Space Exploration Technologies Corp. (NASDAQ:SPCX) wants to put AI compute in orbit, and plenty of Redditors on investment subs think the idea makes little economic sense. The skepticism is intuitive. SpaceX's Starmind AI1 satellites will use Rubin GPUs and Vera CPUs from NVIDIA Corporation (NASDAQ:NVDA), yet Nvidia Corp. (NASDAQ:NVDA) is now delivering major AI hardware upgrades every year. Why launch expensive processors into space when newer hardware can make them economically obsolete within a few years?
The objection looks stronger... until we do the math.
SpaceX says AI1 will deliver 120 kilowatts of average compute while achieving vehicle efficiency of 70 kilowatts per metric ton. That works out to roughly 14.3 kilograms of spacecraft per kilowatt of compute.
SpaceX also says a fully and rapidly reusable Starship could eventually cut launch costs by at least 99% from NASA's historical $18,500-per-kilogram benchmark. A 99% reduction would eventually imply about $185 per kilogram. At AI1's claimed mass efficiency, that translates into only about $2,640 of launch cost per kilowatt of compute.
For perspective, an April study examining orbital data-center economics used roughly $10,000 to $40,000 per kilowatt as the cost of terrestrial data-center infrastructure. The comparison has its limitations: SpaceX's $2,640 figure covers launch cost only.
Satellite manufacturing cost is a different question. SpaceX's Redmond factory averaged roughly 70 Starlink satellites per week from December through April, while the company says AI1 will use proven Starlink technology and be more streamlined than satellites it already mass-produces.
Quilty Space estimates Starlink V3 costs around $1 million to manufacture. If AI1's non-compute hardware approached that level, its roughly 1.7-ton mass would put manufacturing plus launch at about $11,000 per kilowatt, near the low end of terrestrial infrastructure costs. AI1's larger solar arrays and radiators make that an optimistic benchmark, and SpaceX has disclosed no manufacturing cost, but the math suggests the spacecraft need not become extraordinarily cheap for orbital economics to start closing.
Institutional investors appear much less concerned. Insider Monkey's Q2 database showed 119 hedge funds holding long positions in SpaceX. Goldman Sachs recently identified SpaceX as one of only six stocks that both hedge funds and mutual funds are overweight. Cathie Wood's ARK Invest added roughly $27 million of SpaceX shares on August 21.
There is plenty of skepticism on the other side of Wall Street. Short interest reached 207.8 million shares as of July 31, up 25.9% from mid-July. David Einhorn has also questioned SpaceX's valuation, calling its $1.75 trillion IPO valuation a possible marker of a speculative market top.
At roughly $1.86 trillion, the company trades around 80 times trailing revenue. Even against current consensus estimates for 2027, its enterprise value works out to roughly 17 times sales and 28 times EBITDA. Goldman Sachs separately puts the stock at about 93 times projected 2027 earnings. For a company spending heavily on Starship, AI infrastructure and Starmind, the market is already assuming that today's extraordinary capital requirements eventually produce superior long-term economics that SpaceX promises for AI infra.
My read is that Reddit's skepticism overstates the problem. If Starship reaches full reusability and AI1 achieves SpaceX's claimed mass efficiency, launch economics look surprisingly workable. I would give SpaceX the edge on technical and economic feasibility, but not yet on proven economics. SpaceX's valuation gives it almost no room for the economics to disappoint.
While we acknowledge the potential of SPCX as investment, we believe certain other AI stocks offer greater upside potential and carry less downside risk. If you're looking for an extremely undervalued AI stock that also stands to benefit significantly from Trump-era tariffs and the onshoring trend, see our free report on the best short-term AI stock.
READ NEXT: NVIDIA (NVDA): What Foxconn and Super Micro Are Telling Us about the AI Boom and Pony AI Is Scaling Robotaxis Fast—Can the Stock Reach BofA's $17 Target?
Disclosure: None. Follow Insider Monkey on Google News.