White House AI adviser Sriram Krishnan to step down

Krishnan’s departure matters less for the individual résumé change and more for what it signals about how fluid US AI governance remains. He was a central node between the Trump White House, frontier labs like Anthropic, and investors pushing for “American AI dominance”, helping shape executive orders on model reviews, safety testing, and potential government equity stakes in leading firms.([dawn.com](https://www.dawn.com/news/2005854/white-house-ai-policy-adviser-krishnan-to-leave-position))

His exit comes just as Washington is grappling with two conflicting pressures: local backlash to data center build‑outs and power usage on one side, and geopolitical anxiety about ceding ground to China on the other. Losing a seasoned operator in this space increases the risk that short‑term political considerations dominate over coherent, technically grounded AI policy. It also underscores how intertwined frontier labs and government have become: the same adviser who helped broker a détente after the Pentagon’s blacklisting of Anthropic is now expected to keep influencing policy from the outside.([dawn.com](https://www.dawn.com/news/2005854/white-house-ai-policy-adviser-krishnan-to-leave-position))

For the race to AGI, this reinforces that regulatory and political capacity, not just model capability, will be a differentiator. If the US can’t stabilize its own AI governance bench while labs sprint ahead, the vacuum could be filled by ad‑hoc local moratoria, aggressive state‑level rules, or foreign standards that don’t reflect frontier‑lab realities.