DigitalBridge CEO Warns of AI Infrastructure Bubble After $4B SoftBank Sale
Marc Ganzi cautions that high leverage and aggressive spending in AI data centers mirror the fiber-optic overbuild of the late 1990s.
Marc Ganzi, the CEO of DigitalBridge, is warning that the current surge in artificial intelligence infrastructure investment may be entering a dangerous bubble. The caution comes shortly after Ganzi agreed to sell his data center firm to SoftBank Group Corp in a $4 billion all-cash transaction.
The deal, agreed upon on December 29, 2025, and approved by shareholders in April 2026, places Ganzi in a unique position as both a major beneficiary of the AI boom and a skeptic of its current trajectory. Ganzi has described the market as being in a "toppy-esque moment," drawing a direct parallel to the late 1990s fiber-optic overbuild. During that era, massive investments in network infrastructure preceded a significant market crash once capacity far exceeded immediate demand.
The Risk of Overleverage
While the underlying technology of AI is viewed as legitimate, Ganzi's primary concern lies in the financial structures supporting the build-out. He specifically highlighted the risk of high leverage among newer market entrants. According to Ganzi, some newer competitors are operating with loan-to-value ratios between 70% and 80%, a stark contrast to the more conservative approach taken by institutional-grade operators.
This aggressive borrowing suggests that many firms are betting on immediate and massive economic returns to service their debts. If the revenue from AI services fails to scale as quickly as the physical infrastructure is deployed, these highly leveraged firms could face insolvency, potentially triggering a broader contagion across the tech sector.
Industry Implications
This warning arrives as SoftBank and other global giants aggressively pursue AI infrastructure, pouring billions into GPUs and massive data center campuses. The concern is that capital expenditures have become decoupled from actual economic utility. If the industry is indeed repeating the 1990s pattern, the current phase of hyper-growth may be followed by a severe correction in valuations and a sharp contraction in infrastructure spending.
What to Watch
A significant correction would likely hit the providers of the hardware and the real estate that houses it. Investors are now watching to see if AI applications can generate enough sustainable cash flow to justify the current spending levels. While the SoftBank acquisition of DigitalBridge underscores the continued appetite for these assets, Ganzi's warning serves as a reminder that the gap between infrastructure capacity and economic return has historically led to volatility.