David Sacks spent 130 days as the Trump administration's AI and crypto czar, and by his own description, he spent most of that time trying to clear regulatory underbrush. Now that his stint as a special government employee has ended, the official White House AI role is being restructured into something considerably more ambitious. Sacks will co-chair the President's Council of Advisors on Science and Technology, known as PCAST, alongside Michael Kratsios. The council's initial membership reads like a who's-who of Silicon Valley's most powerful figures: Jensen Huang of Nvidia, Mark Zuckerberg of Meta, Sergey Brin, Marc Andreessen, Lisa Su of AMD, Larry Ellison, and Michael Dell, among others.
The transition is worth unpacking, because it represents a meaningful shift in how the administration intends to engage with the technology industry on AI. As a special government employee, Sacks had a defined and time-limited role. PCAST is a federal advisory body with a broader mandate; it can study any issue the president puts before it, produce formal recommendations, and in principle carry more institutional weight than a single political appointee. The near-term agenda covers AI, advanced semiconductors, quantum computing, and nuclear power, with particular attention to the national AI framework released last week.
Bloomberg separately reported this week that Sacks told an interviewer Congress could pass an AI standards framework within months, a surprisingly optimistic timeline given how slowly AI legislation has moved. If accurate, it would suggest the administration sees the current political window, with GOP control of both chambers and a tech-friendly White House, as genuinely usable for passing something before the midterm cycle scrambles the calculus.
What is notable about the PCAST lineup is who is and is not in it. The council is heavily weighted toward hardware and infrastructure: chip manufacturers, hyperscalers, cloud providers, the people who build the pipes. It includes almost no one from the AI safety research community, from labour organisations, from academia (in the traditional sense), or from civil society groups concerned about AI's social effects. The framing is competitive and industrial: America needs to win the AI race, and the people advising on how to do that are the people already winning it.
That is not necessarily the wrong group to convene if the goal is industrial policy. Jensen Huang understands GPU supply chains in a way that no academic economist does. But advisory bodies shaped entirely by incumbents tend to produce advice that benefits incumbents. The history of industry-dominated regulatory processes suggests that what gets optimised for is usually market access, liability limits, and freedom from interference, rather than the harder questions about distribution of benefits and management of risks.
For Sacks personally, the move is a straightforward upgrade. PCAST co-chair is a more prestigious perch than special government employee, it is not time-limited in the same way, and it allows him to engage with a wider range of technology policy questions rather than just AI and crypto. He said explicitly that the broader mandate was part of the appeal. He is also no longer a formal government employee, which reduces some of the conflict-of-interest complexity that attached to his previous role given his venture capital background.
The open question is whether PCAST will function as genuine policy input or as a legitimating body: a way for the administration to point to industry consultation without being bound by it. Federal advisory committees have a mixed record. At their best, they bring technical expertise into policy discussions that would otherwise lack it. At their worst, they become a mechanism for giving powerful interests a seat at the table they were already occupying informally, while providing political cover for decisions already made.
The coming months will reveal which version this is. The national AI framework released last week gives the council something concrete to respond to. If PCAST produces substantive recommendations that shape legislation, it will have been worth structuring. If it produces periodic reports that are politely acknowledged and filed away, it will be one more example of how Washington processes powerful forces without quite governing them.