Quick Take
  • The same week SpaceX took its IPO to Wall Street, its biggest AI client showed up at the same door.
  • SpaceX launched its IPO roadshow this week, targeting $75 billion, the largest public offering in history, with a June 12 Nasdaq debut under ticker SPCX.
  • On the same day, Anthropic, which pays SpaceX $1.25 billion a month for AI compute, filed its own confidential IPO prospectus with the SEC.
  • But the filing contains a clause that changes how investors should read that revenue.

What Happened

SpaceX launched its IPO roadshow this week, targeting $75 billion, the largest public offering in history, with a June 12 Nasdaq debut under ticker SPCX. On the same day, Anthropic, which pays SpaceX $1.25 billion a month for AI compute, filed its own confidential IPO prospectus with the SEC.

SpaceX’s amended IPO filing discloses the full Anthropic deal: 325,000 Nvidia GPUs across its Colossus and Colossus II facilities in Memphis, $1.25 billion a month through May 2029. But the filing contains a clause that changes how investors should read that revenue.

The same IPO that lists Musk’s aerospace company on June 12 carries a clause that lets its most important client walk away in 90 days. Whether investors read that as flexibility or fragility will set the price.

Market Context

The two companies, which compete directly in AI through xAI’s Grok and Anthropic’s Claude, are now officially racing each other for the same institutional capital.

Two IPOs, One Capital Pool

Anthropic’s IPO, targeting a $965 billion valuation, will follow SpaceX and OpenAI into a market already absorbing more capital than any IPO cycle in recent history.

Why It Matters

For a company where Anthropic’s payments approach its entire annual revenue, that termination clause is the most important sentence in the filing. As BeInCrypto’s earlier analysis of the SpaceX-Anthropic relationship noted, the financial entanglement between the two was always the central risk in SpaceX’s IPO narrative.

Institutions choosing between the two face a direct tradeoff: SpaceX’s valuation depends partly on Anthropic’s compute payments, and Anthropic is building its own infrastructure that could make those payments unnecessary.

By May, he had leased them the world’s largest AI supercomputer and posted a full reversal on the same platform: “Everyone I met was highly competent and cared a great deal about doing the right thing. No one set off my evil detector.”

Details

The same week SpaceX took its IPO to Wall Street, its biggest AI client showed up at the same door.

The $15 Billion Contract With a Six-Month Exit

After an initial three-month period, either party can terminate the agreements with 90 days’ notice. What reads as up to $45 billion in contracted future revenue is actually an arrangement either side can exit within roughly three months.

Both companies now compete for institutional allocations in the same window. Goldman Sachs holds the lead position on SpaceX’s offering. Morgan Stanley manages a direct share program that reserves 5% of IPO stock for insiders selected “based on the discretion of our executive officers,” with no lockup restriction, meaning those participants can sell from day one.

Elon Musk’s ‘Evil Detector’

Three months before signing Anthropic to a $15 billion annual compute deal, Musk was doing the opposite. In March, he called Anthropic “evil” and “misanthropic” on X, retweeted a post urging Anthropic employees to quit, and asked: “Is there a more hypocritical company than Anthropic?”

The shift from rival to landlord took roughly 90 days, which happens to be the same notice period that lets either side walk away from the deal.

The post SpaceX’s Biggest Customer Is Also Its Biggest IPO Rival Paying $15 Billion a Year appeared first on BeInCrypto.