What Happened

Sriram Krishnan announced Saturday on X that he is leaving his position as the White House senior policy adviser for artificial intelligence at the end of June 2026. He's been in the role for roughly 18 months.

"It is hard to express how big a privilege it has been to serve the American people," Krishnan wrote in his post. No specific reason for the departure was given.

According to The Washington Post, Krishnan plans to launch an outside institution focused on technology policy — one specifically designed to let him keep shaping the Trump administration's AI direction from outside the building. Planning is described as being in "nascent stages."

Who Is Krishnan

Krishnan ran product teams at Microsoft, Twitter, Yahoo, Facebook, and Snap. He was most recently a general partner at Andreessen Horowitz — the venture firm whose founders Marc Andreessen and Ben Horowitz publicly backed Trump in the 2024 election.

He came in as part of a wave of Silicon Valley figures who embedded themselves in the second Trump administration. His closest working relationship inside the White House, by his own account, was with David Sacks — the investor-turned-AI-and-crypto czar who stepped down earlier this year and now co-chairs the President's Council of Advisors on Science and Technology.

What He Actually Did

Krishnan was one of the primary architects of the administration's AI Action Plan, according to both The Washington Post and the Let's Data Science summary aggregating Post reporting. That plan prioritized domestic data center construction and rolled back regulatory guardrails — a full pro-industry posture.

The Trump White House also signed executive orders on AI during Krishnan's tenure, including one challenging state-level AI regulations and another focused on federal oversight — though that second order was delayed and narrowed after industry pushed back, according to TechCrunch.

One order released this week directs federal agencies to ask leading AI developers to voluntarily submit their most capable models for government cybersecurity testing before public release, according to The Economic Times. The word "voluntarily" carries significant weight in that directive.

The Tension Inside the White House

Krishnan is leaving a fractured White House AI shop, not a unified one.

According to Let's Data Science — aggregating Washington Post reporting — there are documented internal tensions between the pro-industry tech advisers like Krishnan and Trump's populist political allies. Chief of Staff Susie Wiles and Treasury Secretary Scott Bessent have both reportedly grown concerned about advanced AI models from Anthropic, specifically citing demonstrated abilities to identify software security vulnerabilities that adversaries could exploit.

The administration's own chief of staff and treasury secretary are worried about AI models that can crack open cybersecurity holes at institutions like banks, while the policy shop Krishnan ran was focused on loosening regulation and building more data centers. This represents a fundamental split between "move fast" Silicon Valley types and officials responsible for national security.

The Revolving Door

Krishnan's plan to start an outside institution that "influences" Trump AI policy from the private sector follows a standard Washington pattern: serve in government, shape the rules, then set up shop outside where you can maintain influence with former colleagues.

This is legal and happens across both parties. But calling it "building institutions to tackle big challenges for America" — as Krishnan framed it on X — is a generous description of a typical influence-economy move.

The Washington Post reported his new institution is expected to be staffed by engineers, according to The Economic Times. So it's not just a think tank — it's a technical operation with the explicit goal of staying connected to White House AI decisions.

What Comes Next

Who fills Krishnan's role — and whether they come from the Andreessen Horowitz orbit or from a more security-focused background — will determine which direction Trump's AI policy actually goes in the next two years.

AI policy determines whether powerful AI systems get tested for safety before they're released, who builds the data centers powering them and where, and whether the government has leverage when things go wrong.

Krishnan's exit removes the loudest pro-industry voice from inside the West Wing. Whether that creates space for more security-conscious policy or gets filled by someone with the same agenda remains an open question.