SpaceX's Orbital Data Center Plan Slammed as 'Dumb' by Analyst

SpaceX's Orbital Data Center Plan Under Fire as Analyst Calls Idea 'Dumb'

SpaceX's ambitious plan to build AI data centers in orbit has drawn sharp criticism from geopolitical analyst Peter Zeihan, who called the concept "dumb on so many levels" in a recent video. Zeihan argues that a 1-gigawatt orbital data center would require roughly 500 Starship launches just for assembly, and that high-performance chips would need massive radiation shielding—adding an estimated 20 million pounds of weight. He also highlights the cooling challenge: in the vacuum of space, heat cannot dissipate via air, necessitating cooling fins spanning tens of square kilometers, which he estimates would cost around $250 billion.

The criticism echoes broader skepticism. Energy consultancy Wood Mackenzie modeled the costs at about $170 billion for a 1-gigawatt orbital facility—still about triple the cost of a comparable terrestrial data center, according to Benzinga. Prediction platform Kalshi assigns a 14% chance that an orbital data center with at least one megawatt of capacity will be operational before January 2029, well below SpaceX's target of 2028. By 2035, that probability rises to 39%, suggesting long-term but uncertain prospects. These debates unfold as SpaceX, having acquired xAI in February 2026, pitches itself to investors as a vertically integrated AI company, not just a rocket operator.

For now, SpaceX's ground-based operations remain the backbone. Since May 2026, AI firm Anthropic has been paying $1.25 billion per month through May 2029 for capacity in SpaceX's terrestrial Colossus data centers, with a mutual 90-day termination clause. SpaceX's IPO risk disclosures also concede that orbital computing relies on unproven technologies and may never be economically viable.

Why Critics Say SpaceX's Orbital Data Centers Don't Add Up

Why Zeihan's Back-of-the-Envelope Math Matters

Zeihan's numbers are approximations, but they spotlight three unsolved problems: launch cadence, radiation, and thermal management. Even if Starship delivers 100 tons per launch (optimistic for a heavy-lift rocket), 500 launches to loft 1 GW of hardware is a staggering logistical feat. Radiation-hardening high-end GPUs is costly and adds mass, potentially undermining the payload efficiency that makes SpaceX's launch business attractive. Cooling in orbit is purely radiative, and the required fin area would be enormous and fragile, all driving costs far above terrestrial economics.

Wood Mackenzie's More Conservative Cost Model

Wood Mackenzie's $170 billion figure is based on power infrastructure, launch, and assembly costs. While lower than Zeihan's estimate, it still implies a cost premium of roughly $110 billion over a ground-based equivalent. This suggests that orbital data centers would only make sense if energy costs on Earth skyrocket or if in-space compute offers unique advantages—such as lower latency for certain orbital assets or immunity to seismic risks. None of these have been convincingly demonstrated at scale.

What the Kalshi Odds Tell Us

The 14% probability of a 1 MW orbital facility by 2029 reflects deep market skepticism. A 1 MW facility is a tiny fraction of SpaceX's 1 GW vision, yet traders still see it as unlikely. Even the longer-term 39% by 2035 indicates that the consensus is "possible, but far from certain." For SpaceX's valuation, if the orbital data center story is a key pillar of its AI pivot, failure to deliver an early demonstration could weigh on the stock.

The Terrestrial Anchor Is the Real Story

While the orbital concept grabs headlines, SpaceX's earthbound data center business is quietly generating substantial contracted revenue—$45 billion over three years from Anthropic alone, with a termination clause that offers flexibility for both sides. This contract showcases SpaceX's ability to secure large AI clients without the orbital risk, and it may be the real growth engine as spaceborne plans mature.

What the Orbital Data Center Debate Means for SpaceX's Investors and Partners

• SpaceX's valuation hinges partly on the orbital AI narrative, but IPO risk disclosures and market pricing suggest significant uncertainty. Investors should weigh the $1.25B/month Anthropic contract as a steadier revenue baseline against the high-risk orbital venture.
• Competitors building terrestrial data centers can take comfort that orbital economics remain deeply unproven, with costs 3x or more above ground. The terrestrial market for AI compute is likely to remain SpaceX's near-term focus, given the Anthropic deal's scale.
• Regulatory and insurance unknowns could further delay orbital projects. SpaceX's IPO filing mentions unproven tech; obtaining launch licenses, space debris mitigation, and liability coverage for such massive infrastructure could add complexity not yet priced in.