AI Infrastructure Debt Risks Have Raised Concern

Venture capitalist Paul Kedrosky warned that current AI financing levels mirror financial risks seen in 2008.

Updated on Sept. 25, 2026 in Artificial Intelligence

Isometric editorial illustration of a modular steel data center frame in muted teal and mustard, representing structural financial risks in AI infrastructure.
Venture capitalist Paul Kedrosky warned that the high level of debt funding currently supporting AI data center construction mirrors risks seen in 2008. AI Illustration. Upload story photo >

Live Poll

Do you believe the current pace of investment in artificial intelligence is creating a financial bubble?

Venture capitalist Paul Kedrosky has compared the current rise in debt-backed AI infrastructure to the 2008 financial crisis. With over 60% of AI financing now debt-backed, analysts worry about the long-term revenue viability of data centers.

Why it matters

Investors are increasingly concerned that the heavy debt loads supporting AI data center construction may not be justified by future earnings. This skepticism reflects broader fears about whether current capital spending can sustain the rapid growth of AI infrastructure.

Current AI financing that is debt-backed has climbed to over 60%, while required returns on certain data-center projects have risen to 10% to 12% due to higher Treasury yields and credit spreads.

The players

Paul Kedrosky

He is a prominent venture capitalist who specializes in analyzing technology markets and macroeconomic trends.

Amazon

It is a major multinational technology company heavily investing in cloud capacity and AI infrastructure expansion.

NBER

The National Bureau of Economic Research is a private, non-profit research organization that produces empirical studies on economic activity.

The details

Companies are leveraging AI to automate the production of software applications, pitch transactions, and investment decks. While an NBER study of 500,000 GitHub developers found autonomous coding agents boosted activity by 240%, these gains diminished to 80% at the project level and only 30% for actual software releases.

Timeline

  1. Debt-backed AI financing reached 15% to 20% of the total in 2025.

  2. Amazon projects capital expenditures of $220 billion in 2026.

  3. Prediction traders see a 10% probability of an AI industry downturn by December 31, 2026.

  4. A refinancing wall is expected for five-year financing structures in 2029.

  5. Refinancing difficulties are projected to continue through 2030.

The Tech Race

This trend of debt-funded growth parallels the speculative environment of the 2008 financial crisis. It highlights a critical shift as the industry moves from experimental coding gains to the high-stakes reality of securing long-term debt for massive data center capacity.

For users and developers, the focus on productivity may shift as the gap between coding activity and actual software releases persists. Investors should monitor refinancing timelines for data center projects as potential indicators of broader tech sector stability.

The takeaway

The gap between high coding activity and actual software releases suggests that AI-driven efficiency gains may be more complex than raw metrics imply. Investors and firms should prepare for potential financial volatility as the current reliance on debt-backed financing faces future repayment pressure.

Further reading

Learn more about the evolving landscape of Artificial Intelligence.

Source note: This article includes information reported by Benzinga.

Live Poll

Do you believe the current pace of investment in artificial intelligence is creating a financial bubble?