Beware the data centre bubble (stage)

Hello,

This August, Indian M&A rose to $15.9bn driven by record M&A volumes and two landmark cross-border transactions. 

KKR made a few headlines when finance chief Lewin predicted PE firm consolidation. Meanwhile, KKR and Blackstone have chosen to turn India into Asia’s buyout HQ as investors look beyond China, a shift which may indicate more competition for targets and higher valuations.

Further, Vedanta ‘beat’ the Adani Group with a ₹17,000 crore ($2.04 billion) bid to acquire JAL. The deal will see Vedanta pay ₹4,000 crore ($480 million) upfront and the rest over the next six years.

Worth mentioning is that edtech was having a day – interest was invited for resolution of Byju’s insolvency case, while PhysicsWallah filed to raise $437m from an India IPO.

And finally, Urban Company was sold within hours of listing and Amazon completed BNPL lender Axio’s acquisition.

I hope you enjoy this week’s roundup — please connect on LinkedIn to discuss your next M&A deal.

Let’s dive in.

Share your thoughts on AI for a chance to win $100

Please complete this short survey and you’ll be entered into a prize draw to win one of three $100 Amazon vouchers.

How are you currently using AI during M&A deals? We’re surveying top dealmakers on how they use AI as part of a major research project. 

Deal Tracker

Our weekly roundup of confirmed M&A deals in India.

Market Trends

White elephant much?

Aphorisms such as ‘the data centre is the computer,’ reportedly spoken by Chase Lochmiller of US-based cloud provider, Crusoe; or ‘I don’t know any company, industry or country who thinks that intelligence is optional, it’s essential infrastructure,’ from Nvidia’s CEO Jensen Huang – may for now explain why capital is sprinting into data centre (DC) markets globally. 

In India, a flurry of trophy transactions in 2024 and lease demand heavily skewed toward hyperscalers have made the market irresistible to sponsors and strategics. 

JLL projected industry capacity to jump toward 1.8 GW by 2027, while Colliers pegged India’s potential at 4.5 GW by 2030 with $20bn or more in fresh investment.

But the same forces that justify aggressive growth also seed fragility. Are investors chasing it? Could the DC frenzy produce stranded white elephant assets?

Monolithic bets

India’s deal list reads like a who’s-who of infrastructure and hyperscale capital: multi-hundred-million and -billion investments by Adani, STT GDC, AWS, CapitaLand, CtrlS, Sify, Princeton Digital and others have validated valuations and attracted follow-on capital. 

Reports show India crossed the 1 GW milestone in 2024 and is adding roughly a quarter gigawatt per year in the immediate horizon, with multi-GW plans to 2030. 

Demand is concentrated and for now predictable, and leasing is dominated by a small number of high-volume occupiers. Hyperscalers account for the lion’s share of leasing, followed by BFSI and large tech firms, meaning deals with hyperscaler pre-lets or strategic partnerships can command premium pricing and lower revenue risk. 

But that concentration is a two-edged sword: if hyperscalers pause or change strategy, the pipeline will shrink fast.

Bull market or bubble?

There are many reasons to have a positive view of the current market. Rent cash flows can behave like long-dated infrastructure attractive to pensions and infra funds, and strategic owners (hyperscalers, telcos, large corporates) want regional capacity to deploy cloud and AI.

Meanwhile, PE will see classic roll-up plays across smaller operators and development pipelines.

However, there are also clear fault lines. Concentration risk is real: Mumbai alone accounted for nearly half India’s pipeline, so a single snag with land, power or interconnection can delay hundreds of MW.

And power supply is not trivial: data centres are energy-hungry, and underestimated grid upgrades or slow renewables deals can quickly erode the project economics that underpin expected IRRs. 

The AI story that justifies many builds also holds substitution risk. Model architectures, hosting patterns and silicon choices are evolving fast, and a move toward compact on-prem or edge LLMs, or hyperscalers keeping more inference on custom silicon in their own campuses, could shave material portions of projected colocation demand. 

Finally, financing and exit assumptions matter: higher rates or weaker buyer appetite at stabilisation would compress returns and leave developers exposed.

What next for dealmakers?

For M&A teams there are a range of defensive options: insist on take-or-pay terms, lock long-dated power or credible brown-to-green pathways, structure phased capex with option value and seller earn-outs, and stress-test models assuming 20-50% of hyperscaler leasing is delayed or replaced by smaller-model deployments. That approach separates resilient assets from potential white elephants.

India’s DC market is a rare infrastructure-plus-tech opportunity: large, investible and strategically vital. Its upside is a concentrated hyperscaler pipeline and strong policy tailwinds; the downside is concentrated execution and a plausible tech pivot that could leave parts of the pipeline redundant. 

For M&A teams the best strategy isn’t to ignore the build, but it’s to structure deals that buy optionality, secure demand and de-risk power and regulatory execution. Get those three right and you win; miss them and you could be left chained to a white elephant.