The Two-Tier Valuation That Survived a 40% Crash

Starlink generated $4.4 billion in operating profit in 2025. At a generous 15 times multiple — the kind a mature telecom might command — that business is worth roughly $66 billion. SpaceX went public in May at $1.77 trillion. The arithmetic is not complicated: Starlink, the company's only profitable division, accounts for less than 4% of the IPO price. The remaining 96% — more than $1.7 trillion in market value — rests on businesses that lose money. ^[1] What fills that gap is not earnings but architecture. The prospectus claimed a $28.5 trillion total addressable market, $26.5 trillion of it tied to AI — a figure the company called the largest actionable total addressable market in human history. ^[2][3] Goldman Sachs projected AI revenue surging to $322 billion by 2030. ^[4] Musk told the FCC that a million-satellite orbital data-center constellation was a first step toward becoming a Kardashev II-level civilization capable of harnessing the sun's full power. He predicted that within two to three years, the lowest-cost way to generate AI compute would be in space. The company applied for 100,000 Gen3 Starlink satellites — each weighing over two tons, each capable of one terabit-per-second downlinks — as the first tranche of a planned million-satellite Starmind orbital AI megaconstellation. ^[5] The prospectus itself declined to stand behind any of it.

Our initiatives to develop orbital AI compute and in-orbit, lunar, and interplanetary industrialization are in early stages, involve significant technical complexity and unproven technologies, and may not achieve commercial viability. — Space Exploration Technologies Corp.

In separate passages, the company acknowledged it has never demonstrated in-orbit refueling and warned it may not be able to develop or commercialize any of its strategic initiatives on the timelines it anticipates, or at all. ^[1] The AI division, formed by merging xAI into SpaceX, lost $6.4 billion. ^[4] SpaceX as a whole has lost $13 billion since 2023, and S&P projected it will remain cash-flow negative until 2030. ^[6][7] Musk told investors revenue would hit $1 trillion by 2030 or 2031. Every major bank that modeled it came in below half that — Goldman at $470 billion, Morgan Stanley at $330 billion. ^[8] The market priced the stock at roughly 100 times revenue anyway. It surged to $225.64 by mid-June, a $2.82 trillion market cap that briefly exceeded Amazon and Microsoft. ^[2][8] Then the floor gave way. On June 22, SpaceX announced a $25 billion debt raise — senior unsecured notes, layered onto a company already losing money. The stock fell 16.4% in a single session, erasing roughly $600 billion. ^[9] By June 23 it was down 32% from the peak. ^[7] It entered the Nasdaq-100 in early July and kept falling. ^[10] On July 16, Starship Flight 13 — the first launch since the IPO, the redesigned V3 vehicle — scrubbed on the pad with an engine failure. The shares dropped below the $135 IPO price, to $131.11. ^[11] From peak to trough, the decline was roughly 40% in one month. That looks like the market re-coupling the stock to the balance sheet. It was not. Even at $131, SpaceX still traded at roughly 80 times revenue and more than double Morningstar's $780 billion fair-value estimate. ^[2][9] The correction was real but partial — a gap that narrowed without closing. And the two-tier structure that produced the gap did not break. Evercore ISI maintained its Outperform rating through the entire decline, citing the company's near-monopoly access to orbit. ^[6] Ron Baron, SpaceX's largest outside investor with $2.7 billion committed across private rounds and the IPO, predicted the stock could return 10 to 30 times its post-IPO value over a decade, on a thesis that Starlink alone will generate $1 trillion in annual revenue. ^[12]

It's going to be the internet for the entire planet. — Ronald S. Baron

The pattern is now visible: price the profitable segment on fundamentals, price everything else on the story, and when the story wobbles, mark the stock down — but never all the way to what the assets earn. The two tiers survived a 40% crash. The next test is structural in a different sense. In August, 911.5 million insider shares — more than the entire IPO float — become tradable for the first time. More blocks unlock in September, October, and December. ^[10] If insiders sell heavily into the narrative they built, the two-tier system loses its last anchor: the presumption that the people who set the price believe it. If they hold, the gap between what SpaceX earns and what it is worth gets another lease — not because the numbers changed, but because the story did not have to.