SpaceX: A $28.5 Trillion New Frontier?

SpaceX doesn't compete in the launch market so much as define it. By landing and reflying rockets, it has driven the cost of reaching orbit down dramatically, and in 2025 it completed 165 orbital launches, roughly 51% of the entire world's total and about 85% of all US launches, according to BryceTech data. That matters far beyond SpaceX itself. Cheaper, reliable access to orbit is the foundation the entire modern space economy is built on, and the reason satellite constellations, lunar missions, defence networks and commercial space ventures are now viable at all. SpaceX effectively owns that foundation, and the next leap could reset it entirely. Starship is designed to be fully reusable and to carry far heavier payloads, which SpaceX believes could eventually cut the cost of reaching orbit to a fraction of today's levels. If it works as intended, it would do for space what container shipping did for global trade, making it so cheap to move mass into orbit that whole new industries, from satellite mega-constellations to manufacturing and data centres in space, become possible. That is the prize the launch business is reaching for.

Starlink is the part of the story already generating serious cash. Its Connectivity revenue climbed to over $11bn in 2025, from under $4bn two years earlier, and turned an operating profit. It beams broadband anywhere on Earth, reaching homes, ships, planes, militaries and remote regions that ground-based networks will never economically serve. SpaceX's own progress report says Starlink added more than 4.6 million customers and expanded into 35 new markets in 2025 alone, and it has since passed 12 million subscribers. A newer satellite-to-mobile service is now erasing phone "dead zones." This is recurring, global, subscription-style income that scales as launch costs fall.

SpaceX owns the trucks, the roads and the product being delivered. It builds the rockets, runs the launch sites, makes the satellites and sells the connectivity, with each part making the next cheaper. Cheaper launch is why Starlink is economic and a growing Starlink is why launch cadence keeps rising. With the 2026 acquisition of xAI, it extended that logic into artificial intelligence, owning the compute, power and models rather than renting them. That infrastructure is already generating returns: Anthropic, one of xAI's biggest rivals, is paying $1.25bn a month to lease spare capacity at SpaceX's Memphis data centre, suggesting the AI build-out could eventually fund itself. Whether the broader AI bet pays off is unproven, but the integrated model itself, controlling an entire supply chain from factory to orbit, is something no rival has come close to replicating.

• Starship test flights: Each launch is a near-term signpost. A clean lift-off and booster recovery rebuilds confidence in the vehicle underpinning the whole growth case, while a setback sets it back, as the FAA's investigation into the May 2026 V3 "mishap" showed.
  
  

• Starlink subscriber momentum: Continued additions, particularly in higher-margin enterprise and government lines, would confirm the cash engine keeps accelerating and that demand is broadening beyond consumer broadband.

• Satellite-to-mobile and Starship-enabled capacity: Expanding the direct-to-phone constellation across more countries could turn ordinary handsets into satellite devices, while Starship is built to deploy far more powerful next-generation Starlink satellites, strengthening the launch and connectivity cases at once.
  
  

• Defence and government contracts: SpaceX is anticipated to win around 60% of US national-security launch missions under the Space Force's NSSL Phase 3 Lane 2 programme, a share worth roughly $5.9bn across 28 missions through 2029, deepening a high-quality, strategically sticky revenue base.

• Routine, low-cost Starship reusability: If Starship reaches frequent, cheap reuse, it slashes the cost of putting anything in space, which feeds directly back into Starlink margins and opens entirely new markets. It would also cement SpaceX as the logistics-and-connectivity backbone for satellites, lunar activity and future commercial habitats, with NASA's Artemis programme already giving it an anchor role in returning astronauts to the Moon.
  
  

• Frontier AI and orbital compute optionality: Success in advanced AI, including the more speculative goal of building compute in space, alongside the lunar and Mars ambitions, could redefine the company's earnings base entirely. This is upside optionality rather than something today's investor should bank on, but it is the part of the story that could justify the valuation if it lands.

In 2026 SpaceX acquired xAI, Elon Musk's artificial intelligence company, and placed one of the biggest bets in corporate history. The logic is straightforward: own the computers, the power and the models, and you control the infrastructure that the rest of the AI industry depends on. It is an audacious extension of the same vertical integration that made SpaceX dominant in launch and broadband. But unlike those businesses, where the revenues are real and growing, xAI is still largely a promise. Grok, its flagship chatbot, has made genuine strides and xAI now counts Anthropic among its data centre customers, paying $1.25bn a month to lease spare capacity. Yet Grok still trails ChatGPT and Gemini in market share, the AI landscape is shifting fast, and there is no guarantee that the model SpaceX is betting on will be among the winners. The orbital data centre ambition, putting AI compute in space powered by solar energy, is visionary but as yet unproven. Whether xAI becomes the backbone of a new AI economy or an expensive distraction is a central question to the thesis.

At its targeted ~$1.75 trillion market value, the company would trade near 90 times its 2025 revenue, a multiple that demands flawless execution. For context, Nvidia, one of the most profitable technology companies on Earth, trades at around 20 times revenue; SpaceX is asking roughly four times that while still posting losses. Morningstar pegs fair value closer to $780bn, less than half the target, arguing most of the premium rests on unproven AI ambitions rather than the businesses earning money today. If sentiment cools, investors could be exposed well before the long-term story plays out.

SpaceX is structured so that Elon Musk retains decisive voting control, reportedly around 82% of the vote, through a dual-class share system, and it qualifies as a "controlled company" exempt from some governance protections. Public shareholders get economic exposure but little say. The same individual also runs Tesla and other ventures, so key-person risk is acute: the strategy, brand and investor faith are tightly bound to one figure whose attention is divided. Leadership concentration this extreme is rare among companies of this scale.