On June 10, 2026, Elon Musk posted on X, replying to SpaceX and Sawyer Merritt, that "1M tons to orbit should be possible in roughly 5 years." Benzinga reported the same statement. The post followed a June 8 technical update at SpaceX's Bastrop, Texas site, where Musk said SpaceX aims to raise Starship annual upmass from about 2,500 tons to about 1 million tons within roughly three years. 

At 100 tons per reusable flight, one million tons a year requires about 10,000 launches, close to 27 per day, against a 2026 production target of roughly ten vehicle sets. The statement is a target Musk set rather than capacity SpaceX can sell today.

Bloomberg reported on June 8, 2026 that the banks running the SpaceX IPO told investors the offering was well oversubscribed, with several institutions each ordering near $10 billion. The banks will close institutional order books after the New York close on Wednesday, June 11, then price that evening for a June 12 Nasdaq debut under ticker SPCX. SpaceX is selling 555.6 million shares at a fixed $135 to raise about $75 billion, which the underwriters peg at a valuation near $1.8 trillion. President Gwynne Shotwell and CFO Bret Johnsen worked roughly 300 investors at Morgan Stanley's New York offices during the roadshow. A day later, on June 9, Reuters cited a source putting demand above $250 billion, close to four times the offering.

With orders running past supply, the banks can steer shares toward long-only funds. 30% has been allocated to retail investors. SpaceX fixed the price at $135, so the underwriters cannot raise the clip to capture the surplus, and investors take tighter allocations instead of SpaceX banking a larger raise.

At a JPMorgan investor event, Elon Musk said SpaceX will build its planned orbital "AI satellites" to run third-party accelerators, so customers can fly Nvidia GPUs or Google TPUs of their choice. @niccruzpatane posted the quote from the event on X, and trade coverage of the appearance reported the same. Musk made the remark the week SpaceX disclosed its $920 million per month Google compute agreement and days before the June 12 listing. SpaceX has shown an "AI Sat Mini" concept rated near 100 kilowatts per satellite at earlier events. It has not flown orbital compute hardware or published a delivery date.

An accelerator-agnostic design would widen SpaceX's customer set past any single chip vendor, which feeds the AI-revenue case behind the bank forecasts in item 9. SpaceX has filed no design, no power-and-thermal spec, and no customer contract for orbital compute, so the claim works as roadshow positioning rather than a purchasable product.

SpaceX disclosed a Cloud Service Agreement with Google entered June 5, 2026, in a free-writing prospectus tied to its S-1, under which Google pays $920 million per month from October 2026 through June 2029 for access to compute capacity, including roughly 110,000 Nvidia GPUs plus CPUs, memory and related components. The committed payments total about $30 billion over the term. The filing states Google may terminate if SpaceX fails to deliver the committed GPU capacity by September 30, 2026 after a one-month grace period, or accept available GPUs at proportionally reduced payments; after December 31, 2026 either party may exit on 90 days' notice. CNBC and Bloomberg reported the terms on June 5.

The agreement books a data-center revenue line that several outlets project will rival SpaceX's combined Starlink, launch and prior AI revenue, which reframes the company the IPO is pricing. The September 30 delivery condition puts the full monthly run-rate at risk of SpaceX's own hardware buildout schedule, so the $30 billion headline is contingent, not contracted-firm. Disclosing a Google counterparty days before pricing hands underwriters a named, investment-grade customer to anchor the AI revenue assumptions in the roadshow.

Apex raised more than $200 million and roughly doubled its valuation to $2.3 billion, Bloomberg and SpaceNews reported on June 5, 2026. The four-year-old, Los Angeles manufacturer builds satellite buses and cited demand from commercial and government customers, including work tied to the Golden Dome missile defense effort. The round lands days after Northrop Grumman named Apex its bus partner for space-based interceptors.

Pairing a fresh $200 million balance sheet with the Northrop selection gives Apex the working capital to scale production against a 2027 demonstration deadline. The landmark raise also serves as an early indication of private capital flowing into satellite-bus manufacturing.

Planet Labs reported fiscal Q1 2027 revenue of $94 million, up 42% year over year, on June 4, 2026, alongside remaining performance obligations of $816 million, up 81%, and backlog above $906 million. The company disclosed a $21.9 million one-year contract extension from the National Geospatial-Intelligence Agency plus a new NGA award for its Global Monitoring Service. Defense and intelligence revenue grew more than 65% year over year. Non-GAAP gross margin was 56%, down from 59% a year earlier.

The backlog and RPO growth outpaced revenue, which signals demand booked ahead of recognized sales and gives forward visibility on the government side. The margin compression to 56% runs against that, indicating Planet is buying revenue growth partly through lower-margin data and services work. The $21.9 million NGA extension is small against backlog but confirms the agency relationship that anchors the defense and intelligence segment.

Sell-side research from Goldman Sachs and Morgan Stanley, surfaced June 4, 2026, projects SpaceX revenue near $160 billion in 2028, against approximately $18.7 billion in 2025. For 2030, Goldman models above $470 billion and Morgan Stanley about $330 billion, a roughly $140 billion gap the banks attribute to differing AI-compute growth assumptions. Morgan Stanley's longer pitch runs to $3.4 trillion of revenue by 2040. 

These are underwriter projections published during the marketing window for a deal those same banks are running, so they describe an incentive to value the float. The $140 billion divergence at 2030 sits almost entirely in the AI-compute line, which drives the majority of the $1.77 trillion valuation.