Key Takeaways

  • Alexey Teplikhin warns that generalist sponsors entering data center megaprojects lack the engineering and supply chain capabilities required for mission-critical builds.
  • Andrew Thomas outlines a multi-year development cycle where securing power, lining up specialized contractors, and hitting ready-for-service milestones expose late entrants to prolonged execution risk.
  • Teplikhin predicts an inevitable wave of project failures among unseasoned developers who entered digital infrastructure simply because it was in vogue.
  • Stalled megaprojects risk sparking a broad sentiment pullback from institutional LPs if underperforming generalists fail to deliver projected infrastructure returns.

The Multi-Year Trap for Inexperienced Capital

Capital follows hype, and data centers are currently soaking up record allocations. Yet building hyperscale digital infrastructure shares almost nothing with standard commercial real estate. As generalist private equity sponsors and real estate developers scramble to rebrand themselves as digital infrastructure platforms, seasoned operators see mounting operational fragility.

Alexey Teplikhin points directly to this growing knowledge gap among new market entrants. “One of my fears about data centers is that many people are trying to invest in data centers now,” Teplikhin explains. “Some of those people have a lot of experience investing in data centers, have a lot of experience constructing specialized mission critical projects and mission critical buildings. Some of these people may not have that experience, and they may invest in data centers because it is an interesting thing that is in vogue now.”

Andrew Thomas outlines the timeline realities that these new entrants routinely underestimate. “You're looking at a development cycle that is going to last for several years from the time in which you can procure power, line up all of the contractors that are required, actually go through construction, and bring a facility to ready for service,” Thomas observes. In a market constrained by grid queues, long-lead electrical switchgear shortages, and specialized engineering deficits, a three-to-five year development window leaves zero margin for amateur errors.

The Fallout from Flawed Execution

The assumption that rising AI demand guarantees underwriting safety is running into physical bottlenecks. Power availability, substation interconnection, liquid cooling integration, and specialized mechanical contracting require institutional track records. When generalist teams run into procurement delays or miscalculate substation engineering requirements, carrying costs compound and project yields collapse.

Teplikhin expects market reality to separate specialized builders from capital chasers. “I do not know whether all of the developments that have been announced will go forward,” Teplikhin says. “I think there is an inevitability that not all the developments will succeed, and that's okay. But one of the things that I'm concerned about is in building data centers, we should realize these are large projects, they are complicated, they take many years, and if you are not experienced in it, it may not work.”

The secondary risk is reputational contagion. When high-profile developments stall, LPs who backed non-specialist managers often misdiagnose operational incompetence as an asset class flaw.

“If it doesn't work, investors may say, 'Oh, I thought data centers were wonderful and easy,'” Teplikhin warns. “And I think the reality is we think that if you have experience and develop data centers correctly, they can be a good investment. As with any investment, if you don't have those things, they may lose money.”

Why It Matters

The data center market is splitting between specialized platforms with secure supply chains and late-arriving generalists holding stranded real estate options. As novice sponsors stumble on contractor availability and grid interconnections, distressed asset sales and recapitalizations will create aggressive buying opportunities for experienced operators. For LPs, manager selection now hinges on actual construction engineering history rather than generic digital infrastructure exposure.