Sriram Krishnan exits White House AI policy role

The most striking feature of Krishnan's departure is what comes next. According to The Washington Post, he plans to start an outside institution to continue influencing Trump's AI policy while 'building institutions' that address challenges for 'America and its allies' ^[2]. The Information reported the body would be staffed with engineers expressly to support the White House's AI plans ^[3]. This is a clean instance of the revolving door operating in reverse motion: rather than cashing out to a firm with business before the government, the departing official is constructing a private vehicle whose stated purpose is to keep steering the very agenda he just left. That arrangement keeps his influence over federal AI policy intact while moving it to a private seat, which is precisely why the structure raises questions about disclosure, ethics, and oversight even as it is framed as continuity of mission.

Krishnan was a co-author of the American AI Action Plan, not its sole author; Michael Kratsios and David Sacks shared the pen ^[4]. The plan's core commitments are by now institutional rather than personal: prioritizing data center construction over regulation and framing AI as a race against China ^[1]. Krishnan's own farewell gestures at exactly these durable pillars, citing 'energy, data centers or a clear path for Americans to experience the benefits of AI' as the issues that remain to navigate ^[1]. The substance of the agenda, in other words, does not depend on his presence. What his exit changes is less the policy direction than the caliber of in-house technical translation behind it, the rare person who could move fluidly between an engineering brief and a diplomatic one.

Krishnan's path is now a recognizable archetype for how AI is governed in this administration. He arrived from a general partnership at Andreessen Horowitz, where he ran the firm's London office, after operating roles at Microsoft, Facebook, and Twitter, including helping rework Twitter after Musk's acquisition ^[5]. The trajectory runs from venture capital into the West Wing and now back out into a privately funded policy institution, a loop that keeps the same Silicon Valley operators near the levers of frontier-tech regulation regardless of whether they hold a federal badge. It is a model of AI governance in which the people who funded and built the technology also write and now privately steward the rules for it, blurring the line between industry and the state that is supposed to oversee it.

Timing matters. Krishnan worked most closely with David Sacks over his 18 months in the building, and Sacks himself stepped down from the AI and crypto czar role earlier in 2026 ^[1]. Two of the three named authors of the American AI Action Plan have now exited the West Wing within months of each other, leaving the administration's senior AI bench visibly thinner even as the agenda accelerates. Sacks's send-off, crediting Krishnan with the country's lead in the AI race ^[1], doubles as an acknowledgment of how concentrated that expertise had become in a handful of people. The reaction outside government has tracked that significance: the principal's own announcement was gracious, the tech press treated the private-institution plan as the real news, critics politicized the exit, and diaspora-focused outlets framed it as a setback for the administration. The spread of reactions underscores that this was read as a consequential departure, not a routine staffing change.