SpaceX and OpenAI File for Trillion-Dollar IPOs, Oil Sinks on an Iran Truce, and a New Hawk Takes the Fed

Welcome to the first issue of The Memo, our new weekly briefing on the week in finance and the industry around it.

The idea is simple. Markets and the business of finance throw off an enormous amount of noise every week, and most of it does not matter. The Memo does the filtering: each week we pick the five to seven stories that genuinely moved the needle and explain them properly, with the figures, the context and the takes that help you understand what actually happened, not just that it did.

It is built for three kinds of reader at once: people already in the industry who want a fast, intelligent catch-up; people trying to break in who need to sound informed and see how the pieces fit; and anyone who simply follows finance and wants the signal without the firehose. Every item links out to its original sources and, where it helps, to one of our in-depth guides if you want to go deeper on the underlying concept.

Across issues we cover the full sweep of the industry: M&A and deals, IPOs and capital markets, private equity and private credit, the macro and markets picture, careers and recruiting, how AI is reshaping the work, the geopolitics that moves prices, restructuring and distress, sector trends, and the occasional landscape-shifting people move. Now and then we will also share what we are building here, under Inside IB IQ, which is where this note lives.

The Memo lands every Monday, so you start the week already up to speed. This is No. 1, so it will keep evolving; tell us what you want more of. For now, on with the week.

SpaceX filed its S-1 publicly on May 20, weeks after a confidential April 1 submission, setting in motion what would be the largest IPO in history. The company is seeking to raise up to $75 billion at a valuation of at least $1.75 trillion, trimmed from the above-$2 trillion target it floated in April (a cut Musk has publicly disputed, and one that could still move with investor demand). Formal marketing is expected to begin as soon as June 4, with pricing as early as June 11 and a Nasdaq debut targeted for June 12, under the ticker SPCX. Goldman Sachs won the coveted lead-left slot, with Morgan Stanley second and Bank of America, Citigroup and JPMorgan filling out the top of a syndicate of more than 20 banks, none of them European.

As expected, Musk keeps the keys. The offering uses a dual-class structure: public Class A shares carry one vote each, while insider Class B shares carry ten, leaving Musk with about 85% of the voting power and SpaceX classified as a “controlled company” under Nasdaq rules, which exempts it from several governance requirements. Public investors are buying the economics, not a say in how the company is run.

The more revealing detail sits in the use-of-proceeds section. Of the expected haul, about $62.8 billion, the bulk of the proceeds, is already spoken for, pre-pledged to third parties including early backer Valor Equity, creditors tied to Musk’s X and xAI, and EchoStar for a spectrum purchase. A separate disclosure flags a $20 billion bridge loan that must be repaid within six months of the listing. Strip that out and less than $18 billion of fresh capital is actually left to grow the business. For all the talk of funding the next chapter, much of this record raise is a debt refinancing.

The numbers underneath are a study in contrasts. SpaceX booked $18.7 billion of revenue in 2025, with Starlink now around 61% of the top line and the only consistently profitable unit. That base is set to jump, though by how much is contested. In a deal disclosed in the filing, Anthropic agreed to pay SpaceX $1.25 billion a month for access to the Colossus and Colossus II AI clusters and their 220,000-plus GPUs. The S-1 frames the fee as running through May 2029, which Bloomberg put at nearly $45 billion over roughly three years, about $15 billion a year, enough on its own to almost double SpaceX’s current revenue. But Musk has since publicly called it a 180-day lease with 90-day mutual cancellation, a maximum of roughly $7.5 billion, and said the short term “was our request.” That $37.5 billion gap between the filing’s implied commitment and Musk’s description is a live disclosure question as the roadshow opens. Yet the group still posted a $4.28 billion net loss in the first quarter of 2026 and carries an accumulated deficit north of $41 billion, much of it tied to the xAI ambitions Musk has folded into the story. The pitch is no longer just rockets; it is rockets, satellite internet, and a capital-hungry AI bet, increasingly funded by selling compute to other AI labs.

Not everyone is buying it. New Constructs’ David Trainer called the valuation “mathematically indefensible” and told Fortune his firm recommends investors avoid the deal, warning that “in order to compete with the hyperscalers, SpaceX will have to spend like the hyperscalers,” and could end up “diluting investors with additional capital raises in short time.” Prediction markets are less bearish: traders put a 98% chance on SpaceX closing above a $1 trillion market cap, and roughly 72% on clearing $2 trillion.

Either way, a deal this size bends the whole market around it. It rewrites the equity capital markets league tables in a single print, hands Goldman a marquee mandate, and pulls a wall of institutional money toward one name, leaving the rest of a crowded 2026 IPO pipeline to fight for whatever attention and capital is left.

After a month-long US-Israel war on Iran that shut the Strait of Hormuz, Washington and Tehran have reached a tentative 60-day memorandum to extend their ceasefire and begin reopening the waterway. The draft terms are sweeping: shipping through the strait would become “unrestricted,” with no tolls and no harassment; Iran would have 30 days to clear all mines; the US naval blockade would lift in proportion to the return of commercial traffic; and Washington would waive a range of sanctions, letting Iran sell oil freely again. In parallel, the two sides would open talks on Iran’s uranium enrichment. The catch: President Trump has not signed off, and Iranian state media insists nothing is final. Even as negotiators finalized the draft, the fighting had not fully stopped, with Iran firing ballistic missiles at Kuwait and sending drones toward the strait, a reminder of how fragile the arrangement still is.

Markets did not wait for the ink. Brent crude ended the month at about $92.50 a barrel on Friday (May 29), the final trading day of May, capping a fall of roughly 19% over the month, its steepest monthly drop since 2020. Brent had spiked above $120 at the height of the fighting in April, so the retreat into the low $90s marks one of the sharpest reversals in years. The logic is simple: a reopened Hormuz, through which a large share of the world’s seaborne crude passes, would unwind the war-risk premium that had gripped the market.

Energy strategists are far less sure the all-clear has sounded. “Overall this is a victory for Iran, which endured over a month of bombing, closed the strait, kept it closed and only agreed to reopen it via negotiated settlement,” said Gregory Brew of the Eurasia Group. Rapidan Energy’s Bob McNally was blunter on Tehran’s leverage, warning the regime is “ready to eat grass for six months to keep their chokehold on this jugular to wait for those oil prices to go even higher.” Even if the deal holds, the physical recovery would be slow: mines must be cleared, damaged infrastructure repaired, shut-in production restarted and a backlog of delayed tankers worked through.

For finance, the episode is a live case study in how fast geopolitics reprices everything. A double-digit swing in crude in a matter of weeks feeds straight into energy issuers’ financing costs, the timing of oil-and-gas M&A and IPOs, inflation expectations and, by extension, the Fed’s next move. It is also a reminder that deal windows in energy and infrastructure can open and slam shut on a single headline, which is exactly why bankers spend so much time pricing risk that has nothing to do with a company’s fundamentals. Geopolitical risk, quiet for a couple of years, is firmly back on the term sheet.

Kevin Warsh was confirmed on May 13 in a bruising 54-45 Senate vote, the most divisive in the Fed’s history, and sworn in on May 22 as the 17th chair of the Federal Reserve, succeeding Jerome Powell. He inherits a central bank that has held its policy rate at 3.50% to 3.75% through its first three meetings of 2026, and he chairs his own first meeting on June 17.

The market’s reaction has been the opposite of what many expected. Warsh was read as a pick who might usher in cuts, yet futures have drifted decidedly hawkish, now pricing roughly a 40% chance of a hike by December rather than the easing investors had penciled in. The April FOMC minutes had already shown officials leaning toward hikes over cuts, citing inflation still running above target and a resilient labor market.

Warsh himself is harder to pin down. Analysts who watched his confirmation hearings described a more nuanced figure than his hawkish first stint on the Board of Governors suggested, open to the idea that AI-driven productivity, tariffs and oil shocks complicate the old inflation playbook. What reads as hawkish is less his rate talk than his appetite for a radical reduction of the Fed’s balance sheet. Bloomberg’s opinion desk went further, arguing the new chair is “doomed to break with his campaign pitch” as expectations collide with reality.

For June, the consensus is that the Fed holds but strips the easing bias from its statement, a quiet but definitive hawkish signal. April PCE inflation offered little comfort, printing 0.4% headline and 0.2% core, soft on the month but not the decisive cooling markets wanted. A hawkish Fed under a brand-new chair reshapes everything downstream: higher-for-longer rates lift financing costs, compress the multiples buyers can justify, and change the math on every leveraged deal and IPO in the pipeline, including the wave of AI-infrastructure borrowing now running through private credit. The Warsh era starts in earnest on June 17, and the market will read his first statement word by word.

Five days after SpaceX, OpenAI confidentially filed its own S-1 with the SEC on Friday, May 22, lining up the second trillion-dollar listing in a single week. The ChatGPT maker is targeting a public debut as early as September, within a Q4 2026 window, at a valuation reported between $852 billion and $1 trillion. That timeline is contested inside the company: CFO Sarah Friar has reportedly argued for waiting until 2027, warning OpenAI is not yet ready for the audited financials and Sarbanes-Oxley-style controls a public listing demands, while Sam Altman favors moving in 2026. As with SpaceX, Goldman Sachs and Morgan Stanley are leading the deal, handing the same two banks the two biggest mandates on the Street at the same time.

The timing is not a coincidence. The filing came just four days after a federal jury in Oakland dismissed Elon Musk’s lawsuit against OpenAI and Sam Altman on May 18, a suit that had sought to unwind the company’s 2025 conversion to a for-profit structure, the very change that makes a public listing possible. Musk has vowed to appeal, but the verdict cleared the most serious legal cloud hanging over the offering.

What the filing cannot hide forever is the burn. OpenAI is losing roughly $1.22 for every $1 of revenue, an operating margin near -122% in the first quarter of 2026, and on internal projections is on course for a $14 billion loss this year even as revenue runs near $2 billion a month. It last raised at an $852 billion valuation in March. Because the submission is confidential, the detailed financials stay sealed until about 15 days before the roadshow, so the market is being asked to anchor on a near-trillion-dollar price tag well before it sees the full accounts.

The optics are hard to miss. Two of the most valuable private companies on earth, both deeply unprofitable, both betting everything on AI, are racing to tap public markets in the same window, what one analyst called the “trillion-dollar IPO test.” Bulls argue these are generational franchises that public investors have been locked out of for too long; skeptics counter that the timing has as much to do with insiders seeking liquidity and with funding an AI arms race that private markets can no longer bankroll alone.

For Wall Street it is a windfall and a stress test at once. Two megadeals from the same lead banks reshuffle the league tables, soak up a huge share of institutional appetite, and will show whether the 2026 IPO window can absorb two offerings this size without buckling. And whenever OpenAI’s S-1 finally opens to the public, it will be the closest look yet at the unit economics underneath the entire AI boom.

On May 19, Blackstone and Google unveiled a joint venture to build a new US company that will sell data-center capacity and access to Google’s TPU chips, the silicon that powers Gemini, as a compute-as-a-service platform. Blackstone is committing an initial $5 billion of equity from its funds, but including debt the venture is sized at roughly $25 billion, and it expects to bring its first 500 megawatts of capacity online in 2027, scaling from there. Longtime Google infrastructure executive Benjamin Treynor Sloss will run it.

The deal is a clean snapshot of the year’s defining trend: private capital pouring into the physical backbone of AI. Increasingly the largest checks for data centers, power and chips are written by private equity and private credit rather than public markets, and the firms once seen as displacing the banks are now partnering directly with the tech giants they finance.

Blackstone framed it as a once-in-a-generation opening. “We see a generational opportunity to invest capital at scale building AI infrastructure,” said president and chief operating officer Jon Gray, pointing to the “unprecedented demand for compute” and pairing Google’s “world-class TPUs” with “Blackstone’s exceptional strength in energy and digital infrastructure.” The structure matters too: by offering TPUs outside Google Cloud, the venture hands AI developers another route to Google’s chips and hands Google a capital partner to fund the buildout without carrying all of it on its own balance sheet.

For anyone tracking where finance is heading, this is the center of gravity. In the same week that SpaceX revealed it is selling $1.25 billion a month of compute to Anthropic, Blackstone planted a $25 billion flag in the same ground. Energy, data centers and chips are absorbing capital at a scale that is redrawing where the most interesting deal, financing and infrastructure work happens, and pulling the industry’s best toward it. The risk, as in any boom, is that everyone is underwriting the same demand curve at once.