The Bank of England is sounding the alarm on something most people aren’t talking about: who’s actually footing the bill for the AI revolution, and what happens when the check comes due.

Sarah Breeden, the BoE’s Deputy Governor for Financial Stability, used her platform at the European Central Bank Forum on Central Banking in Sintra, Portugal on June 30 to highlight a fundamental shift in how AI infrastructure gets funded. The short version: it used to be Big Tech burning through cash reserves and equity. Now it’s increasingly debt. And the path to paying that debt back is, to put it charitably, unclear.

From cash piles to credit lines

The BoE’s Financial Policy Committee conducted an analysis in April 2026 and found that AI infrastructure investment has shifted from being primarily funded by large tech company cash flows and equity to include rapidly rising and increasingly complex debt financing.

Breeden framed the AI ecosystem through four interrelated drivers: capability, capacity (meaning compute), diffusion (meaning adoption), and financing. She specifically called out “how debt and equity finance fund this historically unprecedented pace of investment” as one of the four key channels shaping AI’s scale and timing.

The regulatory trajectory

Breeden’s speech represents a notable evolution in the BoE’s thinking. Earlier work from the central bank, including a 2024 speech and 2025 parliamentary evidence, focused primarily on the risks of generative AI adoption and whether existing regulation was sufficient to govern it. The financing angle, the question of who’s lending how much to fund all of this, didn’t become a central theme until 2026.

The FPC is scheduled to release an updated assessment of AI-related risks on July 7. Breeden also addressed risks posed by agentic AI within cybersecurity, financial markets, and payment systems.