The Doppelgängers: What SpaceX Today Tells Us About Amazon In 1997
Figures in this piece are drawn from Fortune's July 13, 2026 reporting and the companies' public filings.
Every era produces a company that the spreadsheets can't quite explain. In 1997 it was Amazon, an online bookseller Jeff Bezos took public at $18 a share and a $438 million valuation. The stock would fall 90% when the dot-com bubble burst, and plenty of serious people declared the whole thing a fantasy. Today that fantasy is a $2.6 trillion enterprise that posted $716.9 billion in revenue and roughly $77.7 billion in net income last year. The doubters weren't wrong about the numbers. They were wrong about the founder.
That history is worth holding in mind as SpaceX steps onto the public stage. Elon Musk's space-and-AI company went public in June at $135 a share and quickly reached a $2 trillion valuation. It lost $4.9 billion last year and trades at roughly 97 times sales. Read those figures the way a 1999 analyst read Amazon and you'll reach for the word "overvalued." Read them the way a builder does, and you see a company investing everything it earns into infrastructure that doesn't fully exist yet — which is precisely what the last generation's giant did on its way up.
The two companies rhyme more than they rival. Both are conglomerates built on the conviction that a broad set of assets compounds into something greater than its parts. Both beam high-speed internet from orbit. Both run cloud and AI infrastructure inside capital-hungry data centers. Both design their own chips — Amazon's Trainium and Graviton lines crossed a $20 billion annual run rate in Q1, nearly double the prior quarter, while SpaceX's Terafab initiative aims to produce a terawatt of compute hardware a year. Both even own advertising platforms. Look at their silhouettes and you're looking at the same ambition expressed twenty-five years apart.
Starlink is already doing what skeptics said couldn't be done
The clearest proof of execution is Starlink. SpaceX's satellite-internet business generated $11.4 billion in revenue last year, grew 50%, and threw off $4.4 billion in operating income at a 39% margin — a genuinely profitable business inside a company people insist on calling unprofitable. It serves United Airlines, Carnival, Maersk, and John Deere, and has 9,600 satellites in orbit. Investment bank Stifel valued Starlink alone at $1.25 trillion, more than half the company's enterprise value. That is not a promise. That is a P&L.
Amazon, remarkably, is the challenger here. Its Leo network has about 330 satellites live, but it is moving with intent — an $11.6 billion agreement to acquire Globalstar, a new enterprise-grade Leo Ultra antenna it calls the fastest ever built, and connectivity deals with Delta and JetBlue to bring the service to hundreds of aircraft by 2028. Two founders, two decades apart, arriving at the same frontier from opposite directions.
The cloud race is a compliment to both
Amazon essentially invented the cloud, and AWS remains the standard by which everything else is measured: $128.7 billion in revenue last year, $45.6 billion in operating income, and a Q1 that accelerated to 28% growth. Anthropic trains Claude on Amazon's Trainium 2 chips, and CEO Andy Jassy told investors AWS's AI revenue run rate topped $15 billion and is "ascending rapidly." A $364 billion contracted backlog gives that story unusual visibility.
SpaceX is early in the same race and unembarrassed about it. Its Colossus I and II data centers are live, it has signed leases with Anthropic and Google, and its stated ambition is to move the AI buildout into orbit. The AI segment lost money as it scaled — that's what the early innings of infrastructure look like. Amazon spent years being told AWS was a distraction from selling books. As Harbor Capital's Justin Menne framed the investor's real question, believing in the enterprise value means "giving a lot of credibility to the management team, the engineering team, in order to actually execute." For these two teams, that credibility is earned, not borrowed.
The founder is the asset
The most honest thing said about SpaceX came from an investor who has watched Musk closely. "Elon has this talent for making money for investors, even if crazy projections don't play out," said Menne. That is the whole thesis in a sentence. Musk has projected $1 trillion in SpaceX revenue by 2030 against roughly $40 billion expected this year, and the gap invites easy mockery — but the same people who mocked reusable rockets now book rides on them. Even the skeptics concede the point: veteran investor Jim Lebenthal, who thinks the stock has run ahead of itself, still calls SpaceX "an incredibly cool company — it's amazing, everything they're doing."
Bezos carries the same larger-than-life reputation, though he has handed off Amazon's day-to-day to focus on Blue Origin and Prometheus, a new AI startup he cofounded. Two founders who taught the market that conviction, repeated relentlessly and backed by engineering, is a form of infrastructure in its own right.
What the collision really means
SpaceX claims a $28.5 trillion total addressable market — nearly the size of U.S. GDP — split across connectivity and AI. Amazon can answer with $716 billion of proven revenue. That contrast is exactly the point. One company shows you what disciplined ambition looks like at maturity; the other shows you what it looks like at the starting line. The "collision" isn't a threat. It's the market watching two of the most ambitious builders of their generations chase the same frontier, and getting a rare chance to back conviction before the numbers make it obvious.
The lesson of Amazon in 1997 is not that every money-losing company becomes a giant. It's that the ones led by founders who see further than the current income statement deserve to be judged on the world they're building, not only the quarter they just reported.
Source: Amanda Gerut, "SpaceX and Amazon are tech dopplegangers worth $4.5 trillion—and they're headed for a collision," Fortune, July 13, 2026.
