SpaceX closed day one at $160.95—a $2.1 trillion market cap on $18.6 billion of trailing revenue and a $5 billion annual loss—while 70% of the institutional book went to Saudi and Kuwaiti sovereign wealth funds and long-only allocations that won't touch the bid for months. Musk crossed a trillion dollars on paper and controls 80% of the vote with 42% equity, but a billion of his shares are locked behind a Mars colony SpaceX's own S-1 calls "improbable"—he can borrow against them tax-free in the meantime. Nasdaq's new fast-entry rule forces SPCX into the Nasdaq-100 after 15 trading days, so passive flows hit before price discovery does.

SpaceX prices Thursday at $135 a share in a $75 billion raise—largest IPO ever—but the actual setup is the retail mechanics: $70 billion in retail orders chasing at most $22 billion in retail allocation, which means brokerages leave 70% of orders unfilled and that unmet demand becomes day-one buy pressure. Musk keeps 82% voting control via dual-class shares while the company lost $4.9 billion in 2025 and another $4.3 billion in Q1, and Nasdaq's new 15-day fast-track rule forces QQQ's $460 billion into the stock almost immediately after it starts trading Friday. Fidelity dropped its IPO minimum from $500K to $2K for this deal—the retail tranche is triple the usual size and the structure was designed to let passive flows and FOMO do the price support work institutions usually extract concessions for.

SpaceX set its IPO price at $135 a full week before pricing day and is paying its 23 underwriters roughly $500 million on a $75 billion raise—about 67 basis points, a fraction of typical IPO economics. The bankers took the discount anyway, and JPMorgan is simulcasting a retail roadshow to 2,500 wealthy clients across 90 branches Thursday, bypassing traditional book-building entirely. This isn't price discovery—it's a controlled distribution event engineered around a <5% float and more demand than supply, pricing June 11 and trading June 12 as SPCX.

Anthropic filed confidentially for its IPO Monday, beating OpenAI to the punch and locking in a $965 billion valuation that tripled in roughly three months while revenue went from $30 billion to $47 billion annualized—the multiple expanded faster than the fundamentals, and the confidential route means no one sees the actual loss figure until Anthropic's ready. Going first means Anthropic sets the comp the entire AI sector gets priced against, right as OpenAI is dealing with missed revenue targets and a Musk lawsuit. The DoD sanctions over Claude's refusal to run weapons targeting could cost billions in sales this year per Anthropic execs, and that risk is sitting inside an S-1 the public can't read yet.

The auto industry just lost a million buyers since 2020 and instead of dropping prices they're stretching loans to 84 months so the $757 monthly payment doesn't look as insane as it actually is. The average new car now costs $49,461 and takes 35 weeks of median household income to afford, but analysts are calling it a soft landing because they're comparing it to last year's disaster instead of this month's actual deterioration. Even the safest auto loan securities are starting to crack as delinquencies rise, which means the industry's main trick for faking affordability is about to get way more expensive right when buyers are already tapped out.

Samsung, SK Hynix, and Micron just became shareholders in Anthropic at a near-trillion-dollar valuation, which means the three companies that make the memory chips Claude runs on are now financially locked in to Claude winning the AI race. Anthropic raised what looks like sixty-five billion at a nine hundred sixty-five billion dollar valuation, but fifteen billion of that was recycled hyperscaler money from deals like Amazon's April commitment, so the actual new cash is closer to fifty billion. The wildest part is the valuation went up two-and-a-half times in three months while Anthropic pushed its profitability target back to 2028, and almost no one is pointing out that those two facts happening at the same time is completely insane.

The 30-year Treasury hit 5% nine days before Kevin Warsh was sworn in as Fed Chair at the White House, not the Fed building, with Trump literally demanding lower rates during the ceremony while claiming he wants total independence. The bond market just priced in political capture risk before Warsh has chaired a single meeting, and it's a bear steepener, meaning long rates are rising even as traders expect short-term cuts for political reasons. Warsh's play is probably the balance sheet, not rate cuts, because slashing into a 5% long bond during an oil shock would be an instant credibility collapse the market would punish in real time.

Elon Musk just filed to take SpaceX, Starlink, xAI, and X public together at a two trillion dollar valuation, and the math only works because Starlink made four point four billion in profit last year while xAI lost six point four billion. The S-1 buries a four point two seven billion dollar quarterly loss and gives Musk eighty-three percent voting control with mandatory arbitration and class-action waivers, meaning if you buy shares you get zero ability to vote, sue, or propose anything. Existing SpaceX investors are calling the xAI and X acquisitions bailouts, but nobody can see the exchange ratios or valuations because the filing doesn't disclose them.

A jury in Oakland just cleared OpenAI to go public at a trillion-dollar valuation after rejecting Elon Musk's lawsuit in under two hours, and now two rival AI companies might hit the market within weeks of each other. Musk wanted 134 billion dollars and tried to unwind OpenAI's entire for-profit shift, but the jury ruled he waited too long to sue—he knew about the restructuring back in 2017 but didn't file until 2024. The wild part nobody's talking about: OpenAI still burns cash despite 20 billion in revenue, Greg Brockman's stake is worth 30 billion dollars, and xAI could beat them to Nasdaq as soon as June 12.

SoftBank just reported the biggest profit in Japanese corporate history and basically all 32 billion dollars of it is an unrealized gain on OpenAI shares that can't actually be sold because the company is still private. They funded the whole thing with a 40 billion dollar bridge loan that expires in about 12 months, which means OpenAI's IPO isn't happening when Sam Altman feels like it anymore—it's happening on SoftBank's lenders' timeline or this entire structure starts to unravel. They sold off massive chunks of T-Mobile and NVIDIA to go all-in on a single private AI bet they're now considering using as margin collateral, and if that sounds like Archegos with extra steps, that's because it kind of is.