🤚 The Open-Palm Illumination

Ladies, gentlemen, and institutional investors who haven’t slept since the S-1 dropped — SpaceX is going public. Not in the quiet, dignified way a company usually approaches the public markets, but in the manner of a man who merges his rocket company with his AI company, absorbs a social media platform for good measure, and then files for the largest IPO in human history.

The numbers, because you came here for the numbers: $75 billion raise. A target valuation of $1.75 to $2 trillion. Ticker symbol SPCX on the Nasdaq. Roadshow begins June 4th, pricing June 11th, first trading day June 12th, 2026. A 21-bank syndicate led by Morgan Stanley, Goldman Sachs, and JPMorgan Chase is involved, because when you’re raising the GDP of a mid-sized European country, you need a small army of bankers to carry the paperwork.

For context, Saudi Aramco’s 2019 IPO raised roughly $29 billion. SpaceX is aiming for 2.5 times that. The financial world has not seen anything like this before, and frankly, it looks slightly nauseous.

👐 The Two-Handed Reality Check

Here’s the part the roadshow brochure won’t lead with: the combined entity — which now includes SpaceX launch services, Starship development, Starlink, and the entire xAI infrastructure (Grok, the X platform, the Colossus supercomputer) — posted a $4.94 billion net loss in 2025. This from a company that was profitable in 2024, when standalone SpaceX posted $791 million in profit.

What happened between then and now? An all-stock merger with xAI completed in February 2026, absorbing what is essentially a compute-hungry AI company still in its “spend aggressively and figure out revenue later” phase. Revenue hit $18.67 billion for the combined entity, which is impressive until you remember it costs approximately $18.67 billion to operate a fleet of 7,000+ satellites, build a next-generation super-heavy rocket, and train frontier AI models simultaneously.

And then there’s the governance structure. Elon Musk holds approximately 42–43% of the equity but controls ~79% of voting rights through super-voting shares. At the midpoint valuation, his stake would be worth over $530 billion, making him not just the richest person on Earth but possibly the richest person in any plausible economic model. Investors are not buying a company so much as they are subscribing to Musk’s decision-making, which has historically been either visionary or chaotic, often within the same earnings call.

🌿 The Gentle Awakening

But Peter Diamandis didn’t spend his entire episode on rocket capitalism. He also covered two quieter stories that, in any other week, would have been the headlines.

First: OpenAI’s GPT-5.5, released April 23rd, has been quietly outperforming prediction markets. Polymarket — the crypto-native forecasting platform where people bet real money on real outcomes — had $1.6 million traded on the GPT-5.5 release date and nailed the timing. But the more interesting story is the model itself. GPT-5.5 scored 82.7% on Terminal-Bench 2.0 (agentic command-line workflows), a 13-point gap over Claude Opus 4.7. It matched its predecessor’s per-token latency despite being a larger model. The age of “bigger model means slower model” is, apparently, over.

Second, and this one deserves a moment of genuine respect: an OpenAI reasoning model autonomously solved an 80-year-old math problem. The unit distance problem, posed by Paul Erdős in 1946, asks how many pairs of points in a plane can sit exactly one unit apart. For eight decades, mathematicians believed square grid constructions were essentially optimal. The AI proved them wrong, discovering an entirely new family of constructions using algebraic number theory and higher-dimensional symmetries. Princeton mathematician Will Sawin refined the result, and Cambridge’s Timothy Gowers confirmed that “no previous AI-generated proof has come close” to meeting top-journal standards.

This is not the GPT-5 embarrassment from seven months ago, when it merely rediscovered solutions already in the literature and everyone politely pretended otherwise. This is novel mathematics. The machine didn’t just pass the test — it wrote a new one.

👑 The Crown Verdict

What we are witnessing, distilled to its essence, is the week the future stopped being theoretical. A private rocket-and-AI conglomerate is about to become the most valuable public company on Earth. An AI model is outperforming human prediction markets at predicting itself. And another AI has produced original mathematical work that the field’s most distinguished minds have verified as genuine.

The SpaceX IPO will dominate the headlines — it should, at $75 billion — but the math result is the one that will matter in fifty years. We have crossed from “AI as a tool” to “AI as a colleague,” and the colleague just solved something 80 years of human mathematicians couldn’t. The appropriate response is neither panic nor celebration. It is a quiet, contemplative sip of something expensive, followed by the realization that your retirement portfolio probably needs a chapter on existential risk hedging.

The future is not arriving. It has arrived, filed its S-1, and is currently on a roadshow.

Inspired by SpaceX’ $75B+ Historic IPO, GPT5.5 Outperforms Polymarket, AI Solves 80yr old math problem | EP #257 by Peter Diamandis.

Your portfolio is showing. Diversify wisely.