Despite growing concerns about an AI investment bubble, the world’s largest private equity firms are doubling down on data center investments that power artificial intelligence infrastructure. At the Goldman Sachs Financial Services Conference on Wednesday, industry titans expressed overwhelming confidence in their AI-related bets.
Blackstone President Jon Gray revealed that AI data center investments have become the firm’s biggest moneymaker, while Ares CEO Michael Arougheti reported that international data center investments are accelerating beyond expectations and boosting revenue projections. Blue Owl co-CEO Doug Ostrover declared the firm “incredibly bullish” on its data center portfolio.
The investment thesis centers on a massive supply-demand imbalance in computing capacity. Apollo CEO Marc Rowan noted that wherever he travels globally, the world’s largest compute users consistently tell him they need more capacity—but won’t get it anytime soon due to natural limits, energy constraints, regulatory hurdles, and zoning restrictions. Ostrover emphasized he has “never seen a market” with this level of imbalance, with demand accelerating while supply remains constrained.
Gray framed the opportunity around a transformative vision: “If you subscribe to our worldview that bringing superintelligence at scale at very low cost is going to be transformative, you try to figure out—how can I invest in that and take the least amount of risk?”
However, these investors are carefully managing long-term renewal risk. The critical question is whether data center leases will be renewed 15-20 years from now as AI technology evolves. Rowan candidly admitted that expert projections for 2030 energy usage vary wildly—“the spread is like a child throwing darts”—and uncertainty extends to chip usage, compute demand, and quantum computing’s impact.
To mitigate risk, firms are employing strategic approaches. Blackstone only invests in data centers with 15+ year leases from large market-cap companies, refusing to break ground without secured commitments. Blue Owl leverages triple-net-lease structures with tech giants like Microsoft, Meta, Google, and Apple—A or AA-credit tenants providing identical 20-year terms previously reserved for retail chains like Walgreens or Cracker Barrel, but with significantly better credit profiles.
This strategy ensures profitability even in worst-case scenarios. Ostrover noted that even if facilities become worthless at lease end, Blue Owl can still achieve “teens returns”—though the firm expects meaningful residual value. Rowan summarized the dichotomy: credit investors can profit without renewal risk, while equity investors bet on renewals with potential for massive upsides or total losses. “There will be both great fortunes made and lost in the equity of data centers,” he predicted.
Key Quotes
If you subscribe to our worldview that bringing superintelligence at scale at very low cost is going to be transformative, you try to figure out — how can I invest in that and take the least amount of risk?
Blackstone President Jon Gray articulated the firm’s investment philosophy, framing AI data centers as the safest way to capture value from the anticipated superintelligence revolution while minimizing downside risk.
The experts in this have no idea on energy use, much less chip use, compute, the impact of quantum. Do I really want to, with my credit hat on, take renewal risk?
Apollo CEO Marc Rowan highlighted the extreme uncertainty in AI infrastructure projections, noting that even top consulting firms’ 2030 energy forecasts vary wildly. This uncertainty drives his firm’s strategy to avoid long-term renewal risk in credit investments.
I have never seen a market with this level of supply-demand imbalance, and I see demand accelerating, but I don’t see the supply increasing.
Blue Owl co-CEO Doug Ostrover described the unprecedented opportunity in AI data center investments, emphasizing that structural constraints prevent supply from meeting explosive demand growth.
There will be both great fortunes made and lost in the equity of data centers.
Marc Rowan warned that while credit investors can profit with limited downside, equity investors face binary outcomes—either massive returns if AI demand sustains or significant losses if the market shifts, highlighting the high-stakes nature of AI infrastructure bets.
Our Take
This conference reveals a fascinating paradox: Wall Street’s smartest money is simultaneously bullish and cautious on AI infrastructure. The private equity giants are essentially arbitraging uncertainty—they believe in AI’s transformative potential enough to deploy billions, but structure deals to profit even if their thesis partially fails. The focus on investment-grade tech tenants with 15-20 year commitments is particularly telling; it suggests these firms view current AI leaders as durable, not vulnerable to disruption. However, Rowan’s admission that 2030 projections are essentially guesswork underscores how unprecedented this technological shift is. The credit versus equity divide creates a natural market structure where conservative capital funds infrastructure buildout while risk-tolerant capital bets on long-term value. This may actually accelerate AI development by ensuring capital flows regardless of bubble concerns. The real question is whether today’s hyperscalers will still dominate in 2040—or if these data centers will require expensive retrofits for quantum or other paradigm shifts.
Why This Matters
This story reveals how institutional capital is reshaping AI infrastructure investment despite bubble concerns, signaling confidence in AI’s long-term trajectory while highlighting sophisticated risk management strategies. The massive supply-demand imbalance in data center capacity represents a fundamental bottleneck for AI development, with energy constraints, regulatory limits, and zoning restrictions creating natural barriers to expansion.
The involvement of blue-chip tech companies (Microsoft, Meta, Google, Apple) as anchor tenants with 15-20 year commitments demonstrates corporate America’s conviction that AI computing needs will remain substantial for decades. This institutional backing provides validation for AI’s transformative potential beyond speculative hype.
However, the candid acknowledgment of renewal risk and uncertainty around 2030 projections reveals that even the most sophisticated investors recognize significant unknowns in AI’s evolution. The divergence between credit and equity strategies—with credit investors avoiding renewal risk while equity investors chase higher returns—illustrates how different capital sources are positioning for various AI future scenarios. This bifurcation suggests the market is pricing in both transformative success and potential disruption, with implications for how AI infrastructure develops and who ultimately captures value from the AI revolution.
Related Stories
- Goldman Sachs Hires Google’s Melissa Goldman as Tech Head for AI Push
- CEOs Express Insecurity About AI Strategy and Implementation
- Meta and Nvidia Billionaires’ Wealth Soars $152B in AI Boom
- AI Lessons from the Dot-Com Era: Steve Case’s Perspective on the AI Revolution
- Business Leaders Share Top 3 AI Workforce Predictions for 2025
Source: https://www.businessinsider.com/blackstone-apollo-blue-owl-data-centers-lease-renewals-2025-12