In a Tuesday regulatory filing, SpaceX exercised an option it set in April and agreed to buy Anysphere, the maker of the AI coding agent Cursor, for $60 billion in an all-stock deal expected to close in the third quarter [1]. The currency it is paying with is three trading days old. SpaceX listed on the Nasdaq the prior Friday; by Tuesday the same shares were buying a coding startup [1].

That sequencing is the whole story. The paper's June 15 reporting on how the Google GPU contract put compute delivery and customer concentration on the IPO clock treated SpaceX's new equity as a record-setting tape. Tuesday turned that tape into acquisition currency. The first major use of the stock minted Friday is a land grab against Anthropic's Claude Code and OpenAI's Codex—the two market leaders Cursor was built to rival [1].

It is also a deal that runs in a circle, which is what X seized on. SpaceX is paying entirely in stock and will not use IPO proceeds [1]. Cursor's tools have run model-agnostically on rival frontier systems, including the Anthropic and OpenAI models SpaceX now competes with [1]. And the workload joins the same Memphis Colossus compute that SpaceX already rents to Google and Anthropic for roughly $26 billion a year [1]. The companion report on the coding-agent market now priced by the token traces what that consolidation does to developers. This piece is the balance sheet underneath it.

Start with the price against the business. Cursor, founded in 2022, generates roughly $2.6 billion in annualized business-to-business revenue [1]. A $60 billion all-stock price is about 23 times that revenue—rich, but a long way from the triple-digit multiple SpaceX itself trades at, and a price fixed back in April, not struck at Tuesday's frenzy [1]. The option's terms were a pre-IPO contractual fact: buy for $60 billion, or pay $10 billion to walk away from the partnership [1][2]. Tuesday simply chose the larger number.

The choice is cheaper than it looks, and that is the part the deal's defenders emphasize. Because SpaceX trades at a towering valuation, a $60 billion purchase paid in stock surrenders only a small slice of the company. "One of the things that makes SpaceX so valuable is how valuable it is," the investor Bill Ackman wrote on X. "The Cursor acquisition costs materially less in dilution because of SpaceX's high valuation." [1] The logic is sound and slightly vertiginous: the more the market pays for SpaceX, the more SpaceX can buy with the same dilution. A three-day-old multiple is doing real corporate work.

The related-party structure is the part both the celebration and the skepticism skate past. Elon Musk merged xAI into SpaceX in February in a transaction he valued at $1.25 trillion [2]. The acquirer, the compute landlord, the coding tool and the chatbot it will compete with are now inside one company. SpaceX said it will soon ship its own model on Cursor alongside Grok Build, xAI's coding agent [1]. Cursor's pull was never only its product; SpaceX's IPO filing noted that access to developers' coding requests and design decisions could improve models like Grok [1]. A tool that trains the buyer's models, runs on the buyer's compute, and competes with the buyer's tenants is not a normal bolt-on acquisition. It is a closed loop with an SEC filing wrapped around it.

There are guardrails on the downside. SpaceX owes a $10 billion termination fee if the deal collapses, falling to $4 billion if it fails on antitrust grounds [1]. Its leases to Google and Anthropic carry 90-day termination clauses, so it could reclaim that capacity for Cursor and Grok if usage climbs [1]. "If usage of Grok and Cursor picks up enough it can go back to using their capacity internally," D.A. Davidson's Gil Luria noted, "but it appears that they will provide capacity to Anthropic and Google for the foreseeable future." [1]

The market did not flinch. SpaceX shares jumped about 10% in early trading, adding roughly $247 billion to a $2.53 trillion valuation and putting the company on track to pass Amazon as the fifth-most-valuable in the world; at $211.27 the stock had climbed more than 56% above its $135 IPO price [1]. By the close it held near $202 [3]. The mainstream coverage reads the move as Musk buying an edge in the one corner of AI already producing real enterprise revenue. X reads it as a man paying himself with paper the market printed last week. Both are describing the same 8-K. The unusual thing is not that the readings differ. It is that, this time, they may both be right.

-- THEO KAPLAN, San Francisco