3rd Open Letter to President Trump: The Deal of the Century is Yours to Make.
For Immediate Release
Date: June 26th, 2026
From: Rufo Guerreschi, Coalition for a Baruch Plan for AI(~2,000 words)
(This open letter is included in the 50-page Strategic Memo of the Deal of the Century v2.8)
One-Pager for President Trump
What: Analysis of the feasibility, rationale, and implementation pathways for an extraordinarily bold and timely US-China-led AI treaty, including a citizen co-ownership framework of top AI and robotics firms, through a US-led international "Apollo Program of AI diplomacy".
Why: Secure US economic leadership. Reduce economic and military risks associated with US-China competition. Prevent the fast-rising, immense risks of AI-driven cyber, chemical, and nuclear threats and loss of control. Ensure Americans fairly share in AI wealth and power. Establish an unprecedented legacy while future-proofing America’s long-term economic and technological leadership.
Who: President Xi has repeatedly expressed support for strong international AI governance mechanisms. So have Bannon, Beck, Rogan, Carlson, the Vatican, Christian leaders and most AI scientists. In different ways, all four top AI labs, Amodei, Altman, Hassabis, and Musk, have called for it. By now, 63% of US voters believe (ungoverned) AI will escape human control, 53% believe it will destroy humanity, and 77% call for a strong global AI treaty.
Timelines: Most top insiders and several key reports warn that AI capable of engineering pathogens that kill millions or permanently escaping human control is potentially within the coming years. The head of Anthropic Institute and co-founder, Jack Clark, places the odds of losing control of AI at 30% in 2027 and 60% in 2028.
Roadmap: On an accelerated interagency timetable, appoint classified and unclassified feasibility working groups. Within 2-3 months, make a private case to Xi to set up an unprecedented diplomatic and cooperation effort to determine urgent mitigation measures, treaty-making processes, and mutually trusted treaty-enforcement and ultra-high-bandwidth diplomacy infrastructure, involving thousands of extremely well-resourced officials and experts: a "global AI Apollo Program of AI diplomacy".
Our Help: Our independent, non-partisan Coalition of 10 NGOs, 20 expert advisors, and a unique 356-page Strategic Memo with 24 contributors, is ready to support President Trump and his partners on multiple fronts: feasibility analysis, actionable-path design, and especially in catalyzing support and converging intent among key influencers and powerbrokers of US AI policy, from lab leaders to leading political, media, and religious figures.
Dear Mr. President,
You may not know that, barely 10 minutes after you were born, in Queens at 10.54 am, June 14th 1946, President Truman's envoy, Bernard Baruch, presented across town to the UN what remains by far the boldest and most ambitious international treaty proposals ever presented.
The Baruch Plan prescribed exclusive, international and federal control of all dangerous nuclear activities and sharing of the benefits of nuclear energy among all nations and their citizens. While it technically failed due to a Soviet veto, leading experts believe the root cause was Truman's September 1945 decision to agree to its core terms first with anti-communist allies, ignoring the recommendation of his Secretary of War, Stimson, backed by Wallace and Acheson, to agree first with the Soviets. This irreparably eroded the Soviets' trust and cost precious months, contributing to the Cold War.
While no one can really tell whether this remarkable fact is a coincidence or predestination, it is undeniable that you stand in the same situation — for AI and with China, this time. You can succeed where Truman failed, complete what he started, and close what is by far the greatest deal of human history.
Most striking is the fact that you have already avoided Truman's pitfalls. You already avoided the allure of hyper-ideological China hawks to avoid a deal with "evil" China or pursue an international AI governance framework or treaty with democratic allies, hoping to impose those terms on China later. You already starkly highlighted the radical limitations of UN-led treaty-making processes.
During a key UN Security Council Session, the OSTP Director stated on your behalf that the US "totally rejects" the UN as the conduit to international AI governance, rhetorically supported by China, and called instead to rely on the prudence and " cooperation of statesmen" — nearly exactly the route Stimson suggested in 1946 should be led first.
To address this opportunity, I convened the Coalition for a Baruch Plan for AI in July 2024, a coalition of small US and international NGOs and over 20 top advisors, including those formerly from Yale, Princeton, and the National Security Agency, and launched its Deal of the Century initiative.
Together with 24 global experts, last December 2025, we published what is by far the most detailed feasibility analysis and strategic blueprint for a US-China-led AI treaty: a 356-page Strategic Memo of the Deal of the Century (v2.6) based on over 667 sources. The Memo also painstakingly described how and why a critical mass of political and media figures, AI labs leaders, and power brokers can support you in realising what a large majority of citizens have been calling for ever more loudly. This document constitutes the updated v2.8 version of its Executive Summary.
So far, publicly, there are only minimal hints of dialogue on a US-China AI deal, light-years away from what is needed at this stage for a treaty. Yet, the last four weeks have set in motion dynamics that have radically increased the incentives for and the urgency of a bold and proper US-China-led AI treaty for both you and President Xi.
The last four weeks have brought an earthquake in US AI policy, which moves from a complete hand-off approach to the validation that AI safety is so important as to block the most valuable and advanced AI systems globally, and even over-swing towards prospects and fears of semi-nationalisation of AI labs, as Amodei recently remarked.
Concurrently, both Bannon and Sanders introduced proposals that 50% of the equity of AI labs should go to citizens — one to individual citizens and the other to a sovereign fund — while Amodei and Altman support the idea, if not the percentage, as do the bipartisan Human Movement and the influential One Project, and you commented positively on it.
A growing majority of your voters fear that AI could take their jobs, livelihoods, and even lives. Yet any serious unilateral regulation would hand China an advantage in a winner-take-all AI race. Only a deal with President Xi can solve both problems at once.
No leader in the world has ever faced a challenge that is even close to the scale of this one. You are destined to become humanity's Superhero, to save America and the world, and to set them on a course of long-term prosperity.
As the Summer approaches and you have two or three more meetings with Xi planned for 2026, you face a rare strategic opportunity. You are uniquely positioned to pursue this deal, as it requires breaking out of old, stale, ineffective treaty-negotiation styles and demands a boldness of vision unseen since 1945.
The opportunity is to co-lead with Xi on an extraordinarily bold US–China-led AI treaty that (1) secures American economic leadership and an incredibly abundant AI future, (2) prevents AI systems from escaping human control, and (3) ensures American citizens will share in the wealth and power that AI creates.
You have a chance to succeed where Truman failed, this time for AI, in a strikingly similar political context, and on terms far more favourable than Truman ever had.
Why Times are Screaming for a Radical AI Policy Shift
Your bold AI growth agenda has ensured America stayed ahead, preventing early AI safety that would have been disastrous for US competitiveness.
Your January 2025 EO on Removing Barriers to American Leadership in AI set the basic direction: accelerate innovation, remove unnecessary federal obstacles, and restore U.S. dominance. Your July 2025 EO's, America’s AI Action Plan and Promoting the Export of the American AI Technology Stack, translated that direction into nearly one hundred federal actions to expand AI infrastructure, adoption, standards, exports, and national-security capacity. Your November 2025 EO Genesis Mission laid the foundation for an ingenious, safe and durable model of American Innovation Leadership through AI.
Yet, over the last few weeks and months, the ground has shifted; a radical new AI policy is needed to build on your successes, face new challenges, and seize a historic opportunity.
Concerns about AI concentration of power, job loss, loss of control, human replacement or disempowerment, and post-humanist futures have been fast-rising among scientists, voters, lab leaders, media personalities, and your own advisors.
By now, 63% of Americans believe humanity will lose control of AI; 53% believe AI will destroy humanity. 57% disagree that it will benefit "everyday people". Americans deeply dislike the direction of ungoverned AI and are increasingly alarmed by it. Yet the upside of AI for all Americans remains astounding, if properly steered (!). This is a political liability that could become a major strategic opportunity.
After the worldwide cybersecurity scare surrounding Mythos Preview and mounting safety concerns from experts and leaders, you approved your 1st voluntary AI safety EO last June 2nd, laying a solid foundation for the process by assigning the right offices and agencies.
In recent weeks, after mounting concern about the concentration of wealth and influence among leading AI firms, nearly concurrent proposals by Sanders and Bannon of 50% of AI stocks going to Americans, you publicly expressed openness to arrangements under which the American people would become partners in leading AI companies and share in the immense wealth AI may create.
The Opportunity: A Historic Deal with China for AI Safety and Future-Proofing US Economic Leadership
The recent rapid increase in unwanted AI cyber and biological behaviours, such as Mythos, has led you to issue your June 2nd AI safety EO. This measure can never be fully reconciled with the ASI race with China. The US cannot resolve or even mitigate such concerns on its own, since unilateral regulation loses ground to China without containing it at all. That's why the only way is a historic and bold US-China-led AI treaty.
While presumably justified by undisclosed national security reasons, the Department of Commerce's emergency suspension of Anthropic's latest model, Fable, last June 12th was triggeredbyexportcontrols, via a private channel and based on flimsy public evidence, rather than the neat process outlined in your June 2nd AI safety EO. This has gravely undermined the trust of AI labs and many others in the USG's ability to regulate AI safety properly, competently, in good faith, and without hidden political or economic motives.
Two weeks later, on June 25th, your administration asked OpenAI to stagger the release of its next model, GPT-5.6 — the first pre-emptive restriction of an American model before launch. What these episodes reveal is bigger than any one model: today's giant, general-purpose AI has grown so unpredictable that it trips national-security alarms on contact. That is exactly why a global treaty setting shared standards is not a brake on American innovation but an accelerant — it can channel investment toward AI that is controllable, predictable, and therefore trustworthy enough to deploy in the high-value domains, from medicine to defense, where today's models are still too risky to use.
As it stands 12 days later, it has become a defactoAIlicensingregime —highly arbitrary and obscure for the time being—sitting atop the voluntary process sketched by the June 2nd EO.
The suspension, paired with AI leader safety warnings, has shown that a voluntary oversight regime is wildly inadequate for hyper-critical technologies, as it assumes a huge amount of blind trust in labs that have even declared their intent to build Superintelligence, AI beyond human control, thereby forcing other agencies to step in.
This chaos is ultimately the result of a deep, uncomfortable truth: that serious mandatory AI safety oversight and safeguards would fatally compromise US leadership on frontier AI with China. So, only a proper, clear, trustworthy, and accountable international regime has a chance, if done well, of being trusted by the USG, AI labs, and US citizens, as well as by China, other nations, middle powers, and the rest of the world.
The political ground is ready: 77% of US voters — including 78% of Republicans — support a strong international AI agreement. This is the rare issue where your base, the centre, and even your critics already agree with you.
A wide MAGA and bipartisan movement is growing for a treaty to ban ASI and prevent the concentration of power in AI, including the Pro-human AI movement, The Human Movement, and Humans First. Bannonthinks the loss of human control (i.e., singularity) could happen in 12-18 months and has been calling for "immediate treaties" on AI. With Glenn Beck and many others, he has signed a call to ban Superintelligence. Rogan and Carlson have been similarly vocal on AI risk.
Most AI lab leaders, while rightly pushing forward in the absence of a global deal on AI, have been publicly calling for stronger international coordination on AI safety, but have recently grown sceptical. Sam Altman called for an "IAEA for AI" and said the US should"try really hard" for a US-China AI safety deal. Musk has been calling for stronger AI regulation for 10 years, warning for a decade that AI was more dangerous than nukes. Amodei has been the most outspoken lab CEO on extinction risks and recently said that a treaty would become plausible if it focused on banning the most dangerous architectures and included credible verification.
Yet winning full support from influential AI lab leaders requires a treaty-making process designed to avoid global authoritarianism, anti-innovation bureaucracy, and excessive concentration of power, while still preventing anyone from building ASI. Many of them, alongside Thiel, fear those governance risks as much as, or more than, the immense gamble of letting ASI emerge, even while acknowledging extinction risk (as do Thiel, Musk, Amodei, and Altman).
Christian leaders would also be on your side, as many of them have called for regulating AI. The Pope's key AI advisor, Benanti, has coordinated an appeal for a binding AI treaty, and so did his predecessor, Pope Francis. In May 2025, Vance deferred to the Pope on AI ethics and, seemingly, AI safety. Catholic Rubio is well-placed to help shape a diplomatic opening, given his prudent approach to China and his neutral stance so far on AI.
Effective global enforcement of an AI treaty inescapably requires the treaty's deep, timely adoption by at least a dozen of the most powerful nations. This requires a process that eliminates the veto and ensures highly intensive, time-bound negotiations.
The only way to ensure that is by using themostAmericanoftreaty-makingprocesses, the constitutionalconventionmodel, which, in just 2 months in 1787, produced the US Federal Constitution, as Sam Altman already suggested in 2023. Such a process would reflect realist power asymmetries by applying a vote-weighting based on GDP and technological leadership, thereby providing global stability while durably safeguarding U.S. and Chinese leadership.
Beyond Safety and Leadership: Ensuring that All Americans Share in the Wealth and Power of AI
While a US-China-led AI safety treaty can deliver immense peace and safety to Americans, it would not, by itself, deliver the one other thing that your voters — and many AI lab leaders, increasingly expect: a guarantee that all Americans, not just a few coastal billionaires, share in the wealth and power AI creates.
Business leaders are proposing radical solutions in the form of European-style AI wealth-redistribution schemes, such as Musk's Universal High-Income proposal and Amodei's repeated warnings and proposals. Even Larry Fink, CEO of the largest AI investor, warned of social unrest unless citizens participate more in the equity.
Yet, the idea of "AI dividend" hand-outs for citizens is bureaucratic and un-American.
Any future administration can easily unwind those handouts and give Americans no voice in their AI future. The American way is ownership.
Citizen co-ownership schemes were implemented with mixed success by Reagan's Employee Stock Ownership Plans and Margaret Thatcher's 1980s privatization program, which made most Britons shareholders, and with great success by the AlaskaPermanentFund, which shared the windfall from state oil with Alaskans and benefited from a constitutional lock-in.
But you could and should, Mr President, go far beyond those partial successes, and you are already halfway there. The optimal solution could emerge as a natural evolution of your U.S. Sovereign Wealth Fund, Trump Accounts, and Genesis Mission, ensuring that everyAmericanfamilyholdsadirectstakeinaleadingAIandroboticsfirm whose value largely derives from our collective knowledge.
Sam Altman has argued for variants of citizen AI co-ownership for years; Bannon, Beck, and the emerging Human Movement have MAGA-coded it; you can be the President who actually delivers it.
In the last four weeks, enormous support for it has been growing.
On May 25th, the Pope’s encyclical was entirely centred on sharing AI's power and wealth. Hence, the President can likely count on the Pope’s wholehearted support.
Citizen co-ownership of 50% of top AI and robotics labs was called for by Bannon on May 29th and by Sanders on June 1st, with a radical difference: one to accrue to individual Americans and the other into a federal sovereign fund.
On the same day as Sanders' announcement, Altmancame out very strongly against Universal Basic Income, arguing that it does not give people power and agency.
A few days later, he publicly announced and held an hour-long meeting with Sanders, reportedly stating afterwards that 50% would be too much. On June 6, as Sanders' proposal was all over the news, you commented favourably about Americans' sharing of AI firms' equity.
On June 10, Anthropic CEO Amodeicalled for citizen individual sharing of AI wealth via "universal capital accounts", similar to the US President Trump Accounts. On June 18, in an interview, J.D. Vance backed such comments. He goes further to emphasise the importance of a pre-distribution model rather than re-distribution, which, he says, makes citizens subservient, and states the need for citizens to have a seat at the table, with the "trade union model" being very important. While his reference to trade unions is confusing (as they refer to employees rather than citizens/consumers), it clearly hints at giving citizens co-decision-making power rather than just dividends from a sovereign fund, closer to our proposal and Bannon's than Sanders's.
Yet, to make this measure truly durable and resilient, it would need to be backed by a treaty.
The treaty should mandate that every signatory, including China, grant its citizens a significant ownership interest in its top AI and robotics companies, using methods determined by each nation. This framework would ensure that no future U.S. administration could easily dismantle the progress you achieve.
The role of citizens should be that of shareholders rather than managers. They should benefit from dividends, transparency, and informed voting rights via expert representation, yet remain uninvolved in the day-to-day operations. Such a balance prevents hindering the innovators and founders who are critical to sustaining America's current and future leadership.
The point is mutual check: the treaty restrains any single government, and citizen ownership restrains any single firm, so that no actor, here or abroad, captures what belongs to everyone.
Next Steps: Towards The Deal of the Century
In January 1946, Truman convened a classified Board of Consultants, led by Oppenheimer, Deputy Secretary of State Acheson, and other leading minds, to assess the scientific and diplomatic feasibility of international controls on nuclear proliferation. They delivered the Acheson–Lilienthal Report, a blueprint for the Baruch Plan, after two months.
We propose you do the same: direct your administration to convene both classified and unclassified Boards of Consultants on the AI Deal of the Century. Distinct in mission and tempo from any standing advisory body, such as PCAST, it would have a sole charge: to assess the feasibility, enforcement architecture, and strategic case for a US–China-led AI agreement.
We recommend you ask your senior advisors with relevant portfolios or expertise to participate personally and through their chosen envoys: the White House and the State Dept. National Security Advisor, ODNI, CISA, OSTP, NSA, aligned frontier AI lab CEOs (e.g. Altman, Amodei, Musk, Suleyman/Smith, and Hassabis/Brin), media and financial leaders, and top AI scientists.
And to avoid Truman's mistake, we recommend that you establish a Joint US-China AI Treaty Feasibility Commission and gradually extend it to other middle powers as soon as initial interest is confirmed with President Xi. This way, shared scientific and strategic underpinnings, as well as enforcement mechanisms for a mutually acceptable deal, can emerge early on, boosting confidence and trust.
Our 356-page Strategic Memo of the Deal of the Century v2.6 maps the case, the architecture and likely allies of such a treaty-making process. Our closed-door Washington, DC, convening, September 15–16, 2026 — The Cooperation of Statesmen — is designed to bring key experts and advisors who can help you properly frame, manage and communicate the Deal of the Century to ensure it is a great success.
A treaty of this magnitude would let you finish your second term in triumph in January 2029, as the most consequential American President. You'd achieve a legacy unparalleled by any leader in history, revered at home and worldwide.
A President who prevents the most plausible self-extinction event in history would become politically untouchable. Lawyers or partisans would not defend your legacy, but by the gratitude of a country and a civilization that survived because you acted in time.
The window is measured in months. The history of the 21st century is being written now. It is on you to write it. It is your destiny.
Barely ten minutes after your birth, Truman presented what is still by far the boldest treaty ever proposed — one that would have reliably banned all weapons of mass destruction globally. It is on you to finish what he started, succeed where he failed, by relying on your intuition and unique deal-making skills. Whether you are predestined or not, the greatest deal and legacy in human history is yours to make.
Respectfully,
Rufo GuerreschiFounder & Executive DirectorCoalition for a Baruch Plan for AIrufo@trustlesscomputing.org | cbpai.org
The Coalition for a Baruch Plan for AI brings together 10 US and international NGOs, 20 expert advisors and 24 contributors, including some formerly at the NSA, WEF, Yale, and Princeton.
