How Seriously Should Investors Take SpaceX’s AI Narrative

How Seriously Should Investors Take SpaceX's AI Narrative?

SpaceX lands rockets on drone ships, launches more payload to orbit than every other company combined, and runs the world's largest satellite internet network.

It is, by any measure, at this juncture a space company.

So why does its IPO filing describe 93% of its addressable market as AI, and why is it spending money on graphics chips faster than it ever spent on rockets? The answer is probably valuation. A proven launch business and a profitable satellite network are worth a lot, but they will likely be valued as a telecom or industrial company. An AI infrastructure company, in today's market, is worth more. SpaceX is heading toward a public listing that could value it at $1.75 trillion or more, roughly 100 times trailing revenue and around 300 times its 2025 adjusted EBITDA. To justify those numbers, it needs a bigger story than rockets.

It is a multiple that recalls the peak of the dot-com bubble, and the parallels are worth considering. SpaceX At 100x Revenue: A Warning From 2000

Goldman Sachs has called AI the biggest driver of SpaceX's future value, noting that revenue could 100x by 2030. ^[1] Whether that holds up depends on three very different businesses, each at a different stage of reality.

The AI Business Today

In early 2026, SpaceX absorbed xAI, the artificial intelligence startup Elon Musk founded in 2023, along with X, the social media platform formerly known as Twitter that xAI had acquired. Together they now sit inside SpaceX under a division called SpaceXAI. The consumer-facing products are Grok, a large language model competing with ChatGPT and Gemini, and X itself as the distribution platform. The problem is the competition. Google's Gemini is embedded across Search, Android, and Workspace by default. Anthropic has quietly won enterprise trust on the back of its safety reputation. Grok's only real differentiator is X's real-time data and Musk's profile, neither of which is obviously a business. The divisions are cash-intensive, and xAI spent $12.7 billion on AI infrastructure in 2025, with another $7.7 billion deployed in just the first quarter of 2026. This is the primary reason SpaceX reported an operating loss of $1.94 billion in the first quarter of 2026, against revenue of $4.69 billion.

To offset those costs in the near term, SpaceX is renting out the data center capacity it has been building. The IPO filing claims SpaceX operates the largest AI training clusters on Earth, built around facilities called Colossus and Colossus II. Anthropic is the anchor customer. It is paying SpaceX $1.25 billion per month for compute access through 2029, though this is not a long-term commitment, and either party can exit with 90 days' notice. Other companies are signing on. That said, this does not appear to be the long-term plan. It is a way to generate cash flow while the larger bet is assembled.

Orbital Compute Is What Investors Are Paying For

The long-term plan is to take AI infrastructure off the planet entirely. The premise is straightforward: building data centers on the ground is becoming a bottleneck. They require enormous amounts of electricity, water for cooling, available land, and the patience to navigate local planning approvals. Demand is outrunning all of it. Space sidesteps most of these constraints. Solar power is effectively unlimited in orbit. There is no cooling problem. The stated goal is to launch 100 gigawatts of compute capacity to space every year. To that end, SpaceX has asked regulators for permission to launch up to one million satellites functioning as orbital data centers.

The case for SpaceX specifically doing this is stronger than it sounds. Its launch cost advantage, extended further by the Starship rocket, means it can put hardware into orbit cheaper than any rival. And once the hardware is up, Starlink's existing laser mesh network, already spanning thousands of satellites, provides the communication backbone to connect orbital compute back to Earth. No other company can credibly claim both sides of that equation at scale.

The risks are significant, though. Latency makes orbital compute poorly suited to real-time applications. Chips in space face radiation and cosmic rays that ground facilities never encounter. Servicing or replacing failed hardware in space is harder than swapping out a server rack. And unlike terrestrial data centers, which can be built incrementally, the orbital version requires launching thousands of satellites before the network delivers meaningful capacity. Amazon's (AMZN) cloud chief has publicly said orbital data centers are nowhere close to being practical.

Musk's Track Record Of Changing Narratives

There is a broader pattern worth keeping in mind.

In the past, Elon Musk has re-framed his companies around whatever story investors are most willing to pay for at a given moment. When SolarCity ran into rough weather, it was merged into Tesla (TSLA), which was re-positioned as an alternative energy company. Tesla's story has since shifted again, now centering on physical AI and its humanoid robot, Optimus, just as EV sales cooled.

SpaceX is following the same playbook. The AI pivot positions SpaceX as the infrastructure backbone for a compute-hungry world, with orbital data centers as the long-term answer to limits that terrestrial infrastructure is already hitting.

That being said, at a near triple-digit revenue multiple, execution has to be flawless across some of the hardest engineering problems on Earth, and arguably off it. That makes this a very high-risk, high reward bet.

SpaceX is asking investors to back an AI infrastructure thesis priced for perfection, at multiples that leave little room for the unexpected. Balancing speculative bets like this against proven cash-generating platforms becomes critical. A disciplined portfolio approach helps you stay invested by limiting the impact of market shocks. While consistently beating the market is a challenge, the Trefis High Quality (HQ) Portfolio is designed to make it an achievable goal. The HQ strategy has consistently outperformed its market benchmark since inception, delivering returns of over 105 percent.

1. Financial Times [↩]