Should I Invest as a Private Equity Firm in Data Centers and AI?

Dr. Michael Nolting -

Artificial Intelligence is not just a buzzword anymore—it's a capital magnet. And few sectors reflect this shift more tangibly than data centers. As AI models grow larger, so do the physical requirements to run and retrain them: electricity, cooling systems, real estate, fiber optics, and increasingly, sovereign control over the physical computing infrastructure.

It’s no wonder that private equity firms like Blackstone, KKR, Blue Owl, and Brookfield are pouring billions into this asset class. Blackstone alone spent $10 billion to acquire QTS in 2021 and has been doubling down since. But the question many are asking now is: Should private equity investors continue to bet on this AI infrastructure boom—or is the frenzy getting ahead of fundamentals?

The Case for Data Center Investment

The Wall Street Journal’s deep dive into this topic paints a bullish short-term picture. Here's the summary:

  • Data centers are now “high conviction” investments for many PE firms, with major players like Blackstone and DigitalBridge competing to build or acquire capacity.
  • Firms like Amazon, Meta, Microsoft, and OpenAI are not just renting space—they’re signing long-term leases, buying land, and demanding capacity years in advance.
  • Companies like QTS, which Blackstone acquired, have become the backbone of the AI race—delivering not only physical space but also high-efficiency cooling, power, and fiber connections.
  • Despite growing fears of oversupply, demand still outpaces capacity—especially in the U.S. and Europe, where geopolitical concerns and digital sovereignty are becoming increasingly important.

But PE Isn’t VC: Innovation ≠ Comfort

While the excitement is justified, private equity is not venture capital. PE firms look for clarity, cash flow, and returns within 5 to 10 years. They don’t thrive in high-uncertainty, bleeding-edge innovation environments. And right now, there are three warning lights flashing:

1. Rising Infrastructure Costs

Building state-of-the-art data centers is more expensive than ever. Energy prices are volatile. Permitting is slower. And cooling technologies—vital for the high thermal loads of AI chips—are pushing design complexity up, not down.

2. AI Model Efficiency Will Improve

The current AI hype is driven by massive retraining and inference demands. But future advancements—such as more efficient model architectures, low-power chips, or AI-native compression techniques—could dramatically reduce compute needs. If that happens, today’s demand projections could look inflated in hindsight.

3. Geopolitical Fragmentation

Countries want their own cloud and data sovereignty. The EU, China, and the U.S. are increasingly requiring that AI and cloud providers localize their infrastructure. While this increases demand in multiple markets, it also limits economies of scale—bad news for margin-obsessed PE firms.

The Long-Term Bet: More Complex Than It Looks

The AI boom has made data centers the “picks and shovels” play of the decade, but that doesn’t mean it’s easy money. Yes, space will continue to be a constraint. Yes, compute power is becoming the new oil. But we must remember: private equity thrives on operational efficiency and predictable cash flow—not disruptive tech volatility.

In that light, here’s my perspective:

  • Data centers remain a solid bet in the short- to medium-term, especially when located in strategic geographies like North America and Western Europe.
  • But the expected exponential AI-driven demand is likely to normalize as innovation makes compute more efficient.
  • Sovereignty pressures will create regional winners, but also limit global optimization strategies that boost profit margins.

Final Verdict

If you're a private equity investor, data centers deserve a place in your portfolio—but not without caution. Focus on high-lease-occupancy assets, long-term tenant contracts (especially from AI leaders), and regional positioning in politically stable areas.

In other words: bet on the infrastructure—but don’t mistake it for innovation. AI will transform the world, yes. But the speed of transformation and its technical path are still far from predictable.

And private equity doesn’t like surprises.