Moonshots Summary: June 9, 2026 - by Peter H. Diamandis

Moonshots Summary: June 9, 2026

Brian Armstrong on Bitcoin, Trump Eyes Stakes in AI Labs, and SpaceX Builds the First Node of a Dyson Swarm

TLDR: The containers we built last century are too small for the decade ahead. Three trillion-dollar IPOs are testing whether there’s enough money in the world to price them. SpaceX is manufacturing the first piece of an orbital compute grid in a Texas field. And in a quieter lab, NewLimit is preparing to put an age-reversal drug into humans next year. Bitcoin is becoming the settlement layer for an economy of machines. Washington is improvising a way to give citizens a stake in companies that may soon dwarf the economy itself. Money, government, and death itself are being rewritten at once.

This week I welcomed a special guest: Brian Armstrong, CEO of Coinbase and co-founder of NewLimit. Here’s what we covered.

BITCOIN, QUANTUM, AND THE AGENT ECONOMY

Citibank projected Bitcoin could reach $189,000 by the end of 2026, a number no major bank would have floated five years ago. Brian’s read was measured: it’s never as good or bad as it seems. He thinks Bitcoin bottomed around $60K, sees AI and stablecoins pulling attention away short-term, and remains long Bitcoin as the new digital gold. Salim noted that about 5% of all Bitcoin trades daily versus 0.5% for gold, an accessibility edge that wins over time.

On quantum risk, Brian sees no imminent threat but thinks it’s almost certain someone eventually builds a machine that challenges the cryptography across the whole internet, not just Bitcoin. The chains are already moving: Bitcoin’s BIP-360 proposal, Ethereum roughly 20% through its post-quantum roadmap, and a Coinbase advisory council of cryptographers. The open debate is what to do with the early “Satoshi coins,” an estimated 5 to 10% of all Bitcoin sitting in exposed addresses.

The heart of it was the agent economy. AI agents are becoming paying customers, using crypto wallets to autonomously buy services. Brian put current volume at roughly 100 million transactions and around $50 million in value. His strategy is to make Coinbase “the financial account for AI.” Through the Base protocol, an agent can spin up a self-custodial wallet instantly, no KYC, and settle in USDC in under a second for less than a cent.

“The AI economy will eventually be bigger than the human economy.”– Brian Armstrong

“Once you can give AI agents the ability to transact, you enable the machine-to-machine economy, which is going to be millions of times bigger than the human-to-human economy.”– Salim Ismail

Brian and Coinbase are giving the “great unbanked” population of AI agents an economic voice!

WASHINGTON WANTS A STAKE IN THE LABS

Trump called a government stake in the AI giants “a beautiful thing” and floated giving pieces to the public. There’s precedent: the government already holds stakes in 20+ private companies, including 10% of Intel and 15% of MP Materials. Separately, Sam Altman met Senator Bernie Sanders, who wants 50% of equity from top AI companies moved into a public fund. Altman called 50% far too much, but the principle is now a live negotiation.

“The government has infinite power of taxation. It doesn’t need to own equity to extract any amount of money it wants. Owning equity only inhibits progress.” – Dave Blundin

Salim framed the deeper problem: these labs are becoming civilization-scale infrastructure, like electricity, and no governance structure exists for a technology that touches every job, industry, and government service at once.

THE TRILLION-DOLLAR IPO PILE-UP

OpenAI filed its S-1 to go public later this year, the third trillion-dollar IPO behind Anthropic and SpaceX. CFO Sarah Friar reportedly worried in April about the company going public, given roughly $600 billion in compute contracts. Alex argued OpenAI rebuilt itself since then: leaning into Codex for revenue, cutting side projects, and switching Stargate from ownership to leasing to de-risk from GPU price swings. Dave’s concern is liquidity: three back-to-back IPOs, each roughly 10x larger than anything in history. Is there even enough money in the world to absorb them?

“The demand for intelligence is almost infinite. But the cost is going to fall way more than Moore’s Law.”– Brian Armstrong

Brian doesn’t think the AI tooling layer holds the durable value. He expects 80% of workloads within 12 to 18 months to run on models that are 99% cheaper for the same intelligence.

SPACEX BECOMES A HYPERSCALER

Google, with its own TPUs, is paying SpaceX roughly $11 billion a year through 2029 to access 110,000 NVIDIA GPUs inside xAI’s data centers. One deal took the SpaceX IPO from 100x revenue to 50x. Salim couldn’t quite believe Google needed external compute at this scale. Dave’s answer: everything is sold out. Alex’s read is that Elon has put Grok on de facto life support to play a longer game, become the biggest hyperscaler first, then buy his way back to the frontier.

Then SpaceX unveiled AI1, its first AI satellite built for compute in space: 150 kilowatts of peak compute, a 70-meter wingspan, 110 square meters of deployable liquid radiators. Elon says it’s simpler than a Starlink satellite, mostly solar cells and laser links. To build them, SpaceX announced a Gigasat factory in Bastrop, Texas: 1,000+ acres, 11 million square feet, fully integrated from solar ingots to finished satellites. Alex flagged the reason: SpaceX has shown concept videos of mass-driver launches off the Moon. Vertical integration of solar and compute is what lets you eventually manufacture these on the lunar surface.

“AI infrastructure just stopped being a real estate problem and became a launch problem. SpaceX isn’t a rocket company anymore. It’s a civilizational infrastructure company.”– Salim Ismail

Morgan Stanley projects SpaceX revenue could grow from $18.7 billion in 2025 to $3.4 trillion by 2040, valuing the company at $50 to $100 trillion.

My caution: “This is a fundraising tool ahead of the IPO, and Morgan Stanley profits from the deal. I believe in SpaceX long term, but I hope retail investors don’t get hurt in the frenzy.”– Peter Diamandis

LONGEVITY: BRIAN’S OTHER MOONSHOT

Beyond Coinbase, Brian co-founded NewLimit with CEO Jacob Kimmel and Blake Byers, and it just raised $435 million for age reversal. The idea traces to the Coinbase IPO: in 2021, Brian started hosting dinners asking top scientists what was exciting but underfunded. The answer was epigenetic reprogramming, built on Shinya Yamanaka’s Nobel-winning 2012 discovery that four proteins can turn an old cell young.

NewLimit aims to do half of what Yamanaka did: change a cell’s age without changing its type. AI explores the roughly 10 quadrillion protein combinations, then pooled screens, then functional assays (does a reprogrammed liver cell process alcohol like a young one?), then trials. Brian thought this was a 5-to-10-year effort. Instead, NewLimit has reprogrammed human cells and has its first drug candidate entering the clinic next year. I hold Ray Kurzweil’s 2033 target for Longevity Escape Velocity; Alex argued LEV may be “spiky” like AGI, with early GLP-1 signals suggesting a subpopulation could pass it before we notice.

“The baby that’s going to live to a thousand years old is already alive.”– Salim Ismail

Catch the full episode wherever you get your podcasts, and join us at the Moonshots Gathering in Los Angeles.