NVIDIA Wins Exclusive SpaceX Pact For Space-Based AI Servers
SpaceX set off some financial fireworks in the markets after reporting a surge in quarterly revenue alongside an exclusive hardware commitment with NVIDIA, even as rising capital expenditures temporarily weighed down its stock.
Despite SpaceX narrowing its second-quarter net loss to $0.09 per share from $0.34 a year prior, company revenues nearly doubled to hit $7.81 billion. Growth was heavily driven by its growing AI division, where quarterly revenue leaped 213% to reach $2.56 billion. Fueling this expansion are lucrative cloud hosting agreements with major industry players like Google and Anthropic, alongside growing enterprise demand for Grok and X subscriptions.SpaceX is partnering with @Nvidia to design the Starmind AI1 satellite compute payload.
— SpaceX (@SpaceX) August 4, 2026
Each of the Starmind satellites will include NVIDIA Rubin GPUs and Vera CPUs for datacenter class space compute → https://t.co/4MOQv0DvTQ pic.twitter.com/rC7UBAznAO
However, Wall Street delivered a mixed response to the earnings release. SpaceX shares tumbled nearly 10% in pre-market trading, largely spooked by an extraordinary spike in capital expenditures. Spending exploded from $2.83 billion to $18.37 billion over the last 12 months, with direct AI infrastructure outlays surging more than twenty-fold to $15.83 billion.
Investor anxiety was further compounded by the impending expiration of the company’s post-IPO insider lockup period, which liberates roughly 912 million insider shares for trading and doubles the public float. SpaceX CFO Bret Johnsen attempted to calm nerves by highlighting the efficiency of the capital deployment, noting that payback periods on AI hardware investments remain well under a single year.
SpaceX plans to deploy NVIDIA Vera Rubin NVL72 rack-scale systems (a.k.a. Kyber) across its ground facilities, targeting 2 gigawatts of total compute capacity by the end of this year before scaling to 10 gigawatts by late 2027.
SpaceX has committed to using Nvidia GPUs exclusively because they are the best
— Elon Musk (@elonmusk) August 4, 2026
Crucially, the collab extends beyond terrestrial data centers and into outer space. To wit, SpaceX is actively co-developing the Starmind AI1 satellite compute payload, leveraging NVIDIA's space-certified Vera Rubin modules to process real-time geospatial data, autonomous flight analytics, and orbital intelligence directly in Earth's orbit. By combining orbital data centers with ground-based infrastructure, Musk aims to bypass terrestrial bottlenecks in power, land, and cooling to drive its deep-space computing ambition.
