IPO Giga Boom, World’s Fairs of the Past | Diet TBPN
TL;DR
TBPN framed the coming AI IPO wave as a “giga boom,” not a mini boom — they argued SpaceX, OpenAI, and Anthropic going public could trigger tens of billions in demand, especially after new Nasdaq fast-entry rules push passive index money into the stocks automatically.
SpaceX was the centerpiece because its IPO could reset the whole market’s scale — the hosts said it may debut above Tesla’s current roughly $1.34 trillion market cap, cited a Financial Times claim that Musk could end up controlling two of America’s most valuable companies, and marveled that Tesla itself IPO’d at just $1.7 billion.
The show tied AI demand directly to old-school hardware bottlenecks — Micron’s launch of one-alpha DRAM in Manassas, Virginia was presented as a big deal because agentic AI and longer context windows are driving an industry-wide memory shortage.
A rumor about Microsoft canceling an internal Claude Code license turned into a broader point about token economics — even though a proposed community note disputed that cost was the reason, the hosts used the story to highlight how companies may push developers toward in-house tools and obsess over token price optimization.
They opened on an AI safety flashpoint: Trump reportedly killed a proposed frontier-model review program after David Sacks intervened — the leaked executive order would have let companies submit models for government review 90 days before release, and the hosts used it to riff on whether nuclear-level AI risk should be governed by the state or by trusted companies.
The back half swerved into a genuinely fun history lesson on World’s Fairs — from the telephone and Heinz ketchup in Philadelphia in 1876 to the Ferris wheel and popcorn in Chicago in 1893 to the 1964 Picturephone, they used the fairs to show how breakthrough tech and silly consumer novelty have always arrived side by side.
Summary
Trump, David Sacks, and the frontier-model review fight
The show opens with a Politico scoop: David Sacks reportedly made an 11th-hour appeal to President Trump to kill an executive order that would have created a voluntary program for frontier AI companies to submit models to the government 90 days before release. Trump’s quote was pure competitive framing — he said he didn’t want anything that would “get in the way” of America leading China — and the hosts bounced between real safety questions and jokes about whether people would rather trust the government or “Disney with a nuclear weapon.”
The Microsoft-Claude Code rumor and the dogfooding argument
Next they hit an anonymous post claiming Microsoft canceled an internal Claude Code license because token-based billing got too expensive. They note a proposed community note disputes that explanation and says Microsoft really wants developers using its own Copilot CLI instead, which led to a good riff about “eating your own dog food” and why Cursor, Claude Code, and Codex got strong in the first place: the people building them actually use them constantly.
Micron’s DRAM run and the memory squeeze behind agentic AI
The third news item is Micron starting production of one-alpha DRAM in Manassas, Virginia, which they call the most advanced memory ever made in the US. The hosts connect it directly to the AI agent boom: longer context windows need more memory, creating a real industry shortage, and they’re visibly stunned by Micron’s stock run while reading off eye-popping gains.
Why this isn’t a mini IPO boom — it’s a “giga boom”
From there they dive into the day’s main story: the press calling the coming IPO cycle a “mini boom,” which they think is absurd. With SpaceX expected to go public next month and OpenAI and Anthropic potentially close behind, they argue this is a once-in-a-generation AI-finance event, especially because new Nasdaq rules mean passive funds may be forced to buy these names and sell rivals.
SpaceX, trillion-dollar comps, and enterprise software on Mars
The SpaceX section is the most animated part of the episode. They talk through the company’s filing, its reported $15 billion spend on Starship, the claim that it sees a $28.5 trillion opportunity including data centers in space, and Musk’s performance shares tied to building a permanent human colony of 1 million people on Mars. Their favorite joke is that after all the cosmic rhetoric, the monetization path still somehow sounds like enterprise software.
Starship delays, IPO optics, and Goldman sliding into Musk’s DMs
They also cover the scrubbed Starship launch, which was delayed by a faulty hydraulic pin at the launch tower, and debate whether it really matters for the IPO. One host initially shrugs it off before conceding that if SpaceX itself lists Starship as its top risk factor, investors will absolutely care. Then comes a lighter Wall Street beat: Goldman CEO David Solomon reportedly worked through staff to DM Musk directly while banks jockeyed for pride of place on the IPO.
The World’s Fair history lesson: telephones, ketchup, Ferris wheels, and FaceTime in 1964
The final big segment is a delightfully chaotic walk through American World’s Fairs. They love the juxtaposition of civilization-shifting inventions and random snack breakthroughs: Philadelphia in 1876 had the telephone and Heinz ketchup, Chicago in 1893 gave us the Ferris wheel and popcorn, St. Louis in 1904 turned into a “Cambrian explosion of snacking” with hot dogs, hamburgers, peanut butter, ice cream cones, and cotton candy, and the 1964 New York fair had AT&T’s Picturephone charging $16 for three minutes — basically expensive proto-FaceTime.
Exoskeleton legs, BYD in F1, and a classic Friday sign-off
The episode ends in true TBPN style with quick hits instead of a neat conclusion. They joke about a Hypershell exoskeleton with “thousand-watt hips,” speculate about the embarrassment potential if someone hacked it, mention CATL investing in DeepSeek, and note that BYD is in talks around Formula 1. Then it’s off to the weekend with newsletter plugs and the usual chaotic affection.
Was This Useful?
Share
Keep Reading
Make Alcreon Yours
Tune your feedFive quick questions, and the feed ranks what matters to you first.Or just get notified
The weekly Echo. Signal worth keeping in your inbox.
Every new piece, announced on X.
Read Next
See all
Playbook
The Retirement Email Isn't a Warning
Model retirements now arrive every few weeks; the config-eval-rehearsal loop turns each deprecation email from a fire drill into an afternoon swap.
Playbook
The Cheapest Model That Passes
OpenRouter lists 400 models behind one API. The fix for choosing isn't a better leaderboard, it's a four-step protocol that ends in a real eval.
Playbook
Cheap Models, Hard Tasks
Most agent workflows route every step to the frontier model by default. The bill scales with how chatty the agent gets, even when most steps don't need that brain.