in China
High performance data center
GDS Holdings Ltd 7 12 1,575 1,602 1,835 8% 687 753 842 11% 6.3 NM NM NM 17.5 16.0 14.3
solution provider in China
Source: Bloomberg, Avendus Spark
This file was generated for vishnu.agarwal@wipro.com from Avendus Spark Research website on July 17, 2025 06:30:53 Page 43
Brief Company Profile – Colocation/Cloud Players in India in Listed Space

Currently Listed
Anant Raj - Real estate player entered in DC space in 2023, transforming existing IT park into DC center in NCR
Techno Electric & Engineering - Power Generation/Transmission EPC player, announced foray into DC space in 2020, with its first DC in Chennai in FY26E
E2E Networks - Niche AI-focused hyperscale cloud platform offering advanced Cloud GPUs and a full-stack ecosystem for AI/ML/GenAI workloads

Likely Near-term IPOs

Sify Infinit Spaces - Pioneers in Datacenter space, having launched first commercial DC in India; One of the top 5 colocation data center capacities in India
ESDS - Offers cloud computing infrastructure & managed services to small enterprises & PSUs

This file was generated for vishnu.agarwal@wipro.com from Avendus Spark Research website on July 17, 2025 06:30:53
DATA CENTER INDUSTRY THEMATIC

Anant Raj: Real estate player entered in DC space in 2023, transforming existing IT park into DC center in NCR

Entered in DC business by commissioning pilot DC centre in existing underutilised commercial

C O M PA N Y B A C KG R O U N D
building/IT center
Anant Raj Limited (ARL) was incorporated in 1985. The promoters are engaged in the business
of real estate and infrastructure development since 1969 and is amongst the oldest ▪ Anant Raj owned several IT Parks and SEZs in Haryana, particularly in
development and construction group in the National Capital Region (NCR). Manesar, Rai, and Panchkula, having higher vacancy post-COVID.
▪ It began retrofitting these IT parks to transform them into Tier III-compliant
ARL is currently developing projects in and around Delhi-NCR with major projects in Sector 63A How AnantRaj
data centers by upgrading power infrastructure, Cooling & HVAC redesign,
in Gurgaon. Additionally, the company is also expanding its existing commercial properties in entered in DC
improved for data center load-bearing standards and carrier-neutral
Delhi and is setting up its data center verticals. The company has committed $2.1 bn (Rs. 180bn) business?
connectivity.
capex for its DC vertical aiming to achieve operational capacity of ~310MW by 2031.
▪ These sites benefitted from pre-existing industrial zoning, uninterrupted

The business is run by the 4th generation of the founders. Amit Saran, Ashim Sarin & Aman Sarin power supply and proximity to NCR for Enterprise/cloud demand.
holding key positions in the board. Mr. Ashim Sarin is currently the CEO & Director of DC
division.
▪ Facility Conversion: In September 2023, Anant Raj announced its first live
data center at the Manesar tech park, repurposing existing IT infrastructure
into a Tier-III/IV-capable facility with an initial ~3 MW of IT load and plans
DC capacity to increase from 6MW currently to 28MW in 1HFY26E and 63MW by FY27E to scale to 21 MW on site.
▪ Strategic Footprint: The company plans to retrofit two additional IT parks in
Anant Raj - Expansion Plan (MW)
~ +245MW Haryana—in Rai (200 MW capacity potential) and Panchkula (50 MW)—
Timeline and with a combined 300 MW target by 2032.
Manesar Manesar:
Rai: +20MW 307 Expansion ▪ Partnership with Orange Business: In June 2024, Anant Raj announced a
+15MW
Manesar:
Panchkula: strategy for collaboration with Orange Business (French telecommunications and digital
+15MW
+7MW AnantRaj DC services provider) , aiming to co-develop sovereign cloud and managed
business data center services across its locations. Already launched “Ashok cloud”
with 0.5MW of IT load on current capacity in Manesar and plans to increase
to ~15MW in near term.
63
▪ MoU with Google: In July 2024, Anant Raj Cloud signed an MoU with
28
6 Google to provide cloud-platform–enabled data center services across its
Haryana sites (Manesar, Rai, Panchkula), aligning with their 300 MW
Current FY26E FY27E 2031E ambition

Source: Company, Avendus Spark Source: Avendus Spark

This file was generated for vishnu.agarwal@wipro.com from Avendus Spark Research website on July 17, 2025 06:30:53 Page 45
DATA CENTER INDUSTRY THEMATIC

Anant Raj: Currently, DC biz accounts for less than 5% of revenue largely with PSU clients; Target to reach 40% over 4-5 years;
Revenue target of Rs.5-6bn+ by FY28E with 50% contribution from Cloud
Colocation and Cloud business ramp target over next few years TCIL and Railtel are the two major clients for AnantRaj DC

Revenue contribution from Datacenter segment

Since 2019 Current Offerings Future Offerings
40%
- Cloud Services
- Cloud Services (IaaS with Orange)
- Colocation - Colocation
35% - Colocation
- PaaS and SaaS
Increase in
DC mix

Anant Raj Cloud has focused its efforts on government entities and public-sector clients, particularly through
strategic partnerships that position it as a sovereign-focused provider. Currently, its DCs are leased out to only PSUs
2%
Telecommunication Consultants of
RailTel Corporation of India Ltd CSC (Common Services Centers)
India Ltd (TCIL)
Current 2031

Source: Company, Avendus Spark Source: Company, Avendus Spark

AnantRaj aims to increase its cloud offering given higher margins and return metrics Ambitious revenue target in DC segment of Rs. 6-7bn over the next years
Revenue potential (Rs.bn) from DC segment over next three years
Colocation versus Cloud capacity Colocation versus Cloud revenue
contribution contribution Colocation/Cloud capacity: 6.5MW Colocation/Cloud capacity: 63MW
Revenue: Rs.460mn Revenue: Rs.6-7bn
8% EBITDA: ~250mn EBITDA: ~2-3bn
22% 15%
50%

92% 85% 3.3 Cloud

78%
50% Colocation

FY25 FY27E 3.3

FY25 FY27E
0.4
Colocation Cloud Colocation Cloud FY25 FY28E

Source: Company, Avendus Spark Source: Company, Avendus Spark

This file was generated for vishnu.agarwal@wipro.com from Avendus Spark Research website on July 17, 2025 06:30:53 Page 46
DATA CENTER INDUSTRY THEMATIC

Techno Electric & Engineering: Power Generation/Transmission EPC player, announced foray into DC space in 2020, with its first DC in Chennai in FY26E

Commissioning its Maiden Hyperscaler DC in Chennai in 2QFY26E

C O M PA N Y B A C KG R O U N D
TEECL is a leading EPC company with core capabilities in power generation, transmission (up to ▪ Techno Electric is playing to its strengths by entering DCs exactly
765 kV AIS/GIS), distribution, and industrial turnkey projects. It offers customized, service-heavy where its power/EPC legacy matters.
solutions including design and engineering. The company has executed projects in substations, ▪ In DC capex cost, 60% of cost is in MEP (Electrical and Mechanical)
distribution systems, captive power plants (up to 100 MW), and specialized industrial systems ▪ With four decades of experience in power transmission, renewable
(e.g., smelters, fuel oil, piping). TEECL has international presence in Afghanistan and Africa, was energy, and EPC, Techno was well-positioned to build energy-
Why Techno Electric is
among the early movers in 765 kV substations in India, and pioneered technologies like intensive and reliable DC facilities.
foraying in DC business:
STATCOM. In FY20, it entered electromechanical works (e.g., FGD systems) and is now setting ▪ In 2021, it launched Techno Digital Infra Pvt Ltd, capitalizing on its
up a 40 MW hyperscale data center in Chennai. existing capabilities to design and manage DC infrastructure.
▪ It has an EPC order of Rs.650mn for external clients.
TEECL plans to invest $1bn by FY29-30 in the data center segment. The target is to build & own
▪ It has formally incorporated subsidiary “Techno Digital Infra Pvt. Ltd”
250MW of capacity over the next 5 years. The total deployed capacity is expected to reach
25MW by the end of FY26 & 50MW by FY27. for DC invesment.
Ankit Saraiya, Director & CEO, Techno Electric is focusing on scaling up this business over next
five years ▪
Announced a $1 billion investment to develop 250 MW of
hyperscale and edge data center capacity across India over five
years.
Core EPC business has an order book of ~Rs.110bn- 4-5x of FY25 revenues; Also, ▪ Planning both hyperscale nodes (Chennai, Kolkata, Noida) and edge
entered in DC EPC recently
Timeline and Expansion data centers in collaboration with RailTel, designing over 102 sites in
Data Center, 1bn, 1% Generation, 11bn, strategy for Techno 23 states to support low-latency.
10% Electric DC business ▪ Its flagship Chennai (Siruseri) facility is a 36 MW, Tier-III-capable DC
featuring energy-efficient design, BESS, and adiabatic cooling,
targeting a PUE of less than 1.5.
Distribution, 23bn, ▪ Strategic partnerships (like with RailTel) ensure broad and rapid
TEECL’s Core 21%
segment deployment at scale.
order book
(Rs.)
▪ Best in class PUE of 1.35 (Design) versus industry PUE of 1.5
Transmission, 75bn, What unique about its ▪ Adiabatic cooling towers which reduces water consumption by 75%
68% Chennai DC ▪ AI ready DC and house high density racks with seamless integration,
requiring no architecture changes.

Source: Company, Avendus Spark Source: Company, Avendus Spark

This file was generated for vishnu.agarwal@wipro.com from Avendus Spark Research website on July 17, 2025 06:30:53 Page 47
DATA CENTER INDUSTRY THEMATIC

Techno Electric & Engineering: Planning both hyperscale nodes (Chennai, Kolkata, Noida) and edge data centers in collaboration
with RailTel, designing over 102 sites in 23 states to support low-latency
Colocation and Cloud business ramp target over next few years Plans on Edge data center with Railtel- Execution over next five years
Location Facility Type Planned IT Load Status Setting up of Edge Data Centers (EDC) in 102 Cities
10MW Managed Service at Gurugram
▪ Entered a contract with RailTel to build 102 EDCs across 23 ▪ The project involves setting up a
Phase 1 under development
Chennai (Siruseri) Hyperscale Campus 36 MW states in India. 10 MW data centre in Noida on RailTel-
(Tier III)
▪ The company will be targeting to set up 20 EDCs/year to owned land. It will be executed in two
conclude the project in 5 years. phases of 5 MW each over an
Kolkata Hyperscale/Edge ~30–40 MW (planned) Land identified
▪ This will be done under the Design, Build, Finance, Operate & implementation period of 1.5 to 3
Transfer Model (DBFOT). years.
Noida Hyperscale ~30–40 MW (planned) In planning stage ▪ RailTel will be offering all the inputs including land, fiber & ▪ It is a 30-year agreement based on a
power. fixed percentage revenue-sharing
LOI signed for 102 sites in model. Earnings are currently not
RailTel Sites Edge DC (Tier II/III) ~50–75 MW ▪ The project will have a concession period for 20 year with the
23 states quantifiable as project
option to extend it by 5 years by mutual agreement.
commercialization timelines are yet to
Others (Gujarat, Pune, ▪ The company commissioned an EDC in Gurugram (0.2MW) & is
Edge + Metro Balance (for 250 MW target) To be phased be finalized.
Hyderabad) in the process of commissioning an EDC in Mumbai.

Source: Company, Avendus Spark Source: Company, Avendus Spark

Commissioning first phase of Chennai DC by Q2FY26E; Aim to reach 40-50MW by FY27E Company aims to reach Rs. 3-3.5bn of revenue by FY27 in DC segment
Target DC Capacity of TEECL DC Revenue Targets of TEECL (Rs. Bn)
Completion of
Kolkata Project 3-3.5
Completion of
Chennai Project 40
Siruseri First
Phase (Chennai) 200% YoY

24

5.6

H1 FY26E FY26E FY27E

FY26 FY27

Source: Company, Avendus Spark Source: Company, Avendus Spark

This file was generated for vishnu.agarwal@wipro.com from Avendus Spark Research website on July 17, 2025 06:30:53 Page 48
DATA CENTER INDUSTRY THEMATIC

Sify Infinit Spaces: Pioneers in Datacenter space, having launched first commercial DC in India; Current capacity of ~111MW

Sify leveraged its networking solution legacy to pioneer India’s first commercial data center
C O M PA N Y B A C KG R O U N D
Sify Infinit Spaces Ltd (SISL), a wholly owned subsidiary of Sify Technologies Ltd (STL), operates 13 ▪ Sify Technologies was one of the pioneers of data center
concurrently maintainable data centers with ~111 MW built capacity -spread across Mumbai (7), Noida infrastructure in India, entering the business in the early 2000s as
(2), and one each in Chennai, Kolkata, Hyderabad, and Bengaluru. part of its broader strategy to shift from being just an internet
These facilities support two main services: co-location and managed hosting (including storage, backup, service provider (ISP) to a managed services and ICT infrastructure
monitoring, hardware/software procurement, and network configuration). Originally part of STL until player.
FY20, the data center business was later carved out into SISL. ▪ It launched the first commercial data center in India in Mumbai in
Pioneer in Data
SISL made plans of targeting Rs. 30bn capex over FY25-27 for the expansion of its data center business 2001. Initial purpose was to support internal web hosting and
Centers
predominantly via debt. The company is currently developing ~100MW of data center capacity across enterprise IT services as the internet ecosystem in India was
Noida, Mumbai & Chennai. Notably, the company is also in the planning phase for ~275MW+ across all its growing.
cities of operations by FY30E ▪ It is first in India to achieve NVIDIA DGX-ready data center
The promoter of the company is Mr. Raju Vegesna, a technocrat and holds an 84% stake in the parent certification for liquid cooling to support AI server
entity STL. ▪ Rated best in APAC from operations point of view in the internal
Over FY22-24, Kotak Special Situations Fund (KSSF) and Kotak Data Center Fund (KDCF) cumulatively ratings of leading US based hyperscaler
infused ~Rs 10bn by way of CCDs while Sify Technologies Ltd (STL; SISL’s parent) infused Rs 2bn in SISL.
▪ Healthy customer mix between hyperscalers and enterprise
customers.
Current operational capacity of ~111MW with 70% of share in Mumbai ▪ Unique position with 360 degree customer engagement by providing
entire gamut of ICT services along with Group companies
Offers host of
Chennai, 3% ▪ Provides primarily colocation and edge data center services
Kolkata, 2% integrated services
Bangalore, 4% ▪ Inter-connection business provides zero latency provision for
enterprises wanting to connect to hyperscalers at same location and
Noida, 10% also there are network players like AMS-IX (internet exchange)
present in Sify data centres which help with the inter-connect.
Current
Hyderabad, 13% Capacity Mix Mumbai, 69% ▪ Expansion of capacity from ~110 MW to 275+ MW over the next 5
(%) years
▪ Land acquisition / tie up in place for the future projected capacity
Scalable Infrastructure ▪ 120+ MW planned to be set up in vicinity of existing Rabale campus
Current Capacity: 111MW ▪ Campus style builds across key data centre locations in India –
Mumbai and Chennai which offer the customer opportunity to scale
within the same premises
Source: Company, Avendus Spark Source: Company, Avendus Spark
This file was generated for vishnu.agarwal@wipro.com from Avendus Spark Research website on July 17, 2025 06:30:53 Page 49
DATA CENTER INDUSTRY THEMATIC

Sify Infinit Spaces: Aim to treble capacity by FY30E; Hyperscalers clientele accounts for 60%+ of colocation revenues

Company plans to treble its DC capacity over next five years 65% of revenues come from long-term agreement with Hyperscalers
Sify DC Capacity (MW) Projections

275+
New towers in
Rabale, Enterprises/Rest, 35%
Bangalore &
- New towers in 200+ Noida Sify Revenue
Rabale, Noida, Contribution
Chennai
(%)
111
Hyperscale Mix, 65%

FY25 FY27E FY30E

Source: Company, Avendus Spark Source: Company, Avendus Spark

Spent Rs.9bn annually on DC capacity expansion in last three years; Expect similar run rate Financial Summary: Revenue CAGR of 20%+ over the last 2 years; RoEs compressed owing to
going forward addition of new capacities & low occupancy
Capex (Rs. Bn)
FY23 FY24 FY25
30
Revenue (Rs. Bn) 10 11 14

EBITDA (Rs. Bn) 4 5 6

PAT (Rs. Mn) 835 570 1,260

9 9
5 Net Debt to Equity (x) 1.1 0.7 0.8
2
Net Debt to EBITDA (x) 2.8 2.9 2.6
FY21 FY22 FY23 FY24 FY25-28
RoE (%) 10% 4% 6%

Source: Company, Avendus Spark Source: Company, Avendus Spark

This file was generated for vishnu.agarwal@wipro.com from Avendus Spark Research website on July 17, 2025 06:30:53 Page 50
DATA CENTER INDUSTRY THEMATIC

E2E Networks: Niche AI-focused hyperscale cloud platform offering advanced Cloud GPUs and a full-stack ecosystem for AI/ML/GenAI workloads

What sets E2E Apart?

C O M PA N Y B A C KG R O U N D
E2E Networks, founded in 2009, initially, focused on delivering fundamental computing, storage, and Strategically Positioned in a
networking solutions to Small and Medium-sized Enterprises (SMEs). ▪ Technical Expertise, Tech Mindset, Strong unit economics.
Large Market Opportunity
In 2011-2014, the company received seed funding from Blume Ventures and in 2017, the company
introduced its ‘Self-Service Cloud Platform’ which allowed its customers to manage its cloud services In-house battle tested ▪ Continuous and Rapid Product Improvement Cycles for System
without any human intervention. Proprietary Software Software and GenAI Platform
Currently, E2E Networks is a MeitY-empaneled AI-focused hyperscale cloud platform offering advanced
▪ One of the first providers in India to bring the state-of-the-art,
Cloud GPUs and a full-stack ecosystem for AI/ML/GenAI workloads. The company provides accelerated Long Collaboration with NVIDIA
energy-efficient NVIDIA H100 & H200
cloud computing solutions, via Cloud GPUs like NVIDIA A100, H100, and the newly available H200 GPUs
on the cloud. Full Stack Sovereign AI ▪ Full Stack AI from the Physical Compute Infrastructure, Storage to
E2E Networks was amongst the first few providers out of India providing contract-less computing with Infrastructure GenAI Platform Services
low latency.
Notably, Larsen & Toubro (L&T) entered as a strategic partner in E2E Networks in November 2024, by Industry Leading Performance ▪ Delivering industry-leading performance while being more cost-
acquiring a 21% stake in the company for Rs. 13bn. and Cost efficient effective compared to other solutions in the market
Tarun Dua, the promoter & MD of the company is a seasoned technologist with a strong foundation in
Linux systems, cloud infrastructure, and open-source technologies. ▪ ~100+AI/ML/Software & Cloud engineers with a track record of
Advanced Cloud Solutions Team ▪ building cloud services and Compute & Storage infrastructure
The company has more H200 chips than H100 chips which are scarce & cost 10-30% more than software
H100 chips; The company leases out space at DCs & leased capacity for 10MW
GPU-Backed AI Infrastructure ▪ ~3,700 Cloud GPUs In Delhi NCR and Chennai
DC IT Power Capacities
E2E Chip Capacity No. of Chips Clusters
(Leased)
Strategic partnership with ▪ Access to larger enterprise channel and ability to act like a full stack
Larsen & Toubro Limited player with access to additional business models
H100 ~700 256

Paths for Expansion

H200 ~2,300 1,024
▪ AILaaS is a scalable, cost-effective, and customizable solution
10MW AI Lab as a Service
designed to help educational Institutions
Legacy Chips (A30, A100,
~700 NA
A40, V100, T4, L4 & L40S)
Channel Expansion ▪ Building long term enterprise and government customers
Total Cloud GPU
~3,700 ▪ Leverage partnership with L&T and GenAI startups building
Capacity Strategic Partnerships
SOTA/Frontier models/runtimes for Govt and Enterprise
Source: Company, Avendus Spark Source: Company, Avendus Spark
This file was generated for vishnu.agarwal@wipro.com from Avendus Spark Research website on July 17, 2025 06:30:53 Page 51
DATA CENTER INDUSTRY THEMATIC

E2E Networks: Sharp increase in capex in FY25 as company aims to double the number of GPUs by FY27E

Capex grew ~5x as the company deploys Rs. 15bn that it received from its Preferential The number of GPUs that the company owns is expected to increase 55% YoY in FY26 & by
allotment FY30 it is expected to grow by 22% (CAGR)
Capex (Rs. Mn) Number of GPUs

8,700 22% 10,000

CAGR
5x
Growth 5,700
~55%
Growth
3,700
1,853

249 350

FY22 FY23 FY24 FY25 FY25 FY27 FY30

Source: Company, Avendus Spark Source: Company, Avendus Spark

Sharp increase in ARPU of top 500 customers from Rs.500k to Rs. 860k during Jan’24 to June’25 Financial Summary: EBITDA expanded in FY25 owing to one-off ARPU increase; RoEs
compressed owing to fund raise in FY25 from L&T
ARPU (Rs. '000) of Top 500 Customers
FY23 FY24 FY25
Sharp increase in ARPU from Rs.500k to Rs. 860k during Jan’24-Jun’24; 861
Normalization due to likely completion of training project of a top 5 client. 750 751 Revenue (Rs. Mn) 662 945 1,640
Notably, customer contribution from top 5 clients increased to 45-50%
during this period. 580 EBITDA (Rs. Mn) 331 479 967
515
369 406 EBITDA Margins (%) 50% 51% 59%
296 328
255 279 286
217 232 240 PAT (Rs. Mn) 99 219 475
196

Net Debt to Equity (x) -0.4 1.3 -0.8

Net Debt to EBITDA (x) -0.4 0.8 -6.5

Q1FY22

Q2FY22

Q3FY22

Q4FY22

Q1FY23

Q2FY23

Q3FY23

Q4FY23

Q1FY24

Q2FY24

Q3FY24

Q4FY24

Q1FY25

Q2FY25

Q3FY25

Q4FY25

RoE (%) 23% 36% 6%

Source: Company, Avendus Spark Source: Company, Avendus Spark

This file was generated for vishnu.agarwal@wipro.com from Avendus Spark Research website on July 17, 2025 06:30:53 Page 52
DATA CENTER INDUSTRY THEMATIC
ESDS: Offers cloud computing infrastructure & managed services to small enterprises & PSUs; Company’s colocation revenue mix has risen from 2% in 2022
to 15 in H1FY25
ESDS Product Portfolio; The company has three services – IaaS (58%), Managed Services (21%),
C O M PA N Y B A C KG R O U N D
SaaS (21%)
ESDS is an IT services company focused on cloud computing infrastructure, colocation data centres,
SaaS, and managed services. It operates four integrated data centres across Nashik, Mumbai (Airoli), What Sets ESDS Apart
Bengaluru, and Mohali.
Govt Platforms & ▪ G-SaaS empaneled with MeitY & STQC-audited; products listed on GEM;
Its business spans: (I) IaaS: Including colocation and cloud infrastructure with end-to-end cloud
Marketplace SPOCHUB marketplace for enterprise solutions.
migration solutions (IT and non-IT) (II) SaaS: Subscription-based access to proprietary and third-party
software applications (III) Managed Services: Covering data management, SAP HANA, security
R&D and ▪ Investing in AI/ML-led cloud infra (Autonomous Cloud Platform, Hiring Platform);
operations, and disaster recovery.
Differentiators differentiated by flexible pricing models and patented tech.
The company launched its first data center in Nashik in 2010, with second one being launched in
2016 in Mumbai and in 2020 the company launched its data center in Bangalore. It commissioned
two data centers in 2022 & 2023 as well. Services Details
The company currently has data center capacity of 7MW.
IaaS
Promoted by Piyush Somani, who has a bachelor’s degree in Electronic Engineering from the
University of Pune. After gradation, he worked briefly in procurement at a meter manufacturing firm ▪ Operates DCs in Nashik, Navi Mumbai, Bengaluru, and Mohali; Bengaluru &
in Mumbai before starting ESDS in 2005. Data Centers
Mohali via STPI (asset-light model) for Tier II/III scalability.

▪ eNiight Cloud (patented in US & India); supports multiple environments (public,

ESDS is largely a IaaS company; over the last 3 years it has derived ~45-50% of its revenue Cloud Offerings
private, hybrid, VPC); vertical autoscaling & pay-per-use model.
from that segment; Notably, its Colo DC revenue mix has been increasing over the years
▪ Community clouds tailored to BFSI (152 clients), Govt (e.g. NRC, Smart Meters),
ESDS - Revenue Mix (%) Community Clouds
and SAP HANA (142 clients).
SaaS Managed Services
26% 23% 21%
29%
Managed Services ▪ Covers cloud ops, DC management, cybersecurity (via SECaaS), IT/network
Managed Services
21% services, DRaaS, managed DB, DevOps & automation.
28% 27%
30%
15% IaaS - Colocation & Data Center SaaS
2% 8%
5% Services
▪ Includes eMagic, VTMscan, eNlight WAF, Web VPN, eCOS, IoT, Meet, 360°, IPeG,
43% 37% 42% 43% IaaS - Cloud & Cloud Computing In-house Products
IPAS, Low-code Magic.

FY22 FY23 FY24 H1FY25 ▪ Covers cyber-attack prevention, loan origination, healthcare, and agro-tech
Other Solutions
(Famrut).
Source: Company DRHP, Avendus Spark Source: Company DRHP, Avendus Spark
This file was generated for vishnu.agarwal@wipro.com from Avendus Spark Research website on July 17, 2025 06:30:53 Page 53
DATA CENTER INDUSTRY THEMATIC

ESDS: BFSI accounts for 30% of customers mix versus 11% in FY22; Growth led by bundling managed services with cloud offering

ARPU had been range bound over FY22 & FY23; Existing customers reached highs of 93%, Over the last 3.5 years, 1/3rd of ESDS customers have opted for all three of their services
ARPUs also reached highs of Rs. 1.96mn
Total Customer & ARPU (Rs. Mn) % of Customer
FY22 FY23 FY24 H1FY25
Availing:
85% 78% 93% 76% Existing Customers
1,492
2.5
1,465 All 3 service lines 77% 84% 62% 91%
1.96 2
Total Customers
1,427 1.5
1.37 1.39
1,398 2 service lines 10% 5% 22% 2%
1

0.5 ARPU (Rs. Mn)

0 1 service line 13% 11% 16% 7%

FY22 FY23 FY24 H1FY25

Source: Company DRHP, Avendus Spark Source: Company DRHP, Avendus Spark

BFSI mix has been steadily rising at the expense of enterprises, Government revenue Financial Summary: Company turned profitable after tax since FY24 & registered RoE of 6%;
contribution has remained steady Expected to achieve double-digit RoE in FY25
Change in Customer Mix (%) FY22 FY23 FY24 H1FY25

Revenue (Rs. Mn) 1,954 2,076 2,865 1,722

Enterprises
EBITDA (Rs. Mn) 587 475 1,019 739
48% 39%
60% 54%
EBITDA Margin (%) 30% 23% 36% 43%
Government Customers
29% PAT (Rs. Mn) -27 -225 136 239
34%
30% 34% Net Debt to Equity (x) 0.4 0.7 0.7 0.7
32% BFSI
11% 12% 18%
Net Debt to EBITDA (x) 1.4 3.1 1.4 2.2
FY22 FY23 FY24 H1FY25
RoE (%) -1% -11% 6% 10%
Source: Company DRHP, Avendus Spark Source: Company DRHP, Avendus Spark
This file was generated for vishnu.agarwal@wipro.com from Avendus Spark Research website on July 17, 2025 06:30:53 Page 54
Appendix

This file was generated for vishnu.agarwal@wipro.com from Avendus Spark Research website on July 17, 2025 06:30:53
DATA CENTER INDUSTRY THEMATIC

Key Developments & Milestones in Data Center evolution in India

2005 2006 2007 2008 2009 2010 2011

▪ Data centers largely ▪ Initial signs of ▪ Rise in managed hosting ▪ Financial crisis prompts ▪ Government begins ▪ Surge in e-commerce and ▪ SEBI begins enforcing DR
captive (IT services, enterprise hosting as enterprises seek scale surge in BCP/DR demand investing in NIC data digital services (Flipkart, site norms for brokers
telcos, BFSI) demand ▪ Netmagic expands in ▪ Enterprises look to shift centers IRCTC traffic spikes) ▪ BFSI starts outsourcing
▪ Tata Communications, ▪ High cost and latency Mumbai from captive to managed ▪ UIDAI project kicks off → ▪ CtrlS expands to Mumbai, infra to colocation
Sify, and Reliance lead hinder widespread ▪ Early discussions around models sets stage for future DC begins positioning as Tier players
early commercial DC adoption data localization in ▪ CtrlS launches its first needs IV provider ▪ Netmagic scales facilities
offerings banking begin Tier-IV facility in ▪ Global cloud players (AWS, ▪ Sify launches large-scale DC in Bengaluru and NCR
Hyderabad Microsoft) expand in Chennai
availability via Singapore

2018 2017 2016 2015 2014 2013 2012

▪ MeitY releases Draft ▪ AWS Mumbai region ▪ STT GDC enters India ▪ Microsoft Azure launches India cloud ▪ Growing OTT, fintech (Paytm), ▪ Digital India vision gains ▪ NTT acquires 74% in
Data Centre Policy goes live via Tata regions (Pune & Chennai) — first and app ecosystem drives infra political momentum Netmagic, marking the
▪ BFSI, OTT, and startups ▪ Nxtra and STT Communications DC hyperscaler with local zones demand ▪ Telcos (Bharti Airtel, first major global DC
drive demand expand capacities acquisition ▪ AWS and Google yet to launch local DCs ▪ Microsoft announces plan for Reliance) expand infra M&A in India
▪ CtrlS announces Noida ▪ India DC capacity ▪ AWS announces plan (followed in 2016–2017) India cloud region ▪ AWS ramps up presence via ▪ CtrlS becomes Asia’s
expansion crosses ~400 MW to launch local India ▪ Enterprises shift from managed hosting to ▪ CtrlS, STT GDC, Sify begin Singapore for Indian clients largest Tier IV certified
region IaaS/cloud-first IT models planning major DC parks provider
▪ IBM SoftLayer enters India
▪ Market estimated at ~250–300 MW ▪ Early signs of structured data market (hosted abroad)
capacity localization policies emerge

2019 2020 2021 2022 2023 2024 2025

▪ Yotta, AdaniConneX ▪ COVID-19 accelerates ▪ AI/ML and edge computing ▪ Draft DPDP Bill introduced ▪ Microsoft announces a‚ Rs. ▪ India DC capacity ▪ ~75% of new builds
enter the market digital transformation use cases emerge ▪ Hyperscaler leasing 4x 150bn for 3 new DCs crosses ~900 MW are AI-ready
▪ CtrlS breaks ground on ▪ Hyperscalers lease ▪ Microsoft, AWS, Google, higher than in 2017 ▪ CtrlS, Yotta build out Tier ▪ Edge DC buildouts ▪ Chip supply and
new hyperscale record capacity Oracle expand local zones ▪ India DC capacity IV hyperscale campuses begin in earnest land/power access
campuses ▪ Nxtra, STT, NTT see ▪ Government pushes PSU approaches ~750 MW ▪ AI/GenAI workloads ▪ India crosses 1 GW emerge as constraints
▪ Colocation demand robust growth digitization ▪ DPDP Act passed dominate new demand operational DC capacity
surges
Source: Industry, Avendus Spark
This file was generated for vishnu.agarwal@wipro.com from Avendus Spark Research website on July 17, 2025 06:30:53 Page 56
DATA CENTER INDUSTRY THEMATIC

Evolution of the Data Center Industry: From CPU-Based Computing to Decentralized AI Infrastructure

Key Developments Implications

▪ Early computers (e.g., IBM 704, UNIVAC) required large, specialized rooms- 1950s–1960s ▪ Birth of dedicated computing spaces; early seeds of the data center model
Environments had dedicated power, cooling, and security Mainframe Era

▪ Rise of minicomputers allowed smaller organizations to deploy IT- Initial server 1970s–1980s ▪ Democratization of computing; emergence of in-house IT infrastructure
rooms emerged but lacked scalability and standardization Minicomputer Shift

▪ Local area networks (LANs) and client-server computing spread 1990s ▪ Start of enterprise-grade IT environments; limited reliability/security drove need
▪ Businesses built server closets/server rooms onsite Client-Server Era for better design

▪ Telecom boom led to rise of colocation facilities

▪ Uptime Institute introduced Tier standards (Tier I–IV) Late 1990s–2000s ▪ Commercialization of data centers; emergence of neutral-host, professionally
Colocation & Tiering managed facilities
▪ Focus on redundancy and uptime

▪ Amazon launched AWS (2006), followed by Azure

▪ Shift from capex-heavy IT to opex-based cloud models- 2006 onwards ▪ Explosion in compute availability and global cloud infrastructure
▪ Rapid hyperscaler expansion Cloud Computing Era

▪ Rise of virtualization, containers (e.g., Docker, Kubernetes)

▪ Software-defined networking and storage 2010s ▪ Enhanced scalability and flexibility; IT abstracted from physical hardware
▪ Hybrid & multi-cloud models Virtualization & Software-Defined

▪ Emergence of AI/ML workloads demanding 30-100kW+/rack 2020s ▪ Redefined infrastructure needs; traditional DCs evolve to support HPC & real-time
▪ Liquid cooling, GPU racks, edge deployments gain traction AI & High-Density Compute AI

▪ Growing focus on energy efficiency (target PUE <1.3)

▪ Renewable energy integration, on-site solar, modular designs 2020s–Future ▪ ESG becomes a key differentiator; innovation in energy and cooling crucial for
Green & Modular Future next-gen facilities
▪ Nuclear microgrids (emerging)

Source: Industry, Avendus Spark

This file was generated for vishnu.agarwal@wipro.com from Avendus Spark Research website on July 17, 2025 06:30:53 Page 57
DATA CENTER INDUSTRY THEMATIC

Comparison between various data center business models; Most corporates in India opt for a hybrid between the On-Premise & Colocation model

Pros Cons

▪ Maximum customization and control; data Captive Data Centers (On-Premise) ▪ High upfront CapEx; inflexible scalability;
sovereignty; low-latency compute; aligns Fully owned and operated IT infrastructure hosted within an organization’s premises, long deployment cycles; internal talent
with IP protection needs. offering maximum control and customization. dependency.

▪ Hardware control without infra CapEx; Colocation ▪ Physical access constraints (mitigated by
scalable; high-tier security, connectivity; Renting physical space in a third-party data center to house self-owned IT equipment, Smart Hands); hybrid management
cost-efficient shared infra. combining infrastructure reliability with hardware control. complexity.

▪ Scalability; pay-as-you-go; fast deployment; Cloud Computing ▪ Vendor lock-in risks (esp. SaaS, PaaS); shared
global reach; access to advanced On-demand delivery of computing resources over the internet, shifting infrastructure responsibility security model; cost overruns
technologies. management to third-party providers under scalable service models. without monitoring.

▪ Proactive monitoring; predictable OpEx; Managed Services

▪ Lower customization; dependence on
expert-led management; reduced internal IT Outsourcing of IT operations and management to a specialized provider for proactive provider capabilities; SLA risks.
burden. support, predictable costs, and reduced internal burden.

Source: Industry, Avendus Spark

This file was generated for vishnu.agarwal@wipro.com from Avendus Spark Research website on July 17, 2025 06:30:53 Page 58
DATA CENTER INDUSTRY THEMATIC

Types of DCs on the basis of redundancies: Tier 1 to 4 DCs; Hyperscalers prefer DCs with Tier-III/IV rating; India has 5+ certified Tier IV data centers

Parameter Tier I Tier II Tier III Tier IV

Minimum Capacity Component

N (non-redundant) N+1 (redundant) N+1 (redundant) 2N (fully redundant) or 2N+1
(Redundancy)

Multiple, independent, usually only Multiple, independent, diverse, all are

Distribution Paths Single, non-redundant Single, non-redundant
one is active active

Minimum Availability 99.671% 99.750% (or 99.741%) 99.982% 99.995%

Maximum Annual Downtime <28.8 hours <22 hours <1.6 hours <26.3 minutes

Dual-Powered IT Systems    

Concurrent Maintainability    

Fault Tolerant Power Design    

Compartmentalisation    

Continuous Cooling    
Source: CoreSite, Uptime Institute, Avendus Spark
This file was generated for vishnu.agarwal@wipro.com from Avendus Spark Research website on July 17, 2025 06:30:53 Page 59
DATA CENTER INDUSTRY THEMATIC

PUE (Power Usage Effectiveness) is the industry-standard metric that measures how efficiently a data center uses energy; Global
PUE has been rangebound at 1.5-1.6 over the last 5 years
Average Annual PUE
▪ Power Usage Effectiveness (PUE) is the industry-standard metric for measuring data center energy
efficiency. It is calculated as the ratio of total facility energy to the energy consumed by IT equipment. 2.5

▪ A PUE of 1.0 represents perfect efficiency, where all energy is used solely for computing. In real-world
scenarios, leading hyperscalers operate at PUEs of 1.08-1.2, while global averages hover around 1.55-
1.58.
1.98
▪ Lower PUE not only reduces operating costs and environmental footprint but also unlocks higher
usable IT capacity per megawatt, directly impacting revenue per rack or square foot.
▪ Importantly, PUE is now central to regulatory benchmarks and ESG frameworks - especially in 1.65 1.67
1.58 1.59 1.57 1.55 1.58 1.56
emerging markets like India, where policies are shaping future infrastructure standards. For investors,
operators with lower PUEs are better positioned on cost, compliance, and scalability
CY07 CY11 CY14 CY18 CY19 CY20 CY21 CY22 CY23 CY24

Period PUE Drop Cause

▪ Airflow management: Hot/cold aisle containment, blanking panels, and perforated floor tiles
▪ Basic cooling upgrades: Higher chilled water temperatures, economizers, airflow tuning
2007-2013 High
▪ Better facility design: More compact server layouts and improved white space planning
▪ Monitoring & metering: Introduction of DCIM tools to visualize energy waste

Constraining Factors-
▪ Legacy infrastructure drag: Most global data centers still run on aging architecture, which is hard to retrofit
▪ Capex limitations: Upgrading old facilities offers lower ROI vs. building new ones
2013-2020 Modest ▪ Thermal inertia: Older sites are not optimized for modern cooling or power density needs
Offsetting Factors-
▪ Hyperscale shift begins - large operators like AWS, Google, and Meta begin consolidating workloads into newer, highly efficient facilities, improving capacity-weighted PUE even as the site-level
average plateaus

▪ Liquid cooling (LIC/DLC/immersion): Enables sub-1.2 PUE by dramatically improving thermal transfer
▪ AI/ML-driven energy management: Dynamic airflow and cooling optimization
2020-2024 Stagnant ▪ Server consolidation & virtualization: More IT output per watt
▪ Smart PDUs & VFDs: Fine-tuned energy use across cooling/power systems
▪ However: These are not yet mainstream-adoption is limited to hyperscalers and top-tier colocation builds
Source: Industry, Uptime Institute, Avendus Spark
This file was generated for vishnu.agarwal@wipro.com from Avendus Spark Research website on July 17, 2025 06:30:53 Page 60
DATA CENTER INDUSTRY THEMATIC
Cooling systems are the largest non-IT energy consumers in data centers, often accounting for 30–50% of total energy use; As power densities and AI workloads increase, cooling is
emerging not only as a cost optimization lever, but also a critical enabler of infrastructure scalability and PUE (Power Usage Effectiveness) reduction.
Different forms of cooling methods Comparison of Cooling Methods
Deployment Liquid Cooling
Cooling Method Type Efficiency Metric Air Cooling Chilled Water
Suitability (DLC/Immersion)

CapEx Low Medium–High High

Air Cooling Legacy facilities, Tier
Air-based Standard (baseline)
(CRAC/CRAH) II/III DCs
OpEx High Medium Low (post-deployment)

Energy Efficiency Low Moderate High

Colocation &
Chilled Water Systems Air-liquid hybrid Moderate–High enterprise-grade Cooling Density
facilities < 10 kW/rack 10–25 kW/rack 30–100+ kW/rack
Capability

High (can reduce PUE

Impact on PUE Moderate Moderate
<1.2)
Direct-to-Chip Liquid AI/ML clusters, high-
Liquid Very High
Cooling (DLC) density zones

▪ Water consumption at a data center can be measured using WUE. WUE can be calculated by dividing
Annual Water Usage (liters) by Annual IT Energy Usage (kWh). Hyperscalers achieve values below 0.4
L/kWh whilst the typical data center operates in the range of 1.0-2.5L/kWh.
GPU clusters, HPC, AI
▪ Data centers use large amounts of water for their cooling systems, which include cooling towers, chillers,
hyperscale,
Immersion Cooling Liquid Ultra High pumps, pipes, heat exchangers, condensers, and computer room air handler (CRAH) units. Additionally,
edge/space
data centers need water for their humidification systems and facility maintenance.
constrained
▪ Currently 80% of data centers around the world use traditional air cooling as the average rack density is
at 8-10kW per rack. As share of AI data centers increase & rack densities increase expect a decrease in
traditional air cooling.
▪ The expectation is that when there is widespread adoption of liquid cooling, it will done via close loops.
High (climate Colder geographies,
Free Cooling Air or water For instance, OpenAI’s project ‘Stargate’ would require ~4mn liters of water everyday for cooling but
dependent) cost-sensitive builds
given that is a closed loop it will only require ~4mn liters of water once.

Source: Industry, Avendus Spark

This file was generated for vishnu.agarwal@wipro.com from Avendus Spark Research website on July 17, 2025 06:30:53 Page 61
Exchange and Currency Performance AVENDUS SPARK RESEARCH
08 July 2025

Equity Performance (%) Chg. Chg. Currency Performance (%) Chg. Chg.
Global Indices from from Currency from from
Today 1m 3m 6m 12m YTD Today 1m 3m 6m 12m
52WH 52WL 52WH 52WL
Developed Developed
US (Dow Jones) 44,373 3.8 16.9 4.3 12.7 4.3 -1.6 21.2 Dollar Index# 97.3 -1.9 -5.8 -10.3 -7.2 -11.7 1.0
UK (FTSE100) 8,807 -0.4 14.3 6.8 7.3 7.8 -1.1 16.7 Pound 1.4 0.5 7.0 9.1 6.3 -1.2 12.5
Japan (Nikkiei 225) 39,588 4.9 19.9 -1.0 -2.9 -0.8 -6.7 28.6 Yen 146.0 -1.0 1.3 8.3 10.2 -9.8 4.6
Germany (DAX) 24,074 -0.9 21.6 18.4 30.3 20.9 -1.7 41.4 Euro 1.2 2.6 7.4 13.3 8.3 -0.9 15.1
BRICS BRICS
Brazil (IBOV) 139,840 2.7 11.3 15.4 10.7 16.3 -1.2 18.3 Real 5.5 1.6 8.1 11.5 0.0 -13.4 1.8
Russia (IMOEX) #N/A N/A #N/A N/A #N/A N/A #N/A N/A #N/A N/A #N/A N/A #NAME? #NAME? Ruble 78.6 0.8 9.9 36.7 11.6 -31.7 2.2
India (Sensex) 83,443 1.5 12.4 6.8 4.4 6.8 -2.9 16.8 Rupee 85.9 -0.3 0.0 -0.2 -2.8 -2.4 2.9
China (SHCOMP) 3,473 2.6 10.4 7.5 18.8 3.6 -5.5 29.1 Renminbi 7.2 0.1 2.1 2.1 1.3 -2.4 2.4
South Africa (Jalsh) 97,363 1.0 18.5 15.7 20.5 15.8 0.1 26.2 Rand 17.8 -0.2 10.7 5.2 2.0 -10.9 4.3
Asian Asian
HK (H S I) 23,888 0.4 18.7 23.9 36.3 19.1 -4.0 45.3 HK Dollar 7.8 0.0 -1.0 -0.9 -0.5 0.0 1.3
Korea (Kospi) 3,059 8.8 31.1 21.4 7.1 27.5 -2.4 33.9 Won 1,376.2 -1.6 6.9 5.6 0.5 -7.5 5.6
Singapore (Straits) 4,032 2.5 16.2 3.7 18.4 6.4 0.0 26.1 SG Dollar 1.3 0.6 5.8 6.7 5.5 -7.0 0.7
Malaysia (KLCI) 1,538 1.4 6.5 -4.8 -4.6 -6.4 -8.7 10.9 Ringgit 4.2 -0.1 5.8 5.9 11.2 -10.2 3.4
Indonesia (Jakarta) 6,901 -3.0 15.1 -2.5 -4.8 -2.5 -12.8 17.3 Ind Rupiah 16,240.0 0.3 3.6 -0.6 0.1 -5.7 7.8
Commodities Performance (%) Commodities Performance (%)
Brent ($/bbl) 69.3 4.2 7.9 -10.1 -20.0 -7.2 -20.3 18.6 Indonesian Coal ($/MT) 107.4 8.9 -10.7 -13.4 -17.7 -18.2 8.9
WTI ($/bbl) 67.7 6.4 13.0 -5.1 -10.5 -2.6 -13.7 25.1 S Africa Coal ($/MT) 94.6 5.4 6.9 -10.3 -12.2 NA NA
Copper ($/MT) 9,960 3.0 14.3 13.6 2.3 15.1 -2.4 17.9 Australia Coal ($/MT) 110.0 0.9 11.4 -8.1 -18.8 NA NA
Zinc ($/MT) 2,702 1.1 2.2 -5.5 -7.7 -8.5 -18.9 7.4 Gold Spot $/Oz 3,323 -0.1 11.4 25.5 40.9 -5.0 41.4
Aluminium ($/MT) 2,587 4.5 10.3 4.9 4.6 2.4 -5.5 19.9 GOLD INDEX (Rs./10g) 96,317.0 -0.3 9.1 25.4 33.1 -2.6 41.9
Iron Ore ($/MT) 91 -1.4 -6.5 -3.4 -14.6 -4.9 -12.0 6.7 Silver Spot $/Oz 36.7 -0.2 21.9 22.1 19.2 -1.7 38.7
Lead ($/MT) 2,034 3.3 8.5 7.4 -6.7 5.7 -8.4 11.8 MCX Silver (Rs./KG) 107,068.0 2.0 19.1 19.8 18.2 -1.9 36.5

This file was generated for vishnu.agarwal@wipro.com from Avendus Spark Research website on July 17, 2025 06:30:53 Page 62
 Exchange and Currency Performance AVENDUS SPARK RESEARCH
08 July 2025

Particulars Today 1D 1M 3M 6M 12M

NSE / BSE Category wise turnover FII DII Others
Liquidity (Rs. Bn) (Rs. Bn)
Revers e Repo
Repo 5.50 6.00 6.00 6.25 6.50 6.50
MSF 13 12 20 46 159 47
523 543
Net l i qui di ty* (3,924) (3,926) (3,166) (1,829) 1,161 (1,300) 592 552
Bond Market Change in BPS 576
439
Indi a 10 yr 6.29 (0) 0 (19) (46) (70) 304 295
251 240
US 10 yr 4.38 3 (13) 19 (31) 10 216 204
251 295 248 309
Sprea d (Indi a 10Y-US10Y) 192 158 176
Indi a 10YR AAA corp 7.14 (1) 22 3 (11) (34) 1-Jul 2-Jul 3-Jul 4-Jul 7-Jul 1M avg
Sprea d (Indi a 10Y-AAA10Y) 85
91D T.Bi l l s 5.37 0 (25) (93) (123) (143)
1Y T.Bi l l s #N/A N/A #VALUE! #VALUE! #VALUE! #VALUE! #VALUE! Nifty Top 10 Deliveries
Ca l l ra te 5.26 (3) (16) (90) (147) (129) Rank Company Delivery % 30D Del. % Price Chg
LIBOR, MIFOR Change in BPS
1 HOUSING DEVELOPMENT FINANCE 83.3 71.1 1.5%
LIBOR 4.96 11 28 (108) #VALUE! #VALUE! 2 DR. REDDY'S LABORATORIES 73.9 58.0 1.0%
MIFOR 6.56 (39) (67) (138) 92 30 3 BHARTI AIRTEL LTD 73.8 64.4 0.8%
OIS 5.37 (4) (9) (14) (11) (31) 4 MARUTI SUZUKI INDIA LTD 71.9 61.4 2.6%
12m OIS fwd 5.51 1 3 (38) (96) (125) 5 MAHINDRA & MAHINDRA LTD 70.7 62.1 4.2%
* Positive number represents liquidity deficiency/Negative number represents surplus – in the system 6 HINDUSTAN UNILEVER LTD 70.5 63.3 0.6%
7 INFOSYS LTD 69.7 65.6 3.3%
8 BAJAJ AUTO LTD 69.2 57.8 -0.4%
9 SUN PHARMACEUTICAL INDUS 68.5 62.8 2.9%
10 HCL TECHNOLOGIES LTD 66.5 61.5 1.5%

Market Activity FII & DII - Provisional (INR Bn) Bulk Deals (INR Mn)
Last 5 Day FII Buy FII Sell Net DII Buy DII Sell Net Date Script Client Name Type Qty Price
7-Jul -25 89.6 86.4 3.2 111.3 92.8 18.5 7-Jul -25 INFIBEAM MAYUR MUKUNDBHAI DESAI BUY 1,00,00,000 15.2
4-Jul -25 75.2 82.8 -7.6 103.1 113.3 -10.3 7-Jul -25 INFIBEAM SONAL DESAI SELL 1,00,00,000 15.2
3-Jul -25 116.7 131.5 -14.8 126.9 113.6 13.3 7-Jul -25 INFIBE-RE CRAFT EMERGING MARKET FUND PCC- CITADEL CAPITAL
BUY FUND 68,30,000 4.3
2-Jul -25 139.5 155.2 -15.6 167.0 136.6 30.4 7-Jul -25 INFIBE-RE CRAFT EMERGING MARKET FUND PCC- ELITE CAPITALBUY
FUND 62,00,000 4.3
1-Jul -25 115.6 135.3 -19.7 129.2 121.5 7.7 7-Jul -25 INFIBE-RE VANGUARD EMERGING MARKETS STOCK INDEX FUND SELLA SERIES 37,31,001
OF VIEIF 4.3

This file was generated for vishnu.agarwal@wipro.com from Avendus Spark Research website on July 17, 2025 06:30:53 Page 63
 Avendus Spark Focus Stocks AVENDUS SPARK RESEARCH
08 July 2025

Delivery Delivery
Returns (%) Returns (%)
Volume % inc/dec Volume % inc/dec
Company Price 1D 1M 3M 1Y ('000) to 30D avg Delivery % Rating Company Price 1D 1M 3M 1Y ('000) to 30D avg Delivery % Rating
AgroChemicals Cement, Building Material & Real Estate

PI INDUSTRIES LTD 4,162 (1.8) 6.8 27.0 10.8 57.5 (1.0) 60.7 SELL DALMIA BHARAT LTD 2,154 (2.5) 1.8 19.1 16.1 64.4 (0.6) 56.1 REDUCE

COROMANDEL INTERNATIONAL LTD 2,268 1.5 (1.3) 9.9 40.3 128.6 (1.0) 45.0 SELL ASTRAL LTD 1,489 (0.3) (2.5) 17.3 (36.1) 346.9 (0.9) 62.1 BUY

Automobiles & Logistics CENTURY PLYBOARDS INDIA LTD 737 0.1 (6.7) 8.3 0.7 18.9 (1.0) 50.7 BUY

AMARA RAJA ENERGY & MOBILITY 964 0.6 (3.6) (0.3) (42.8) 496.0 (0.8) 70.8 ADD KAJARIA CERAMICS LTD 1,167 2.3 15.3 48.1 (18.9) 161.1 (1.0) 55.2 ADD

ASHOK LEYLAND LTD 250 (0.3) 3.2 23.8 10.5 1441.5 (0.6) 59.7 ADD RAMCO CEMENTS LTD/THE 1,086 0.9 8.0 16.2 36.6 131.5 (0.3) 54.5 SELL

MAHINDRA LOGISTICS LTD 355 2.4 8.1 31.3 (31.8) 120.9 (1.0) 56.8 SELL SOBHA LTD 1,516 0.2 (9.7) 31.1 (25.1) 74.0 (1.0) 34.7 BUY

UNO MINDA LTD 1,094 (0.5) 0.4 30.6 (2.7) 125.6 0.0 52.1 BUY Financials

TIMKEN INDIA LTD 3,353 0.7 1.3 45.5 (22.3) 64.2 (1.0) 66.6 ADD AU SMALL FINANCE BANK LTD 814 0.2 8.6 49.2 26.6 1127.2 28.6 63.3 ADD

SUNDRAM FASTENERS LTD 1,047 (0.8) 4.5 23.4 (25.6) 32.9 (0.6) 41.2 ADD SBI LIFE INSURANCE CO LTD 1,808 (0.1) 1.6 21.5 19.4 244.9 (0.5) 64.1 ADD

ZF COMMERCIAL VEHICLE CONTRO 13,137 (1.6) (5.2) 3.2 (16.5) 7.2 (1.0) 51.7 ADD CHOLAMANDALAM INVESTMENT AND 1,514 (0.8) (5.0) 3.5 7.2 1994.0 1.3 77.5 BUY

Consumption & Media CITY UNION BANK LTD 219 0.9 8.8 35.2 33.4 944.1 (0.7) 55.0 BUY

ADITYA BIRLA FASHION AND RET 78 2.0 0.4 (13.9) (33.5) 2245.1 (0.6) 39.0 – CENTRAL DEPOSITORY SERVICES 1,779 (1.5) 0.1 51.8 54.2 717.0 (0.9) 26.5 ADD

JUBILANT FOODWORKS LTD 684 (3.2) (1.6) 0.3 19.2 1259.2 3.4 37.6 BUY FEDERAL BANK LTD 215 (0.9) 3.5 12.9 14.4 3225.8 (0.5) 60.0 ADD

BERGER PAINTS INDIA LTD 586 (2.4) 0.8 9.4 15.0 191.9 (0.6) 56.8 SELL KARUR VYSYA BANK LTD 270 1.9 13.3 27.9 34.4 1834.0 (0.1) 66.2 BUY

VMART RETAIL LTD 793 (0.6) (6.7) 1.2 (3.5) 36.1 (0.0) 55.1 BUY ICICI LOMBARD GENERAL INSURA 2,026 (0.4) 1.0 14.0 10.1 541.6 1.9 65.7 BUY

TATA CONSUMER PRODUCTS LTD 1,102 1.1 (1.2) 3.0 (3.1) 612.1 0.5 52.0 ADD SHRIRAM FINANCE LTD 671 (0.6) (2.5) 4.7 19.5 934.5 (0.9) 52.0 BUY

ZYDUS WELLNESS LTD 1,986 0.3 2.0 14.3 (3.8) 16.3 (0.9) 42.5 ADD SUNDARAM FINANCE LTD 5,167 0.5 1.9 10.2 11.4 10.2 (1.0) 38.0 ADD

LA OPALA RG LTD 257 (1.4) 4.2 18.8 (22.7) 23.2 (0.7) 42.7 BUY IT Services

PVR INOX LTD 974 (0.4) (4.5) 10.3 (33.2) 155.4 (1.0) 52.5 BUY CYIENT LTD 1,297 (0.1) (2.3) 12.4 (27.7) 137.0 (0.3) 53.2 ADD

TRENT LTD 5,499 (11.2) (4.8) 15.4 (1.7) 436.1 (0.9) 28.9 ADD TEAMLEASE SERVICES LTD 1,990 (1.8) 3.5 14.5 (32.6) 24.5 (1.0) 74.1 ADD

V.I.P. INDUSTRIES LTD 415 0.4 21.0 56.5 (12.9) 121.2 (1.0) 46.0 REDUCE INTELLECT DESIGN ARENA LTD 1,149 0.9 (1.2) 81.3 6.2 130.7 (0.3) 48.0 REDUCE

Capital Goods & Infra KPIT TECHNOLOGIES LTD 1,269 0.6 (3.8) 14.0 (24.6) 437.6 (0.4) 56.8 SELL

AIA ENGINEERING LTD 3,353 1.3 (4.3) 8.2 (20.3) 84.6 (1.0) 89.7 REDUCE Pharma

DIXON TECHNOLOGIES INDIA LTD 15,433 2.4 3.9 18.8 23.6 224.0 (1.0) 34.5 REDUCE APOLLO HOSPITALS ENTERPRISE 7,616 0.7 9.7 12.7 20.6 333.2 0.5 72.9 ADD

TUBE INVESTMENTS OF INDIA LT 2,940 (0.4) (4.1) 14.4 (32.1) 39.4 (1.0) 39.7 ADD BIOCON LTD 372 (0.6) 12.7 15.5 1.5 1860.0 (0.7) 53.6 ADD

BLUE STAR LTD 1,827 (0.8) 14.9 (7.6) 10.3 382.4 1.8 53.9 ADD DR LAL PATHLABS LTD 2,855 3.6 0.3 12.6 (1.8) 99.3 (1.0) 60.9 REDUCE

CARBORUNDUM UNIVERSAL LTD 984 (0.8) 3.6 3.2 (42.0) 103.5 (0.7) 60.0 SELL MAX HEALTHCARE INSTITUTE LTD 1,300 (0.1) 11.2 21.1 40.3 427.0 (0.9) 65.3 REDUCE

ELGI EQUIPMENTS LTD 532 (1.0) 0.5 27.3 (28.2) 75.6 (1.0) 39.7 SELL NARAYANA HRUDAYALAYA LTD 1,990 (0.8) 13.7 18.4 60.3 99.3 (1.0) 44.5 ADD

KNR CONSTRUCTIONS LTD 221 (1.0) 4.8 0.7 (38.2) 321.4 (1.0) 40.7 REDUCE DR. REDDY'S LABORATORIES 1,311 1.5 (0.8) 18.3 0.3 304.9 (1.0) 52.2 BUY

AMBER ENTERPRISES INDIA LTD 7,431 1.2 16.2 15.1 65.2 148.5 (1.0) 37.1 SELL Oil & Gas

V-GUARD INDUSTRIES LTD 397 0.7 5.8 10.2 (14.9) 69.4 (0.5) 43.9 SELL GUJARAT GAS LTD 485 (2.1) 1.0 21.4 (25.8) 91.4 (1.0) 43.0 SELL

VOLTAS LTD 1,366 (0.1) 7.4 3.4 (6.4) 453.4 (0.9) 50.4 REDUCE PETRONET LNG LTD 308 2.2 0.5 7.8 (8.8) 4785.3 4.0 51.6 SELL

This file was generated for vishnu.agarwal@wipro.com from Avendus Spark Research website on July 17, 2025 06:30:53 Page 64
 Avendus Spark Disclaimer

BUY Stock expected to provide positive returns of >15% over a 1-year horizon REDUCE Stock expected to provide returns of <5% – -10% over a 1-year horizon
Absolute Rating
Interpretation
ADD Stock expected to provide positive returns of >5% – <15% over a 1-year horizon SELL Stock expected to fall >10% over a 1-year horizon

DISCLAIMER:
Spark Institutional Equities Private Limited (“SIE”) holds a SEBI Research Analyst license (Registration No.INH200009722) and is primarily engaged in the business of institutional equities broking and research. SIE is registered with SEBI as a Stockbroker (member of BSE Limited and the National
Stock Exchange of India Limited) and a Research Analyst.
SIE’s affiliates include (1) Avendus Capital Private Limited, holding Company of SIE, is a SEBI registered Merchant Banker (2) Avendus Wealth Management Private Limited is a SEBI registered Portfolio Manager & Investment Advisor, an investment manager to a Category III Alternative Investment
Fund ("AIF”), AMFI registered Mutual Fund distributor and registered as Authorised Person and registered with United States Securities and Exchange Commission (US SEC) as a Registered Investment Advisor (3) Avendus PE Investment Advisors Private Limited is an investment manager to a
SEBI Registered Category II AIF. (4) Avizo Advisors Private Limited is an investment manager to SEBI Registered Category II AIF (5) Avendus Capital Public Markets Alternate Strategies LLP is a SEBI registered Portfolio Manager and an investment manager to SEBI registered Category III AIF (6)
Avendus Investment Managers Private Limited is a SEBI registered Portfolio Manager and an investment manager to SEBI registered Category III AIF (7) Avendus Finance Private Limited which is registered with RBI as Non-Banking Finance Company and registered with SEBI as Depository
Participant (with CDSL and NSDL) (8) Avendus Capital Inc which is Broker dealer registered with SEC-Financial Industry Regulatory Authority (FINRA)-USA and Registered Investment Advisor with State Securities Regulator of New York, New Jersey, Texas and California (9) Avendus Capital Pte
Limited, Singapore, registered with Monetary Authority of Singapore ("MAS") and has Capital Markets Services (CMS) Licence in Fund Management activity.
We offer our research services to clients as well as our prospects.
The information and opinion expressed in this Report does not constitute an offer or an invitation to make an offer to buy or sell any securities. None of the Analysts of SIE accepts any liability or responsibility whatsoever for the accuracy or completeness of any such information. All estimates,
expressions of opinion, and other subjective judgments contained herein are made as of the date of this Research Report. This Research Report does not claim to contain all information that an investor / potential investor may require for making an investment. The past performance of a
product or portfolio does not in any manner indicate the surety of the performance in the future. This Report also does not constitute a personal recommendation or take into account the particular investment objectives, financial situations or needs of individual clients. SIE also has a separate
team that puts together Technical Analysis / Alternative Research and each recipient of this report must note that the views of the Technical Analyst is always based on short term market variables and will be materially different from the views of the other Sector Analysts in SIE, whose reports
are based on fundamental analysis of the Company.
Each recipient of this document should make such analysis as it deems necessary to arrive at an independent evaluation of an investment in the securities of the companies referred to in this document and should consult their own advisors to determine the merits and risks of such an
investment. The price and value of the investments referred to in this Research Report and the income from them may go up or down and investors may realize losses on their investments. This Research Report is provided by SIE on a strictly confidential basis for the exclusive use of the
recipient and should not be reproduced, redistributed, or passed on, directly or indirectly, to any other person or published, copied, in whole, or in part, for any purpose. SIE does not provide tax advice to its clients, and all investors are strongly advised to consult with their tax advisers regarding
any potential investment.
This Research Report is not directed or intended for distribution to or use by any person or entity who is a citizen or resident of or located in any locality, state, country, or other jurisdiction, where such distribution, publication, availability, or use would be contrary to law, regulation or which
would subject SIE and/or its affiliates to any registration or licensing requirement within such jurisdiction. The securities described herein may or may not be eligible for sale in all jurisdictions or to a certain category of investors. Persons in whose possession this document may come are
required to inform themselves of and observe such applicable restrictions. This material should not be construed as an offer to sell or the solicitation of an offer to buy any security in any jurisdiction where such an offer or solicitation would be illegal. Recipients shall be solely liable for any
liability incurred by them in this regard and will indemnify SIE and/or its affiliates for any liability it may incur in this respect.
Emerging securities markets may be subject to risks significantly higher than more established markets. In particular, the political and economic environment, company practices, and market prices and volumes may be subject to significant variations. The ability to assess such risks may also be
limited due to significantly lower information quantity and quality. By accepting this document, you agree to be bound by all the provisions as may be applicable pursuant to it. Certain statements made herein may not be based on historical information or facts and may appear to be “Forward
Looking Statements” including those relating to general business plans, future financial condition, and growth prospects. The actual results may differ materially from any “Forward- Looking statements” due to a number of factors, including socio, political, competitive environment, force
majeure, etc. SIE makes no representation or warranty, express or implied, as to the accuracy, completeness, or fairness of the information and opinions contained in this document. SIE, its affiliates, and the employees of SIE and its affiliates may, from time to time, effect or have effected an
own account transaction in or deal as principal or agent in or for the securities mentioned in this document. They may perform or seek to perform investment banking or other services for or solicit investment banking or other business from, and any company referred to in this report.
This Report has been prepared on the basis of the facts, figures and information, which is already available in publicly accessible media or sources believed to be reliable or developed through an independent analysis by the Analysts of SIE. Such information has not been independently verified
and no guaranty, representation of warranty, express or implied, is made as to its accuracy, completeness or correctness. While we would endeavor to update the information herein on a reasonable basis, SIE and its affiliates are under no obligation to update the information. Also, there may be
regulatory, compliance, or other reasons that prevent SIE and its affiliates from doing so. Neither SIE nor its affiliates or their respective directors, employees, agents, or representatives shall be responsible or liable in any manner, directly or indirectly, for views or opinions expressed in this
report or the contents or any errors or discrepancies herein or for any decisions or actions taken in reliance on the report or the inability to use or access our service in this report or for any loss or damages whether direct or indirect, incidental, special or consequential including without
limitation loss of revenue or profits that may arise from or in connection with the use of or reliance on this report.
SIE and/or its affiliates and/or employees may have interests/positions, financial or otherwise, in the securities mentioned in this Report. To enhance transparency, SIE has incorporated disclosure of interest statement in this document. This should however not be treated as an endorsement of
views expressed in this Report:
DISCLOSURE OF INTEREST & MATERIAL CONFLICT OF INTEREST STATEMENT:
SIE with its affiliates (reference shall individually or collectively refer to its respective relevant SIE’s affiliates as the context may require) :
i. Are engaged in multiple financial service businesses inter-alia investment / merchant banking, lending, broking, distribution, wealth management, asset management, portfolio management, equity capital markets services including institutional equities & research;
ii. Actual, potential, or apparent conflict of interest may arise as a result of various business activities carried out by SIE and its affiliates.
iii. SIE and its affiliates and / or its respective personnel and / or its advisors may in future engage in activities that may result in conflict which may not have been captured herein.
iv. SIE and its affiliates has policies and procedures in place for addressing such conflicts. SIE, its affiliates and its directors/partners/employees shall exercise a standard of good faith in its dealings.
This file was generated for vishnu.agarwal@wipro.com from Avendus Spark Research website on July 17, 2025 06:30:53 Page 65
 Avendus Spark Disclaimer VIJAYARA Digitally signed by
VIJAYARAGHAVAN S

GHAVAN S 06:49:22 +05'30'

Date: 2025.07.08

This file was generated for v i s h n u . a g a r w a l @ w i p r o . c o m from Avendus Spark Research website on July 17, 2025 06:30:54

SIE and/or its affiliates and/or employees and/ or their relative(s) may have a financial interest in the subject company. SIE and/or its affiliates may have actual/beneficial ownership of one percent or more securities of the subject company or may from time to time, have a long or short position
in, act as principal in, and buy or sell the securities or derivatives thereof of companies mentioned herein at the end of the month immediately preceding the date of publication of the Research Report. SIE and/or its associates and/or its employees have not been engaged in market making
activity for the subject company.
In the last twelve-month period ending on the last day of the month immediately preceding the date of publication of the Research Report, SIE and/or its affiliates and/or employees may have:
i. managed/co-managed public offering of securities for the subject company;
ii. received compensation for investment banking/merchant banking/brokerage services / other products/ services of the group from the subject company of this Research Report;
iii. engaged in any other transaction involving such securities and earn fees or other compensation as an advisor or lender/borrower to subject company or may have any other potential conflict of interests with respect to any recommendation and other related information and opinions at the
time of publication of Research Report or at the time of public appearance.
iv. provided services to the subject company as a client.
v. Research Analyst views on subject company may vary based on Fundamental research and Technical Research. Investment Advisory (both as portfolio manager or investment advisor or Fund management activities) or Proprietary investment / trading by SIE’s affiliates maintains arm’s length
distance with Research Team as all the activities are segregated from SIE’s research activity and therefore it can have an independent view (including take contrary view / investment advise / investment exposure) with regards to subject company for which Research Team have expressed
their views.
The affiliates of SIE has not received any compensation or other benefits from third party in connection with the research report
However, the above shall have no bearing on the specific recommendation made by the Analyst(s), as the recommendation made by the Analyst(s) are independent of the view of the affiliates of SIE. No part of the analyst's compensation was, is, or will be related, directly or indirectly, to the
specific recommendations or views as expressed in this report.
Research Reports are not exchange traded products, and hence disputes relating to research activities of SIE do not have access to exchange investor redressal or Arbitration mechanism. We also hereby declare that our activities were neither suspended nor have we defaulted with any Stock
Exchange authority with whom we are registered in the last five years. We have not been debarred from doing business by any Stock Exchange/SEBI or any other competent authorities, nor has our certificate of registration been cancelled by SEBI at any point in time.
ADDITIONAL DISCLAIMER FOR US INVESTORS:
SIE is not a registered broker - dealer under the U.S. Securities Exchange Act of 1934 and under applicable state laws in the United States, as amended. Accordingly, in the absence of specific exemption under the Acts, any brokerage and investment services provided by SIE, including the
products and services described herein are not available to or intended for U.S. persons. This report is intended for distribution only to "Major Institutional Investors" as defined by Rule 15a-6(b) under the Securities Exchange Act of 1934 and interpretations thereof by SEC (henceforth referred to
as "major institutional investors"). This Research Report must not be acted on or relied on by persons who are not major institutional investors.
SIE has entered into a chaperoning agreement with a U.S. registered broker-dealer, Avendus Capital Inc (“ACI"), affiliate a FINRA registered broker-dealer (CRD # 150160) in the State of Delaware with its principal office at 445 Park Avenue, 19th Floor, New York 10022. Any investment or
investment activity to which this Research Report relates is only available to major institutional investors and will be engaged in only with major institutional investors. In reliance on the exemption from registration provided by Rule 15a-6 of the U.S. Securities Exchange Act of 1934, as amended
(the "Exchange Act") and interpretations thereof by the U.S. Securities and Exchange Commission ("SEC") in order to conduct business with Institutional Investors based in the U.S., Any business interaction pursuant to this report will have to be executed within the provisions of this chaperoning
agreement. This research report is distributed in the United States by ACI, accepts responsibility its dissemination in the United States. Further, ACI has not authored this research report nor has independently verified any of the data presented within this report. The information provided is
based on the research and analysis conducted by SIE. ACI does not recommend nor endorse this report. Furthermore, ACI does not have any direct business relationships with SIE nor does receive any direct compensation for the distribution of this report.
The Research Analysts contributing to the report may not be registered /qualified as research analyst with FINRA. Such research analyst may not be associated persons of the U.S. registered broker-dealer, ACI, and therefore, may not be subject to FINRA Rule 2241 restrictions on communications
restrictions on communication with a subject company, public appearances and trading securities held by a research analyst account.
ADDITIONAL DISCLAIMER FOR UK INVESTORS:
For the purposes of distribution within the United Kingdom, this communication is exempt from the financial promotion restriction in Section 21 of the Financial Services and Markets Act, 2000 relating to the communication of an invitation or inducement to engage in investment activity on the
grounds that it is made to those persons falling within the following Articles of the Financial Services and Markets Act 2000 (Financial Promotion) Order 2005, as amended: Article 19 (Investment Professionals) and Article 49 (High Net Worth Companies). Any investment to which this
communication relates is only available to investment professionals and high net worth companies.
CERTIFICATION BY EACH OF THE AUTHORS OF THIS REPORT:
The Analyst certifies that the views expressed in this Research Report are a representation of the Analyst's personal opinions on the stock or sector as covered and reported. The Analyst is principally responsible for the preparation of this Research Report and does not have any material conflict
of interest at the time of publication of this Report. The Analyst(s) has not served as an officer, director, or employee of the subject company in the last 12 month period ending on the last day of the month immediately preceding the date of publication of the Research Report. The Analyst does
not have actual/beneficial ownership of one per cent or more of the securities of the subject company, at the end of the month immediately preceding the date of publication of the research report.
Registration granted by SEBI, and certification from NISM in no way guarantee performance of the intermediary or provide any assurance of returns to investors.

Investment in securities market are subject to market risks. Read all the related documents carefully before investing.
Spark Institutional Equities Private Limited is a financial services company incorporated in India, with its registered address at EA Chambers Tower II, No 49, 50, 5th floor, Whites Road, Royapettah, Chennai – 600 014
Website: www. avendusspark.com; CIN: U65999TN2022PTC149473; SEBI Registration No. INZ000307037 (Member of NSE and BSE); SEBI Registration No. INH200009722 (Research Analyst)
Compliance Officer details: Mr. Mahesh Agal; Contact on +91 8080023220 or email: ie.compliance@avendusspark.com; Grievance Officer: T K Ramaswamy - investorgrievance@avendusspark.com, Tel: +91 44 4344 0078

This file was generated for vishnu.agarwal@wipro.com from Avendus Spark Research website on July 17, 2025 06:30:53 Page 66