Futures
Access hundreds of perpetual contracts
TradFi
Gold
One platform for global traditional assets
Options
Hot
Trade European-style vanilla options
Unified Account
Maximize your capital efficiency
Demo Trading
Introduction to Futures Trading
Learn the basics of futures trading
Futures Events
Join events to earn rewards
Demo Trading
Use virtual funds to practice risk-free trading
Launch
CandyDrop
Collect candies to earn airdrops
Launchpool
Quick staking, earn potential new tokens
HODLer Airdrop
Hold GT and get massive airdrops for free
Pre-IPOs
Unlock full access to global stock IPOs
Alpha Points
Trade on-chain assets and earn airdrops
Futures Points
Earn futures points and claim airdrop rewards
Elon Musk said that space AI data centers "will happen even if you just think about it," but SpaceX's IPO warns that it may not proceed commercially.
SpaceX warns investors in its IPO application that the company’s orbital AI computing project “involves significant technical complexity and unproven technology, which may not achieve commercial viability.” The space data center will be exposed to the unique risks brought by the unpredictability of the space environment. This statement directly contrasts with CEO Elon Musk’s public remarks at the World Economic Forum (WEF), where he said the space data center is “obviously the only expansion path.”
(Background: Breaking! SpaceX acquires xAI with a valuation of $1.25 trillion, Musk aims to build a “space + AI” giant for IPO readiness)
(Additional context: The biggest wealth-creating IPO in history is about to begin: SpaceX, OpenAI, and Anthropic lead the charge)
Table of Contents
Toggle
The application is a legal commitment to investors, while Musk’s speeches are meant to communicate his vision to the audience.
When Musk talks about space AI data centers in public, he is describing the future he envisions; on the other hand, when SpaceX’s lawyers and CFO write risk disclosures in the S-1 filing, they emphasize the company’s legal obligation to present a realistic assessment.
This application, obtained and verified by Reuters, contains content that has never been publicly reported before.
What did Musk say? What does the IPO reveal?
In January 2026, Musk stated at the World Economic Forum that space AI data centers will become the “cheapest option” within two to three years, calling it the “obvious answer.”
In February, he further said, “Space AI is clearly the only expansion path.” His reasons include that the solar power prices in low Earth orbit are only about $0.005 per kWh, roughly one-fifteenth of ground wholesale prices, and that the vacuum environment directly eliminates the need for cooling water.
In the S-1 filing, SpaceX described another scenario: “We are developing orbital AI computing and plans for in-orbit, lunar, and interstellar industrialization, which are currently in early stages and involve significant technical complexity and unproven technology, which may not achieve commercial viability.”
The document further warns that the space data center will operate “under harsh and unpredictable space environment conditions, exposed to extensive and unique space-related risks, which could lead to failure or malfunction.”
The technical challenges have not disappeared
The cautious language in the application is justified, as SpaceX proposed in its FCC application to deploy up to 1 million solar-powered data center satellites in low Earth orbit. The scale of this project presents engineering challenges that have been widely discussed in the industry.
Thermal management is one of the core difficulties. GPUs generate heat during computation that cannot be dissipated through convection in a vacuum; it can only be cooled via radiation, which remains an unresolved issue at commercial scale for high-density computing deployments.
Network bandwidth is another bottleneck. Even with Starlink providing satellite connectivity, the latency and throughput between in-orbit data centers and ground stations remain significant limitations in training large models. Maintenance and upgrade costs are also not comparable to ground data centers, especially since there’s no Rocky to repair them in space.
A former NASA engineer publicly stated that putting data centers into space is “the worst idea I’ve ever heard.” The wording in the application aligns more with this assessment than Musk’s optimistic predictions.
Further reading: Former NASA engineer: Building space data centers is the worst bad idea I’ve ever heard
The two narratives around $1.75 trillion
SpaceX is preparing for an IPO with an estimated valuation of about $1.75 trillion. If the $75 billion fundraising target is achieved, it will surpass Saudi Aramco’s record as the largest IPO in history. In this scale of market storytelling, the space AI data center is a key narrative element, representing a story of computing power growth without terrestrial limitations.
The risk disclosures in the S-1 will not eliminate this story, but they legally require SpaceX to clearly inform investors that this technological dream has not yet been proven to be profitable.