The episode opens in medias res with Kratsios and Diamandis already mid-conversation, cutting together the most provocative lines from the interview to come — a nuclear reactor in space by 2028, an AI-native reimagining of science, and the bold declaration that 'this is the golden age of America.' Kratsios's framing is immediately clear: he believes government must have strong opinions about what matters most for the future, not just step back and let things happen. Peter cues up the formal introduction with his signature line — 'That's a moonshot, ladies and gentlemen' — before rolling into the episode proper.

Peter Diamandis introduces Michael Kratsios in full, describing him as the principal architect of three initiatives shaping America's scientific and technological trajectory: the America AI Action Plan (the administration's AI race roadmap), the Genesis Mission (a Manhattan Project-style AI-for-science effort), and 'Science: A New Golden Age' (a blueprint for unconventional scientific funding). Recording inside the White House itself, Diamandis frames this conversation as one he's conducting on behalf of his 'Moonshot Mates' — his community of forward-thinking listeners. The segment closes with a light-hearted mock press conference bit between the two.

With Ray Kurzweil's prediction of a century of progress in the next decade as backdrop, Peter pushes Kratsios on whether the US government can possibly keep pace with exponential technological change. Kratsios responds not with defensiveness but with a framework: technologies are either 'born free' — like the 1990s internet, where government's best move is to step back — or 'born in captivity,' like AI-powered medical diagnostics and commercial drones, where the technology cannot reach Americans unless government affirmatively changes the regulatory landscape. ^{[1] — Michael Kratsios "Kratsios splits all technologies into two categories: 'born free' (like the 1990s internet) where government should do nothing, and 'born i…" 03:05 Play} The EU AI Act becomes his sharpest illustration of failure: it was written and finalized before ChatGPT was even invented, making it structurally inapplicable to the large language models it was supposed to regulate. For Kratsios, this is the original sin of technology governance — regulating before you understand what you're regulating.

Peter raises the example of the UAE channeling 50% of government operations through AI and Malaysia deploying an AI representation speaking in every language. Should the US have an AI czar? Kratsios says no — firmly. His reasoning is structural: AI is not a single domain but a horizontal technology that will transform every agency simultaneously, from drone regulation at the FAA to financial services oversight at the SEC. The idea that one person could provide the best policy answer across all those domains is, he says, 'a tall, tall order.' He notes that the Trump administration was ahead of this conversation from the start — President Trump signed the first presidential executive order on artificial intelligence in history in 2019, years before the technology dominated the front page. The conversation surfaces an uncomfortable admission: large language models are currently banned from use on White House systems due to the Presidential Records Act.

Peter delivers a brief sponsored read for Google for Startups, directing listeners to the Google Startup Technical Guide for Generative Media — a blueprint for deploying DeepMind's generative AI models in production across images, video, and audio. Link available in the show notes.

Peter opens with sobering polling: three-quarters of Americans fear AI, and 71% oppose data centers in their neighborhoods — a higher share than those who oppose nearby nuclear power plants. Kratsios doesn't deflect. 'AI has a massive PR problem,' he says flatly, before tracing its origin. ^{[1] — Michael Kratsios "Three-quarters of Americans fear AI, and Kratsios blames a decade of government-driven doom narratives — culminating in the Bletchley Park …" 08:44 Play} His diagnosis: the entire prior government narrative around AI was fixated on what could go wrong — bio-risks, job losses, existential dangers — rather than what could go right. The Bletchley Park AI safety summit, which he attended, almost entirely framed the technology through a lens of fear. The result, he argues, is unsurprising: of course people are skeptical when every signal from government and industry leaders emphasizes danger. His proposed antidote is healthcare — the domain where Americans' emotional connection to AI's potential is strongest and most immediate. Meanwhile, he notes the economic reality being ignored: Jensen Huang told him that NVIDIA's order to Corning is the largest in Corning's history, factories are being built, and people are being hired — stories that never reach the public.

Peter presses Kratsios on the most politically charged AI question: is AI destroying jobs or creating them? Kratsios offers a nuanced personal view — he's long-term optimistic about AI's impact on employment, noting that fears of automation-driven job loss in the first Trump administration were never realized, with only employment increases following. But he acknowledges the current data is poor. His solution: a Department of Labor initiative called for in the AI Action Plan to collect better, more granular data on AI's actual labor market impact, including from companies that typically don't report to Labor. He makes a pointed observation — many companies are strategically attributing layoffs to AI because it plays well in the press and boosts stock prices, even when AI had nothing to do with it.

Peter asks Kratsios to paint the long-term vision for America, pushing beyond incremental healthcare wins. Kratsios frames his answer in terms of national missions — bold, specific goals that generate national pride and inspire young people the way Apollo once did. ^{[1] — Michael Kratsios "Returning humans to the Moon by 2028, building the first elements of a lunar base by 2030, and launching a nuclear reactor into space — all…" 16:25 Play} The list is staggering: humans on the Moon in 2028, the first elements of a lunar base by 2030, a nuclear reactor in space (with Mars propulsion capability) by 2028, a scientifically relevant quantum computer by the end of the president's term (directed through executive order), and an evolving national fusion strategy backed by more private investment than at any point in history — including 37 venture-backed fusion companies. Kratsios is most personally excited about quantum's applications in pharmaceuticals, where the ability to model molecular interactions could be transformational. The conversation reveals a government that is, for the first time in decades, willing to stake out ambitious, falsifiable timelines for moonshot-level achievements.

Peter flags a geopolitical issue that's dominating Washington: open-source AI and the risk that China uses it as an instrument of global influence. Kratsios is unusually candid: the US leads in closed models, but its open-source ecosystem is underperforming, and cash-strapped American entrepreneurs cannot be blamed for using cheaper Chinese models — because right now, Chinese models are cheaper. ^{[1] — Michael Kratsios "Kratsios makes a blunt admission: right now, the cheapest open-source AI model available to bootstrapped American entrepreneurs is Chinese.…" 19:55 Play} His response is the American AI Exports Program: a Commerce-led, whole-of-government effort to package the best American chips (NVIDIA, AMD), models, and applications into a complete, turnkey AI stack for foreign governments, backed by Export-Import Bank and Development Finance Corporation financing. The explicit motivation is avoiding a repeat of the Huawei telecom playbook, where a subsidized Chinese stack was rapidly proliferated across the Global South before America could respond. This time, Kratsios argues, America has the advantage — its chip lead over Chinese competitors is widening year over year due to EUV export controls.

A sponsored read for Blitzy describes its AI-driven software development platform, which uses thousands of specialized AI agents to understand enterprise-scale codebases and autonomously deliver 80% or more of development work — with enterprises reporting a 5x engineering velocity increase. Listeners are directed to blitzy.com to schedule a demo.

Peter raises the humanoid robot gap: China has over 150 humanoid robot companies, aggressively subsidized and showcased at national events, versus a handful in the US. Kratsios acknowledges the disparity and points to dramatic new action: the administration just restricted the importation of all non-US humanoid robots that haven't already shipped, a bold protectionist move designed to kickstart domestic supply chain investment. He draws the parallel to a similar drone import restriction in December that immediately triggered significant investment into US drone supply chain companies. The subtext is clear: the administration sees humanoid robots as strategically critical to future manufacturing and national security, and is willing to use trade policy as an industrial accelerant.

Peter invokes Elon Musk — 'the most brilliant engineer on the planet' — who has predicted double-digit GDP growth in 18–24 months, triple-digit growth in 5 years, and a post-capitalist world by 2036 where AI and robotics satisfy virtually any material want. Kratsios doesn't dismiss the prediction but admits it's hard to conceptualize at that velocity. He points to Treasury Secretary Bessant's serious engagement with post-AI financial system implications — specifically around a super-capable cyber AI model that could threaten systemically important financial institutions — as evidence that the administration's senior leadership is genuinely grappling with rapid change rather than ignoring it.

The interview pivots to the substantive heart of the Golden Age Report: why, despite record budgets and better tools, is scientific productivity declining? Kratsios traces the problem to institutional inertia — a system designed decades ago that nobody wants to change, especially those who have succeeded within it. ^{[1] — Michael Kratsios "The NIH budget has ballooned to $45 billion yet drugs are more expensive than ever. Researchers spend 45% of their time on grant paperwork.…" 31:30 Play} The data is damning: the NIH budget has swelled to nearly $45 billion yet drugs are more expensive than ever. A National Academies study found researchers spend 45% of their time on grant administration. The median age of an intramural NIH scientist is 71 — while the average age of Nobel-winning scientific work is in a researcher's 20s. The scientific community, Kratsios observes wryly, is perhaps the community least interested in experimenting with new ways of doing science. The irony is not lost on either speaker.

Beyond the data on administrative burden, Kratsios draws a concrete regulatory example that illuminates the broader problem: the US has had a blanket ban on supersonic flight over land since the 1970s, framed as a noise concern. But research and Boom Supersonic's own demonstrations showed that modern aircraft can exceed Mach 1 without creating a sonic boom — yet the law prevented any testing because flying over Mach 1 was illegal regardless of actual noise produced. The administration fixed it with an executive order replacing the speed limit with a noise limit: if you can keep it quiet, fly. Kratsios frames this as a template — the regulatory system is full of rules written for a world that no longer exists, and the act of replacing the rule's intent (noise reduction) rather than its blunt mechanism (speed cap) unlocks decades of stalled innovation. The same logic, he implies, applies throughout the science funding system.

Peter walks Kratsios through the specific mechanisms in the Golden Age Report for reforming scientific funding. First: long-duration grants — most government grants are roughly 18 months, not because that's the right timeline for discovery but because it's administratively convenient. Some ideas need 5 years to properly explore. Second: fast grants — inspired by Tyler Cowen and others who pooled private capital during COVID to make grant decisions in hours rather than months, with no reason this model should be restricted to emergencies. Third and most structurally important: meta-science units — new departments at NIH and NSF dedicated to empirically evaluating whether funding mechanisms themselves are working and course-correcting when they're not. Kratsios notes that government essentially never runs experiments on its own funding processes, and the meta-science units — already announced at NIH and NSF — are the first step toward treating science funding as a system that can itself be optimized.

The most creative mechanism in the Golden Age Report is the golden ticket: on any merit review panel, each member is given 1–3 golden tickets allowing them to unilaterally fund any proposal, independent of what the rest of the committee thinks. ^{[1] — Michael Kratsios "Each reviewer on a grant panel gets 1–3 golden tickets, allowing them to unilaterally fund a proposal regardless of what the rest of the co…" 39:50 Play} The insight is elegant — the current system systematically defunds the most original ideas because they require unanimous or near-unanimous consensus across a committee of specialists who are personally incentivized to protect the status quo. A golden ticket creates two virtuous loops: applicants with genuinely bold ideas now have a credible path to funding (one true believer is enough), and better people are incentivized to join review panels precisely because they get to exercise independent judgment. Peter shares a Harvard professor's experience of being penalized in reviews for being too successful — a perfect illustration of how the current system punishes excellence. The peer review process, both agree, is 'extraordinarily broken.'

Peter saves what he calls his favorite idea for last — the section of the Golden Age Report that imagines a fully AI-native scientific marketplace. The vision: funders post bounties for breakthroughs, AI agents identify promising leads and 'hire' autonomous labs to run experiments, results are cryptographically signed and verified, and smart contracts automatically release payment as milestones are met. ^{[1] — Peter Diamandis "The most radical proposal in the Golden Age report: a decentralized scientific marketplace where AI agents post bounties, run autonomous la…" 43:20 Play} Prediction markets guide grantmakers toward the highest-value problems; reputation systems identify reliable agents; human experts remain essential only for judgment calls about which questions matter. The entire system operates at machine speed — continuously, 24/7 — replacing the institutional coordination overhead of committees, journals, and grant cycles with market incentives. Kratsios acknowledges this is blue-sky thinking but insists the technical building blocks now exist and that he genuinely believes it's achievable. Peter frames it as the mechanism that could make Elon Musk's 2036 post-capitalist prediction come true.

With the Golden Age Report explicitly endorsing prize challenges and advanced market commitments, Peter steps into his home turf: over 30 years, XPRIZE has launched $600 million in prize competitions and driven approximately $30 billion in R&D — a roughly 50:1 leverage ratio on prize capital. He pitches Kratsios on the idea of 10 $1 billion prizes focused on the nation's most important challenges, including one targeting longevity. Kratsios is enthusiastic but pragmatic: the federal budget can absorb prizes in the $100 million range for sufficiently significant challenges, but $1 billion for a single prize is a stretch for the current fiscal environment. Peter then introduces the $101 million XPRIZE Healthspan — the organization's largest-ever prize, with 830 competing teams — as an example of what's achievable, arguing that reversing aging by 20 functional years would be the single highest-impact intervention on the US economy.

Nick Leonard, CEO and co-founder of VoiceRun, delivers a sponsor read describing VoiceRun as a developer-first platform for building, deploying, and evaluating voice AI agents — built CLI-first for full developer control and enterprise extensibility. The platform is described as capable of building agents, simulating test scenarios, and analyzing performance at scale. Listeners are directed to voicerun.com.

Peter notes that Kratsios's six national technology missions — AI, quantum, commercial fusion, lunar exploration, robotics, and next-generation semiconductors — conspicuously omit longevity and biotech. Kratsios acknowledges this directly, noting that longevity wasn't explicitly listed, though much of the biotech work is nested under the Genesis Mission's AI-for-science umbrella. He confirms that roughly $5 billion in Genesis Mission grants were announced at a recent summit, with many focused on biotech applications. The conversation then surfaces Dario Amodei's claim of potentially doubling human lifespan in 5–10 years, and Demis Hassabis's assertion — delivered to Kratsios personally — that AI could cure all diseases within a decade. Kratsios recalls being genuinely unsure whether Hassabis was joking. He was not.

Peter presents a map his team created ranking US states by their openness to autonomous vehicles, drones, nuclear, and data centers — what he calls 'singularity zones.' He asks Kratsios whether the US could formalize this into designated innovation regions where regulations are relaxed enough to accelerate research, similar to what China and Chile have done. Kratsios enthusiastically endorses the concept, tracing its lineage to the first paper he got President Trump to sign in 2017: the UAS Integration Pilot Program, which created 10 state and local government zones paired with drone operators to run real-world delivery tests with FAA regulatory clearance. That model is now being replicated for eVTOLs through the EIPP. Texas, both agree, is currently the leading state for innovation-friendly policy. Kratsios connects this to data center siting, describing the president's Ratepayer Protection Pledge — requiring AI companies to bring or buy their own power — as the mechanism that turns data center buildouts from community liabilities into community benefits.

Peter opens this segment with visible frustration: as the father of two 15-year-olds, he finds the US education system structurally misaligned with the world his children are about to enter. Still tied to an Industrial Revolution factory model, resisting AI integration, and hobbled by teacher unions and public school boards. Kratsios frames his response around parent-first optionality — the government's role is to provide choices, not mandate a single approach. He highlights Alpha School as a compelling example: 3–4 hours of AI-driven personalized learning in the morning, followed by social skills and other activities in the afternoon. But he's realistic: most Americans get neither the Alpha School model nor a tech-free classical education — they get 'a broken middle' where majorities of students graduate without basic math proficiency. On the STEM side, declining numbers of American students pursuing STEM degrees is, he says, genuinely dangerous for national security and economic growth. He hopes the Apollo-style national missions and space programs will inspire the next generation the way Apollo inspired Diamandis.

Peter and Kratsios discuss the autonomous self-driving lab concept — a fully closed-loop scientific system where AI proposes hypotheses, robotic platforms run experiments continuously, and models immediately read results and design follow-up experiments without human intervention. Lila Sciences is cited as a real-world example, reportedly building out a million square feet of autonomous lab space. Kratsios wants federal funding to test experiments on these platforms, acknowledging that the near-term bottleneck is hardware capable of running a sufficiently broad range of experiment types. He believes universities will be early movers — they have strong incentives to offer the most powerful tools to their scientists. Peter adds the democratizing dimension: these platforms could eventually give a high school student access to national lab equipment, collapsing the institutional gatekeeping that currently controls who gets to run experiments.

Peter challenges the Genesis Mission's stated goal of doubling scientific productivity over the next decade, calling it unambitious by moonshot standards. Kratsios's response is disarming: he agrees. The 2x target was set in late 2025 when the mission framework was being built, and even then he had internal debates about whether to raise it. But events have overtaken the original number. ^{[1] — Michael Kratsios "The Genesis Mission's stated goal of doubling scientific productivity was set before the AI explosion of the last six months. On air, Krats…" 1:04:22 Play} Anthropic's revenue grew from roughly $10 billion to over $70 billion in just seven months — making it the fastest-growing company on the planet. The Mythos and Fable AI releases happened. OpenAI is about to release its 6th model. The pace of AI innovation, he says, is 'just insane.' His conclusion, delivered with evident conviction: 'We've gotta aim big, we've gotta do 10x.' Peter previews the Solve Everything paper he co-authored with Alex Wissner-Gross, which envisions AI progressively cooking every scientific domain — math first, then physics, chemistry, biology — in a cascade of accelerating discovery.

One of the most concrete and actionable insights in the conversation: the US national labs — DOE, NIH, and others — have accumulated 70 years of scientific data across every domain, from nuclear physics to biomedical research, that has never been made AI-ready or incorporated into any model. ^{[1] — Michael Kratsios "US national labs hold 70 years of scientific data across every domain that has never been made AI-ready. Getting that data into models is t…" 1:07:50 Play} Kratsios frames this as an enormous untapped accelerant for the Genesis Mission. His analogy is clarifying: when the federal government made NOAA weather data freely available, it became the foundation for every commercial weather application ever built. The same cascade could happen with national lab data — except instead of weather apps, the downstream products would be drug candidates, materials discoveries, and physics breakthroughs. He also addresses a question Peter raises about who captures the value from AI-driven discoveries built on public data: those are public goods, and the value should flow back to the American people who funded the research in the first place.

Peter raises Argentina's Milei government, which has offered to grant AI agents personhood and eliminate taxes on AI companies — an aggressive attempt to position Argentina as a global AI hub. Kratsios's response is measured: it's an 'interesting take,' but the US is not ready for AI personhood. The administration's focus is on delivering AI's benefits to American citizens first and creating a regulatory environment where American companies — the best in the world — can continue leading. The subtext is one of the episode's cleaner closing frames: for all its ambition, the administration's AI philosophy ultimately reduces to 'let our horses run.'

After a warm close with Kratsios — 'this was so fun, thanks for your work' — Peter transitions to the Moonshots health segment, sponsored by Fountain Life. He introduces Dr. Dawn Musalem, Fountain Life's Chief Medical Officer, for a discussion about early cancer detection. The data point is striking: 3.3% of Fountain Life members who believe themselves to be healthy are found to harbor an undetected cancer through full-body MRI and liquid biopsy screening. Dr. Musalem emphasizes that these are cancers found at early, curable stages — precisely the window that conventional insurance-covered care misses. The segment closes with a direct call to action: visit fountainlife.com/peter to learn about membership and early detection capabilities.