SpaceX vs. Anthropic: One Trillion-Dollar IPO Is Priced on Cash, the Other on a Promise

Two near-two-trillion-dollar companies are fighting for the same headline: the largest stock debut in market history. One has already taken the title; the other plans to snatch it. The headline is entertainment. The investor's question is whether either price is anchored to economics the company actually owns — or is merely a fee for joining a story.

So here is the card.

	SpaceX (SPCX)	Anthropic
Status	Traded since June 12, 2026	Pre-IPO, S-1 filed
Market cap / valuation	~$1.9 trillion	~$0.97–1.8 trillion
Its one claim	Largest IPO in history (already)	Largest IPO in history (planned)

Same starting line, same clock, one question: which of these two trillion-dollar books can compound into the price it is asking you to pay? Both open at 100 paper points on the shared valuation line of roughly $1.8–1.9 trillion. Because Anthropic cannot be traded yet, the live score here runs on the mechanism board — the economics behind the multiple — not on a price race this card cannot fairly run. Opening odds: SpaceX 55/45, on demonstrated cash flow. Those are editorial confidence odds, a handicap for debate, not a sportsbook line or a promise of return.

Forget who sounds smarter. The match turns on what counts.

SpaceX finished its debut day as the biggest IPO on record — priced at $135, raising $75 billion, closing at roughly $2.1 trillion, or about 112 times the prior year's revenue. The market has since been respectful rather than reverential: shares ran to an intraday $225 in June, gave most of that back, and trade near $142 in early September at a market cap around $1.9 trillion.

What does that price actually buy? A company that booked $18.67 billion of revenue in 2025, about 61% of it — $11.4 billion — from Starlink, the satellite internet service that had roughly 12 million subscribers by mid-2026 and grew nearly 50% from 2024. Starlink is the one part of the business that produces real, recurring cash flow, and it is why SpaceX still reports positive adjusted EBITDA of about $6.58 billion for 2025 even while booking a net loss of roughly $4.9 billion.

Now the other side of that ledger: the rocket-launch division is described as burning cash while it waits for the next-generation Starship to deliver better unit economics, and the AI business absorbed through the xAI acquisition is a near-term money pit. SpaceX has talked about up to $119 billion of capital expenditure for a data-center buildout. Its wider addressable-market pitch — $28.5 trillion across space, connectivity, and AI — drew a blunt dismissal from valuation professor Aswath Damodaran, who called the figure a "hallucination". Within three months of the debut, CFRA had a sell rating on the stock, with a $115 target.

The opponent's case: what the forecast is paying for.

Anthropic's numbers tell a different story because they are mostly future tense. The Claude developer confidentially filed its S-1 with the SEC on June 1, after raising $65 billion in a May round at a $965 billion post-money valuation. Its revenue run-rate went from roughly $9 billion at the end of 2025 to $47 billion by mid-May to $65 billion by the end of July, and growth the company says ran more than 10-fold a year in each of the three years through early 2026, and now ahead of OpenAI's reported $40 billion pace.

Here is the catch in that trajectory: Anthropic just recorded its first quarterly operating profit, about $559 million in the second quarter of 2026, against revenue of at least $10.9 billion for the same period. That is an operating margin in the single digits, and it comes only after a massive buildout of compute capacity — commitments to up to 5 gigawatts with Amazon, multiple gigawatts of TPU capacity with Google and Broadcom, and more than 300 megawatts with 220,000-plus Nvidia GPUs through SpaceX itself. The business is also strategically close to the hyperscalers who both fund it and lend it capacity, and it showed its regulatory surface in June when it suspended some models for foreign nationals after a U.S. government directive.

The handicap keeps the duel honest.

This is where a fair match gets its tension, because the cheap versus expensive reading runs opposite to the "real economics" reading.

SpaceX is the side with cash that exists — but at roughly 81 times trailing sales and more than 300 times adjusted EBITDA, it is priced like the cash already has to grow at Starship speed just to defend its multiple. Anthropic is the side with the seductive number — about 15 times its July run-rate, far cheaper-looking on paper — but that number is a forecast, and the profit supporting it is a rounding error beneath the revenue. Two American bankers are even underwriting against a 2028 revenue projection of $190 to $200 billion, meaning the price the IPO reportedly targets — around $1.75 to $1.8 trillion, or roughly 27 times the July run-rate — asks you to believe the current hypergrowth persists two more years without a misstep. A flagged risk in that diligence is an accounting restructure from gross to net revenue, which could shave a fifth to two-fifths off the headline number, though the forecasts assign that a low probability.

So the scoreboard splits: SpaceX leads on demonstrated cash flow and margin durability; Anthropic leads on headline growth and forward cheapness. Neither is a walk, and that is why the duel is worth running at all.

What to watch, not what to bet.

For a retail owner trying to understand the stakes, the shared, uncomfortable fact is that both of these are pay-later-for-tomorrow stocks, and both are currently outside the broad passive benchmarks — SpaceX is not even eligible for the S&P 500 because it reports a net loss. Public investors get founder-heavy governance either way: Elon Musk controls more than 82% of SpaceX's voting power, and Anthropic is a public benefit corporation with its own charter trade-offs between growth and safety.

The mechanism checkpoints are the ones that will separate the two. For SpaceX, do Starlink's margins hold and does Starship actually fly at the promised economics? For Anthropic, does its customer set repeat, does the 2028 revenue projection get underwritten with real renewals, and does a gross-to-net restatement ever surface in the public S-1?

The next checkpoint is Anthropic's listing, expected around late October. That debut is the first point where the paper contest can have a real return scoreboard at all. At the shared line, the mechanism board still edges toward the company whose trillion-dollar valuation is backed by cash that exists rather than revenue that is promised.

In one line for the final bell: one side is expensive for what it earns today; the other is priced for what it has yet to earn at all. The better "own," judged on demonstrated economics, is the expensive one that already ships cash — but at this altitude, neither valuation leaves you room for the story to stumble.