THIS WEEK IN AI
Each week I synthesize the key developments in AI and public policy so you don’t have to. Here’s what I’m watching:
• AI companies increasingly relying on debt to finance compute
• A draft White House order targeting state-level AI laws paused
• Europe delaying enforcement of high-risk AI rules
Together these stories show how AI is reshaping the global regulatory landscape
1. Debt-Fueled AI Development Comes Into View
The Daily reported that many AI companies are funding compute and model development with debt, often at high interest rates, as the real costs of training frontier models continue to exceed revenue. Economists interviewed likened the trend to other leverage-heavy sectors where investment cycles outpaced fundamentals.
What I’m watching:
Early policy discussions around AI tend to focus on privacy, safety, and national security. But leverage could eventually become part of the conversation too, with more scrutiny on transparency and monitoring as the industry grows more capital-intensive.
2. White House Pauses Draft Order Challenging State AI Laws
Reuters reported that the White House has paused a draft executive order that would have directed the Department of Justice to challenge state-level AI laws on interstate commerce grounds. The draft also contemplated an AI Litigation Task Force and a Commerce Department review of state rules, with the possibility of withholding certain federal funds.
The order hasn’t been released publicly, but reporting indicates it would have moved toward a more federalized approach to AI policy.
What I’m watching:
Even with this order paused, the tug-of-war between federal and state authority on AI isn’t going away. This draft order looks like a second attempt at the 10-year ban on state AI laws that failed in the Big Beautiful Bill earlier this year. Expect continued friction as both levels of government test how much influence they should have.
3. EU Delays Enforcement of High-Risk AI Rules to 2027
The European Commission announced it will delay enforcement of major “high-risk” provisions of the EU AI Act—from biometric identification to health, hiring, and law enforcement tools—pushing deadlines from 2026 into late 2027. Supporters say the delay will give industry more time to adapt and reduce compliance burdens; critics warn it could weaken digital protections. The move comes as part of a broader “Digital Omnibus” effort to simplify overlapping digital laws, including GDPR, e-Privacy rules, and the Data Act.
Several member states have pushed for these adjustments, arguing that Europe risks falling further behind the U.S. and China in AI if its regulatory environment becomes too complex or burdensome.
What I’m watching:
Europe is known for strict digital regulation, setting a high bar for privacy, competition, and consumer protections. This moment suggests the EU may be recalibrating, weighing competitiveness more heavily than before. If Europe shifts its posture, the overall balance globally may shift with it
The views expressed here are Candice's and not those of any organization she is affiliated with.
