The Trump administration has formalized what many in Washington already suspected: Silicon Valley isn't just lobbying the White House anymore — it's sitting inside it. On March 25, the administration announced the first members of the President's Council of Advisors on Science and Technology (PCAST), placing Nvidia CEO Jensen Huang, Meta CEO Mark Zuckerberg, and Oracle executive chairman Larry Ellison in formal advisory roles on AI and emerging technology. Notably, conspicuously absent from the list are two names that once dominated the conversation: Elon Musk and OpenAI CEO Sam Altman.
What PCAST Actually Is — And Why This Version Is Different
Every president since Franklin D. Roosevelt has assembled some version of a science advisory council. The format has varied — some convening regularly, others gathering dust. Under Trump's executive order signed in January 2026, the reconstituted PCAST was explicitly designed to focus on AI and emerging technologies with a mandate tied to economic competitiveness and national security.
What makes this iteration distinctive isn't the institutional wrapper. It's who's in it — and who's running it. The council will be co-chaired by David Sacks, the White House's AI and crypto czar, and Michael Kratsios, the administration's senior technology adviser. That pairing is deliberate: Sacks is the dealmaker and narrative-setter; Kratsios is the policy architect. Together, they represent the operational axis through which Silicon Valley relationships translate into regulatory and legislative posture.
The council can hold up to 24 members. The initial 13 appointments span a striking cross-section of the tech industry's current power structure.
The Full Roster — And What Each Seat Signals
The White House announcement named the following individuals:
Marc Andreessen — Andreessen Horowitz co-founder. One of the most aggressive voices pushing for AI deregulation and minimal government interference in frontier model development. His presence signals that PCAST will lean pro-innovation over precautionary in any debate about AI guardrails.
Sergey Brin — Google co-founder, who has re-engaged personally with Google's AI work after years out of day-to-day operations. His appointment is a nod to the foundational generation of Silicon Valley — and possibly a signal about the administration's ongoing dialogue with Alphabet, despite antitrust tensions.
Safra Catz — Former Oracle CEO, now a key executive in the Oracle-SoftBank-OpenAI Stargate coalition. Her presence bridges enterprise cloud, federal contracting, and the $500 billion AI infrastructure bet that Stargate represents.
Michael Dell — Dell Technologies founder and CEO. As AI compute demand reshapes enterprise hardware cycles, Dell is one of the few legacy infrastructure players successfully pivoting its server and storage business to capture AI workloads.
Jacob DeWitte — Co-founder and CEO of Oklo, a nuclear fission startup backed by OpenAI's Sam Altman (more on that irony shortly). Oklo is building micro-reactors designed in part to power AI data centers. His inclusion puts energy infrastructure — not just software — at the heart of the AI policy conversation.
Fred Ehrsam — Coinbase co-founder, and an anchor of the crypto-to-AI pipeline that Sacks has championed. His seat reinforces the administration's view of digital assets and AI as part of the same deregulatory reform agenda.
Larry Ellison — Oracle's executive chairman, and arguably the most senior enterprise tech voice in the room. Ellison has become a crucial figure in the Stargate AI infrastructure coalition, alongside Microsoft and OpenAI, and has spoken publicly about using AI for large-scale surveillance applications. His presence reflects how intertwined the administration's AI ambitions are with federal contracting.
David Friedberg — Entrepreneur and investor, known primarily through the All-In podcast circle that includes Sacks. His appointment is the most politically candid of the group — a network effect more than an industry credential.
Jensen Huang — Nvidia CEO and, effectively, the hardware king of the AI era. Every major frontier model runs on Nvidia GPUs. Huang's placement on PCAST cements what his frequent White House visits already suggested: he is the single most important industrial figure in U.S. AI strategy. His comments about declaring AGI achieved will carry different weight with a formal advisory seat.
John Martinis — The lone academic researcher on the list, a physicist and UC Santa Barbara professor who led Google's quantum computing team and achieved the landmark quantum supremacy demonstration in 2019. His inclusion is the council's nod toward scientific credibility.
Bob Mumgaard — CEO of Commonwealth Fusion Systems, a fusion energy startup backed by a coalition of investors including Bill Gates and Khosla Ventures. Like DeWitte, Mumgaard's seat treats energy as integral to AI strategy, not secondary to it.
Lisa Su — AMD CEO, and one of two women on the council (alongside Catz). Su's AMD has become an increasingly credible alternative to Nvidia in AI accelerator markets, and her appointment signals the administration's interest in supply chain diversification across AI silicon — not just reliance on a single chipmaker.
Mark Zuckerberg — Meta CEO and architect of the open-source AI strategy that produced Llama. Zuckerberg has spent the past year positioning Meta as the pro-openness counterweight to closed-model incumbents. His PCAST seat puts open-source AI interests directly into White House deliberations.
The Notable Absences
The omissions are as informative as the appointments.
Elon Musk is not on the list. The Tesla, SpaceX, and xAI CEO previously led the Department of Government Efficiency (DOGE) before a public split with the administration. Musk's xAI is a direct competitor to the OpenAI ecosystem that dominates much of this council's orbit. His absence removes a volatile and unpredictable voice — but also one with arguably the most comprehensive stake in AI infrastructure, from compute to frontier models to autonomous systems.
Sam Altman is absent despite OpenAI's central role in the Stargate coalition. Altman's complicated relationship with the administration — marked by Stargate partnership announcements on one side and ongoing federal procurement tensions on the other — appears to have kept him off the formal list, at least for now. Oklo's Jacob DeWitte being included (Altman backs Oklo) is the closest the list gets to the OpenAI orbit.
There are also no Microsoft representatives on the council — a striking gap given that Microsoft's $13 billion OpenAI stake and Azure cloud dominance make it one of the most consequential actors in U.S. AI infrastructure. Whether that absence reflects political calculus or simply the first round of appointments remains unclear.
The Federal AI Bill: What PCAST Is Really For
The timing of the PCAST announcement wasn't incidental. It coincided with the administration's push at the Hill & Valley Forum in Washington — an event that brought together defense tech executives, venture capitalists, and government officials to align on technology and security priorities.
At the forum, Kratsios laid out the administration's most ambitious legislative goal: passing the first comprehensive federal AI law before the end of 2026. The framework, released the week before, targets child safety provisions for AI-facing products, federal preemption of state AI laws, and protections for citizens from data-center-related rate increases. "We want to create an environment where innovators have certainty about the way that they can develop their products," Kratsios said, "and it's something that only Congress can provide."
The federal preemption goal is the most politically loaded element. The Senate voted 99-1 last year to reject an AI regulation moratorium, showing bipartisan reluctance to strip states of their AI oversight authority. A patchwork of state laws — from California's SB 1047 successor bills to Texas and New York enforcement frameworks — has made compliance planning difficult for companies operating nationally. PCAST members like Andreessen have been vocal critics of state-level regulation; their formal advisory role now gives that preference institutional weight.
Sacks, speaking at the forum, offered a blunter framing. Connecting the Iran war's drone strikes on UAE and Bahrain data centers — which disrupted Amazon's AWS infrastructure in Bahrain — to the urgency of U.S. AI dominance, he argued that America's AI partnerships abroad are now military stakes. "If UAE's data centers were serving China, I don't think they would be getting bombed right now by Iran," Sacks said.
Advisory Power — Or Something More?
Advisory councils have a historically mixed record. They convene, they produce reports, and sometimes those reports become policy. More often, the relationship is more diffuse: the council's existence normalizes executive engagement with an industry's priorities, signals to regulators which voices carry weight, and provides political cover for decisions already being made.
This PCAST will probably function in all those modes simultaneously. The members have a direct line to Sacks and Kratsios — two officials with operational authority over AI policy execution. The first meeting date hasn't been announced, but the White House indicated additional appointments are still coming, suggesting the council isn't fully constituted.
What's clear is the intended direction. Every member of this council has a material interest in AI development proceeding with minimal friction — whether from regulation, foreign competition, or energy constraints. The White House AI framework released last week aligns with that interest. PCAST formalizes the relationship between the policy document and the industry that benefits from it.
For the companies not represented — and for the critics who argue this council prioritizes industry incumbents over public interest — the next several months will reveal whether PCAST functions as genuine advisory input or as a legitimizing apparatus for decisions already taken.
