SpaceX’s $2 trillion reality check

SpaceX’s $2 trillion reality check

SpaceX’s upcoming IPO promises to be not only a watershed for the space industry, but a landmark in financial markets. The company—which includes the world’s leading space-launch business, its largest satellite broadband provider, and a money-burning AI division—aims to raise about $75 billion, at an estimated valuation of $1.75 to $2 trillion. 

Only a decade ago, “the concepts of reusable rockets and low-cost miniaturized satellites were perceived as a bit of a curiosity,” says James Bruegger, chief investment officer at space-focused Seraphim VC. “Now we’re on the cusp of the biggest IPO in history being a space business.”

The company’s valuation is tied to a grand vision of space’s future, rather than the fundamentals of SpaceX’s business today. In SpaceX’s IPO prospectus, the company states that its mission is nothing less than “to understand the true nature of the universe, and to extend the light of consciousness to the stars.” Its current space efforts, it claims, could one day “lead to the emergence of new trillion-dollar markets on the Moon, Mars, and beyond.”

But if you want to understand SpaceX’s business today—and what its IPO will mean for the broader space economy—you need to cut through the froth and grapple with the underlying economics, technological realities, and geopolitics that are actually shaping the space race. 

Launch capacity remains a massive bottleneck

The start of the private space era began 18 years ago, with the launch of SpaceX’s Falcon 1 rocket. Today just two entities outside of China and Russia are delivering payloads into orbit frequently and reliably: SpaceX and Rocket Lab. 

There is an order of magnitude difference between them. SpaceX completed 165 orbital launches in 2025—accounting for nearly 51% of total launches worldwide and sending aloft 85% of all satellites last year. Rocket Lab had 21 successful launches, including orbital and suborbital missions. With a maximum capacity of 13,000 kilograms, SpaceX’s Falcon 9 rocket makes Rocket Lab’s Electron, with a maximum payload of about 300 kilograms, look rather puny.

Both companies are transitioning to next-generation rockets—and both are behind schedule. Developed at a cost of more than $15 billion, SpaceX’s Starship is designed to deliver 100 metric tons to Earth orbit in its current fully reusable configuration, and double that in future versions. Starship is central to all of SpaceX’s future aspirations, “unlocking growth through our unique vertically integrated business model,” as the company says in its prospectus. 

Starship has made 12 R&D flights since 2023, six successful; its most recent launch, last week, was mostly successful. The company says it expects to start payload delivery to orbit in the second half of 2026. 

Meanwhile, the inaugural launch of Rocket Lab’s medium-lift Neutron, originally expected in 2024, has been pushed back to the end of 2026, at the earliest, after a first-stage propellant tank ruptured in January. 

The only alternative launch providers in the U.S. are United Launch Alliance (ULA) and Blue Origin, and both are also struggling to put new rockets into regular service. ULA’s Vulcan Centaur—the heavy-lift successor to its 20-year-old Atlas V—has been effectively grounded after three flights due to debris shedding from a booster during a February launch.

Blue Origin’s New Glenn put one NASA payload into orbit last year, but in April missed the orbital target for a satellite for AST SpaceMobile. The misplaced satellite had to be de-orbited.

These setbacks have created a launch backlog with real consequences. Amazon is nearing a July 30 deadline with the FCC to have half of its proposed LEO communications satellite constellation deployed. But though it’s contracted 38 Vulcan launches and 12 with New Glenn (along with 18 through Airbus subsidiary Arianespace and three with SpaceX), so far, it’s only been able to deploy about 10% of its planned 3,232-satellite constellation. The company is citing “a near-term shortage in launch capacity” in its appeal of the FCC deadline.

“In the age of AI and abundance of capital and all the rest of it, [launch] is still freaking hard,” says Rocket Lab founder and CEO Peter Beck. “It’s really rare to get all the things to come together to make it work. There’s no partial success. It’s success or fail.”

Bruegger is confident that additional competitors will emerge eventually. The market has “never been one-size-fits-all,” he says. “As more rockets of different sizes, classes, and capabilities come online, SpaceX isn’t going to have the market all to itself.” 

It’ll be hard, however, for any upstart to compete with SpaceX’s Starship on price: The company’s IPO prospectus states that fully loaded Starships could eventually reduce the historical average launch cost per kilogram of $18,500 by 99% or more.  

Orbital AI data centers won’t be online anytime soon

To justify a possible $2 trillion valuation, SpaceX is trying to sell investors on the idea of space-based AI infrastructure, including massive data centers orbiting the Earth. 

Presently, AI looks like the company’s Achilles’ heel. SpaceX’s xAI division—acquired in February—ran an operating loss of $6.35 billion in 2025, while consuming 61% of total capital expenditures. To bolster its AI offering, SpaceX in April secured an option to buy Cursor, the AI coding startup, for $60 billion; layoffs and staff departures have followed. 

Musk is banking that his idea to fuse his space company’s fortunes to AI will pay off exponentially—the SpaceX prospectus forecasts a total addressable market of $26.5 trillion for its AI initiatives. (In the shorter term, Musk is hedging his bets by selling unused compute capacity to Anthropic in a deal that could bring in $1.25 billion a month to xAI through May 2029.) 

There’s already a lot of data being collected in space—for intelligence, surveillance, reconnaissance, and other purposes. But we don’t have the capacity right now to beam that all back to Earth.

“If you could process it [in space] and just send down the interesting insights, that is the near-term capability to lean into,” says Taylor Sargent, a partner at Industrious VC. Companies are beginning to deploy on-orbit edge computing nodes, but they’re running into launch bottlenecks as they try to get their assets into space. That’s much less of a problem for SpaceX.

Building large-scale AI data centers in space is an altogether different proposition. Proponents argue that as it gets harder to find land, water, and energy for new AI data centers on Earth, space—with its abundant solar power and no neighbors—becomes more appealing. According to the SpaceX prospectus, space is “the only truly scalable solution to the challenge of accelerating demand for compute relative to terrestrial energy constraints.” 

The scale of SpaceX’s ambitions borders on the surreal. The company ultimately hopes to deploy 100 gigawatts of computing capacity in space each year—an effort that, its prospectus says, “will require thousands of launches per year and the transport of approximately one million metric tons to orbit annually.” The company filed plans with the FCC in January for an “orbital data center” with up to 1 million satellites to power advanced AI models and applications. 

SpaceX and its data-center competitors will need exponentially more launch capacity to pull off these massive structures. They’ll also have to figure out how to build in space—using robots and automation, presumably—and solve operational issues, such as heat dissipation and radiation hardening.

Space investors remain bullish, despite the daunting task. “These are all solvable problems, in our view,” says Bruegger at Seraphim. “I have a high degree of confidence that a time will come where a meaningful amount of AI compute is being done in orbit, by the back end of this decade. The step change will be these mega launch vehicles flying regularly.”

“It’s not a scientific or engineering question,” says Beck at Rocket Lab, which manufactures solar arrays that could power space computing. “If the cost of electricity on Earth drops a little bit, then it’s vastly superior to build [data centers] on Earth. If the cost of electricity remains the same, then it is possibly more economical to build them in orbit.”

The biggest SpaceX business is broadband

In 2025, SpaceX made nearly $4.1 billion from its space segment (at an operational loss of $657 million)—but almost $11.4 billion from its satellite broadband service, Starlink, which remains by far the company’s biggest revenue generator. 

SpaceX has more than 9,600 Starlink satellites already in orbit, accounting for roughly two-thirds of all active satellites circling Earth, and is planning for a total of 42,000. About three-quarters of SpaceX missions last year were for Starlink alone. According to the company’s prospectus, Starlink now has 10.3 million broadband subscribers in 164 countries, up from 8 million in November.

With Starlink, SpaceX disrupted terrestrial internet providers and the niche industry of satellite television. For its next phase of growth, it’s tackling mobile connectivity, which Bruegger calls “a bigger market by several orders of magnitude.” 

SpaceX’s vertical integration is a massive advantage. Through partnerships with more than 30 mobile network operators, SpaceX’s direct-to-device cellular service, Starlink Mobile, provides data, voice, and messaging services to approximately 7.4 million monthly unique devices in about 30 countries, eliminating coverage “dead zones.” Since January 2025, it’s grown its constellation of  V1 Mobile satellites from 360 to approximately 650. Starship will be able to deploy approximately 50 mobile satellites per launch, accelerating expansion.

The SpaceX prospectus claims a total addressable market of $1.6 trillion for Starlink, including $740 billion for Starlink Mobile. But there is growing competition in the satellite-to-smartphone race.

AST SpaceMobile now delivers direct-to-device cellular service using larger satellites that could let it scale quickly, and Amazon Leo’s $11.5 billion acquisition of Globalstar in April 2026 gives it licensed L-band and S-band spectrum, critical frequencies for connecting to phones on the ground.  

Defense remains the backbone of the space economy

Defense has been the earliest and biggest customer for commercial space. SpaceX’s first two Falcon 1 launches were purchased by the U.S. Department of Defense. The company now finds itself in a commanding position as space-defense budgets soar.

The Trump administration has requested $71 billion in total Space Force funding for FY 2027, a 77% increase over 2026, including $40 billion designated for R&D. SpaceX is the leading launch contractor for the Space Force’s program focused on high-priority military and intelligence payloads; last April, it secured a $5.9 billion award for 28 launches through 2029.

ULA was awarded nearly $5.4 billion for 19 launches of its Vulcan Centaur, and Blue Origin, $2.4 billion for 7 missions. SpaceX also stands to earn another $2 billion in satellite work tied to the Golden Dome defense system. 

Rocket Lab, meanwhile, has U.S. Space Development Agency contracts worth a combined $1.3 billion to design and build secure communications and missile-warning satellites, plus a $190 million hypersonic-test launch contract. Earth monitoring satellite companies such as IceEye and Planet Labs have also seen revenues surge on defense demand; their civilian business lines are marginal in comparison.

The SpaceX IPO will be a watershed—and a wakeup  

As the first major space industry exit, the SpaceX IPO is expected to catalyze the sector, releasing a wave of capital and experienced talent as early investors take their payouts and longtime employees cash in their equity and fan out through the broader space ecosystem.

Beck, of Rocket Lab, sees the IPO as “net positive” for the industry—and a potential boon to his own company’s market cap. “As kind of the number-two behind SpaceX, we will be the comp. I’d take some of their multiples,” he says. 

In the unlikely event that the SpaceX IPO is a flop, Beck says, “space will become less interesting for everybody, as well. But Elon doesn’t generally have a habit of shitting the bed, so I think everything will be fine.” 

The SpaceX IPO could also bring needed scrutiny to the space industry, says Beck. He describes the industry as a place where, traditionally, “a smart CEO can wave their arms and tell a story, and unless you’ve got a PhD in mathematics, you can’t prove or disprove it. That’s led to some large failures.” 

Public market analysts asking real questions will make it harder to sell the purely speculative. That’s good for the industry, even if it’s uncomfortable for some of the players in it. As head of a public company since 2021, Beck says he’s gotten used to answering to analysts. “It’s basically super-easy, my job: Tell the truth, do what you say you’re going to do, and hit your numbers.” 

Part of running a public space company, though, is making sure that if you don’t deliver something in the promised timeframe, the rest of the business doesn’t suffer. SpaceX is taking the opposite approach. A massive chunk of its proposed valuation rests on Starship and how quickly it can become a reliable ride for the company’s business arms.

As the company’s prospectus notes, any failure or delay in the rocket would hamper SpaceX’s “deployment of next-generation satellites, global satellite-to-mobile connectivity, and orbital AI compute.” Sky-high ambitions are well and good. But on-the-ground fundamentals are essential.