Back to Archive
Sent to 912 readers

Race to AGI - Tuesday, December 9, 2025

Share:

TLDR

Trump moves to centralize US AI regulation

Read national AI rule coverage →

US taxes but allows Nvidia H200 exports

Details on H200 export decision →

BIDU +3.5% amid relaxed chip export stance

View BIDU profile →

TSLA -3.4% despite broader AI policy tailwinds

View Tesla profile →

HN questions real demand for enterprise AI

Join the HN discussion →

The Full Story

US AI governance took a clearer shape today: Washington intends to both write the rules at home and meter access to compute abroad. On the domestic front, Trump signaled an executive order to create a single national AI rule and to block state-level AI regulations, effectively pre-empting California-style experimentation in favor of centralized federal authority. A companion "ONE RULE" framing reinforces a trend we've been tracking in our narrative on federal dominance over AI regulation. Externally, the US will allow Nvidia's H200 accelerators to be exported to China, but with a 25% fee layered on top. This keeps Washington's hand on the tap of high-end compute while preserving significant revenue for Nvidia and some access for Chinese hyperscalers. The mixed message—permission with a price—underscores that compute remains a primary instrument of industrial and national-security policy. Markets leaned into the idea that China's AI ecosystem won't be fully choked off: Baidu gained roughly 3.5%, while US chip peers were mixed, with Broadcom modestly higher and Intel lower. Tesla, often traded as an AI proxy, fell more than 3%, reminding investors that autonomy narratives do not shield from execution or valuation concerns. Downstream, IBM's agreement to acquire Confluent for $11B positions IBM as a more integrated enterprise AI data platform provider, betting that compliant, well-governed data streams will be a prerequisite in a federally defined AI regime. Yet social sentiment is more cautious. A widely read Hacker News thread argues there is limited real demand for enterprise AI products today. The resulting picture: policy and infrastructure are centralizing rapidly, but sustainable pull from end users is still unproven. That gap will define which AI platforms convert today's regulatory and hardware advantages into durable economic moats.

Get This Delivered Daily

Join thousands of AI professionals who start their day with Race to AGI.

No spam, ever. Unsubscribe anytime.