Summary

(Last updated on September 16th, 2026)

The Deal of the Century Initiative is the flagship initiative of the Coalition for a Baruch Plan for AI, uniting 10 small NGOs, 20 expert advisors and 24 contributors in a high-leverage, high-urgency effort to secure a safe and fair global governance of AI.

Our prescient vision was framed in the Open Call that accompanied our December 2024 launch of the Coalition:

"We call on all heads of state, President-Elect Trump and President Xi, and their advisors and security agencies, to engage in an open global treaty-making process for safe and fair AI of radically unprecedented scope, urgency and effectiveness."

Since February 2025, we have further narrowed our focus to persuading Trump with the launch of this Deal of the Century initiative. By September, we had produced a ground breaking 356-page Strategic Memo of The Deal of the Century. 

Since then — via that memo and its updates, direct advocacy, and closed-door roundtables — we have been fostering a convergence of key U.S. AI policy powerbrokers and influencers, and the White House around a whole new, durable and cohesive US AI policy centered on a US-China-led AI treaty-making process that is exceptionally bold, timely, and well-designed. 

For Q4 2026 onwards, our main focus is to hold in Washington DC, a series of close-door DC Roundtables on a Bold US-China AI Treaty, with DC-based think tank and advocacy NGO partners.

How Events Are Imposing a Radical New US AI Policy

The hands-off AI policy maintained by the White House until June 2026 — that arguably proved crucial to safeguarding at least partially US advantage over China, while avoiding grave harms — has ultimately been rendered obsolete by the piling up of evidence of ever more dangerous AI capabilities, and increasing safety and power concentration concerns from all sections of American society.

In a  few weeks, the Overton window has radically shifted and widened. The range and scope of possible solutions have multiplied, all over the map. Previously unimaginable proposals have gone mainstream.

• While the Administration has implemented stop-gap safety measures, bold calls for an international treaty to prevent the permanent loss of AI control have emerged from both President Xi and a coalition of U.S. AI leaders and leading labs.

• At the same time proposals by Bannon and Sanders for 50% citizen co-ownership of AI labs have been supported in principle by Amodei, Altman, Vance and Trump.

• Despite this momentum, most AI lab leaders remain highly concerned that a treaty might be a net negative: fail to materialize; fail to prevent Artificial Superintelligence (ASI); inadvertently create an authoritarian dystopia; or overly regulate, foreclosing exceptional benefits may might bring, including protection from other daunting existential risks.

Although President Xi advocated for a UN-led framework, Trump maintained that AI governance must primarily be established by the two leading AI powers. OSTP Director Kratsios explicitly rejected the UN as a viable platform for AI governance during his address to the UN Security Council, urging last September direct "cooperation of statesmen" instead. This echoes Secretary of War Stimson's 1945 advice to Truman — a recommendation that, had it been followed, might have saved the Baruch Plan.

Towards a Timely, Widely-Shared, and Proper US AI Policy

While developments of the last weeks and months are overall very positive and promising, the policy landscape is dominated by fragmentation and lack of development. Key stakeholders advance a wide array of under-developed solutions, while core concerns are not even named or confronted.

There is an urgent need for extremely high-bandwidth, effective, closed-door processes to enable AI policy confrontation and convergence around among the Administration and key domestic actors, then bilaterally with China, and then eventually with the global community.

Given the short timelines to ASI and the risks of global authoritarianism, the treaty-making process must be: extraordinary in scale, scope and effectiveness; grounded in historical lessons from past international treaties; and guarantee the suitable participation of competent and accountable actors. The "how" is for key actors to determine in the process, but three principles appear inescapable for success:

1. Learning from the 1946 Baruch Plan's failures, and considering global power asymmetries, this effort must begin with a U.S.-China-led AI safety treaty, executed via an extraordinarily bold "global Apollo Program of AI Diplomacy".

2. The initial bilateral phase must be followed by a GDP-adjusted global constitutional convention for AI, as history has shown that this is the only treaty-making model that is even remotely capable of achieving the necessary consensus on such a wide-scoped, within a short-enough and predictable timeframe.

3. The prevention of authoritarianism and immense concentration of power — and the full participation of top AI lab leaders and adherence of middle powers — requires that the treaty include a mandate of citizen co-ownership of large AI and robotics firms and other citizen accountability measures. 

Our Action Lines

• Research. A 356-page Strategic Memo of the Deal of the Century and its executive summaries provide feasibility analysis, path design, and profiles of the US AI policy powerbrokers who will decide.

• Advocacy. Direct, sustained engagement with those powerbrokers' advisors and staff — over 95 to date — and media engagement that keeps the treaty question in front of them.

• Consortium and Roundtables. We are forming a steering board of US AI treaty-advocacy organizations — with full decision rights and a joint fundraising plan — whose first product is the Inaugural Edition of DC Roundtables on a Bold US-China AI Treaty, Washington, D.C., early Q4 2026, anchored by a Washington institutional co-host.

Momentum & Traction

To date, we have engaged over 95 relevant staff and advisors of key US AI policy powerbrokers, whom we have identified though our research in high-ranking officials (e.g. Rubio, Vance, Kratsios), media figures (e.g. Bannon, Carlson, Rogan), and lab leaders (e.g. Altman, Amodei, Musk). Most notably, our engagement with four of these powerbrokers has been direct and sustained, involving over 30 hours of consultation with staff or advisors directly reporting to them.

Four national-security think tanks, one European foundation's DC office and five AI-safety governance advocacy organizations — all of them DC-based or with significant DC presence — have expressed preliminary interest in partnering or co-hosting the DC roundtables.

The Administration: on the right direction, but from the right pace

The Administration's direction on AI governance risk is broadly right on three counts:

1. It favors an initial bilateral framing over a UN process — Kratsios' 'cooperation of statesmen' line at the UN Security Council.

2. Bessent's US-China AI dialogues open on the 'proliferation of dangerous AI among non-state actors' — the risk both sides can most readily agree on, and the one that puts the most competent safety and security agencies at the table

3. Trump, along with Vance, Amodei and Altman, supported in principle, while not in quantity, proposals of last June by Bannon and Sanders for the citizen co-ownership of 50% of large US AI and robotic firms, a proposal that would, if properly implemented, not only prevent, but radically reverse the AI-driven concentration of power and wealth. 

Yet, the scope and plausible timing of immense risks call for such dialogues to be expanded by at least two orders of magnitude in scope and timeliness. 

In our strategic memo, we make a detailed case for it, and provide analysis and insights that we hope can be useful if they choose such a route.

Trump's Recent Statements Against AI Existential Risk

While Trump has increasingly warned of the biological and cyber risks of AI, he has been traditionally vague or neutral on the catastrophic, loss of control and extinction risks to date.

That changed radically since last September 13th. Responding to Amodei/Altman/Musk's "pace the frontier" call he said, "We can put guardrails. We can do this and that. But I think you have a lot of negative forces that are bringing it up that shouldn't be bringing it up, and they're bringing up things that won't happen.". The next day he said fears of "AI taking over the World, destroying Humanity, and all other things bad, is a HOAX," putting it on par with Russian election interference claims and his own impeachments. The same day, he said, "The people that say AI is going to destroy the World... are the same people that said, just a short time ago, that the World would be extinguished by 'Climate Change.'. Trailing along. Vance the same day called the warnings from AI leaders a "Trojan horse."

While those statements could be durable, as his statements about climate change being a hoax, they could primarily derive from conjunctural political conditions, turn out to be temporary, and be easily reversed in a month or so. Here is why and how:

• The timing of the statements is likely in reaction to in a few days of: (1) an upcoming widely-supported US federal bill to ban Superintelligence (the kind of AI that poses existential risks) has been backed overwhelmingly by Democrats, (2) Hillary Clinton calling for a Operation Warp Speed for AI diplomacy and Obama calling for AI to be at the center of Democrats agenda, and (3) possibly Bannon and Sander sharing a stage in DC with mostly Democratic congressmen about such a bill.

• Upcoming treaty negotiations with Xi on September 24th, declaratively also on AI and AI safety, make it strategically apt for Trump to pretend to undervalue something that he knows Xi greatly values (as per consistent Xi Statements) so that he can turn it into a concession to him  during negotiations.

• A face-saving path exists: within a month Trump could distinguish himself from the Democratic bill by citing 'new evidence,' making any US federal ASI ban conditional on a reciprocal treaty with China, while implementing a temporary ASI pause the interim. Our memo is built to make that path concrete.

Roadmap

With the U.S.-China AI safety dialogues commencing in September 2026 led by Treasury Secretary Bessent, we recommend the US government expand it by two orders by appointing urgently but carefully an extremely highly-resourced, multistakeholder AI treaty feasibility working groups on a wartime timetable (supported by expert AI safety and governance non-profits and philanthropic entities). 

These working groups - akin in function to the 1946 Acheson-Lilienthal Report, but much more extensive - will produce  strategic memos that enable the Administration to acquire confidence on a preferred actionable path towards a proper AI treaty.

The process must include temporary bilateral urgent risk mitigation measures (i.e. an initial emergential narrow targeted treaty) and the creation of sufficiently trustworthy and globally trusted treaty-enforcement and diplomatic-communication infrastructure.

China and Xi personally have consistently signaled readiness, and increasing concerns and calls for deeper international coordination. Most recently, on Sept 9th, 2026, while denying accusations by the US of distillation of its models, China redoubled on the dialogues stating, the “two sides should step up cooperation in this field". 

Within two to three months, the US should be ready to fully engage President Xi, launching bilateral feasibility working groups. Given the stakes and timelines, this must be an extremely intense, high-bandwidth endeavor, an Apollo Program for an AI Treaty — first national, then bilateral and then global. 

Download the Latest Strategic Memo · The Roundtables Webpage · Donate · Join