Blackstone AI Chief Jas Khaira Says AI Buildout Debt Fears Are Overblown

TL;DR

• Blackstone's head of AI Jas Khaira says fears of an unsustainable debt bubble in AI infrastructure are overblown, arguing hyperscaler and enterprise demand remains far ahead of supply.
• The private equity giant points to contracted cash flows, low vacancy rates, and Big Tech's balance sheets as reasons data center spending is sustainable despite soaring capex forecasts.
• For investors, Blackstone sees AI data centers, power, and connectivity as a multi-decade opportunity, urging focus on fundamentals over short-term market volatility.

A Confidence Play Amid Market Jitters

Wall Street has spent the past few weeks fretting about how Silicon Valley is paying for the artificial intelligence boom. With Oracle, Meta, Microsoft and others taking on billions in debt to fund massive data center buildouts, analysts and investors have warned of an AI infrastructure bubble built on leverage.

Blackstone isn't buying it.

Jas Khaira, Blackstone's head of AI and a senior managing director in its technology investing business, this week pushed back hard on the idea that AI-related borrowing has become dangerous. Speaking amid a fresh wave of volatility in tech credit and equities, Khaira argued the spending underpinning the AI buildout is not speculative — it's tied to real, contracted demand.

In Blackstone's view, the market is confusing aggressive spending with unsustainable spending.

Why Blackstone Remains All-In on Data Centers

Few firms have more at stake in the thesis than Blackstone. The world's largest alternative asset manager has emerged as one of the biggest private funders of the AI revolution through portfolio companies like QTS Data Centers in the U.S. and AirTrunk in Asia-Pacific, two of the largest data center platforms in the world.

Khaira said occupancy and forward leasing trends across Blackstone's portfolio remain exceptionally strong, with hyperscalers routinely pre-leasing capacity years before it comes online. That visibility, he argued, is what makes the debt financeable.

Unlike the dot-com era's speculative fiber buildout, today's AI data centers are typically built only after long-term contracts are signed with investment-grade tech tenants. The result: predictable, utility-like cash flows that can support higher leverage than traditional real estate or corporate debt.

The Debt Question: Sustainable or Systemic Risk?

Market anxiety has centered on a sharp rise in AI-linked borrowing. Oracle's aggressive bond issuance to fund cloud expansion, CoreWeave's debt-fueled GPU purchases, and reports of Big Tech exploring off-balance-sheet financing vehicles have revived memories of past credit cycles.

Credit spreads on some data center and AI-exposed borrowers have widened, while equity investors have punished stocks viewed as overextending for AI capex.

Khaira countered that the absolute numbers look large only because the opportunity is large. Global data center capex is projected to run into the trillions over the next decade, but it is backed by the world's most cash-rich companies and underwritten by secular growth in cloud, AI inference, and enterprise digitization.

He also pointed to the diversity of financing now available — bank debt, investment-grade bonds, private credit, and equity joint ventures — as a sign of market maturity, not distress.

What Big Tech Balance Sheets Signal

A key pillar of Blackstone's confidence is who is ultimately paying the bills. The primary customers for AI infrastructure are Microsoft, Amazon, Google, Meta, and other hyperscalers with massive cash reserves and strong free cash flow generation.

Even as capital expenditures hit record highs, Khaira noted these firms continue to frame AI infrastructure as a necessary investment to avoid capacity shortages, not optional expansion. Executives at Microsoft and Alphabet have repeatedly told investors that the risk of underbuilding outweighs the risk of overbuilding.

For lenders and equity investors behind the physical infrastructure, that Big Tech backstop significantly de-risks the debt stack.

Beyond the Hype Cycle: Power, Chips, and Inference

Khaira also sought to reframe the AI trade as broader than training large language models. While the first wave of demand was driven by training clusters packed with Nvidia GPUs, the next wave will be driven by inference — running AI models for millions of users and businesses — which requires distributed, always-on capacity.

That shift, Blackstone argues, extends the investment runway well beyond 2026. Power access, land, cooling systems, and grid connectivity are now the binding constraints, not just chips.

Blackstone has increasingly linked its AI infrastructure strategy to energy investment, betting that access to cheap, reliable power will determine winners in data centers.

What It Means for Investors

For public market investors whipsawed by AI volatility, Khaira's message was clear: don't confuse funding volatility with demand destruction.

Blackstone advises investors to distinguish between high-quality, contracted AI infrastructure assets and more speculative corners of the AI trade. Data centers with long-term leases to hyperscalers, strong power positions, and prime locations look very different from unleased developments or highly levered intermediaries without contracted revenue.

Private markets, Khaira suggested, are better positioned to fund long-duration assets like data centers that require billions in upfront capital and years to scale — precisely the kind of patient capital Blackstone and other private equity giants aim to provide.

The Bottom Line for the AI Buildout

Whether Khaira is right will depend on whether AI revenues eventually catch up with AI capex. Skeptics warn that enterprise adoption and monetization are still uneven, and that even creditworthy tenants could retrench if returns disappoint.

But for now, Blackstone — with tens of billions committed to digital infrastructure — is signaling it sees the current selloff as noise, not a turning point.

In its reading, the AI buildout isn't a leverage-fueled bubble. It's the early innings of a multi-decade physical rebuild of the internet, and debt is simply the tool to build it.

AndroGuider Team

Articles written by the AndroGuider team. We try to make them thorough and informational while being easy to read.