SpaceX Plans 10GW AI Compute Buildout by 2027

SpaceX delivered a bombshell announcement during its inaugural earnings call as a public company on August 4, revealing the largest single artificial intelligence compute commitment in history. The aerospace giant plans to install 2 gigawatts (GW) of AI compute capacity by the end of 2026, scaling to approximately 10 GW by the end of 2027. Elon Musk acknowledged that while a 20 GW power and cooling infrastructure target exists, 15 GW represents the realistic outcome. This aggressive expansion effectively transforms the company into the world’s most ambitious compute infrastructure provider.

The commitment centers on an exclusive partnership with Nvidia, building future compute capacity on the chipmaker’s Vera Rubin architecture. Musk characterized Vera Rubin as “the best AI computer” available, signaling a strategic bet on density and performance to maintain competitive advantage. According to Melius Research analyst Ben Reitzes, even the initial 2 GW deployment could add roughly $100 billion in incremental revenue for Nvidia, potentially pushing the chipmaker toward a $1 trillion valuation milestone. SpaceX has already secured $6.7 billion in forward cloud services contracts starting in October 2026, demonstrating tangible market demand.

The financial architecture supporting this expansion reflects SpaceX’s rapid growth trajectory. The company reported Q2 2026 revenue of $7.8 billion, representing a 92% year-over-year increase that comfortably beat analyst expectations of $6.93 billion, according to CNBC. While the AI division posted an operating loss of $1.26 billion, the core business model is crystallizing around GPU rental, which generates 95% of AI revenue. CFO Bret Johnsen noted the company successfully monetizes available compute capacity to generate high incremental EBITDA margins.

Major Cloud Tenants Drive Revenue Growth

SpaceX has secured commitments from industry heavyweights that validate its infrastructure strategy. Google currently pays $920 million per month for access to roughly 110,000 GPUs, while Anthropic contributes approximately $1.25 billion per month at the Colossus 1 facility in Memphis. These contracts demonstrate the extraordinary demand for large-scale compute infrastructure among AI developers. Industry analysis suggests that frontier AI companies operating GB300 clusters can generate over $100 billion per GW annually in revenue when selling API inference services, making GPU rental at current neocloud prices highly profitable.

The economics of AI inference explain why companies willingly pay premium rates for compute access. Using conservative rental pricing of approximately $3 per GPU-hour, annual costs reach around $12 billion per GW. However, companies like OpenAI and Anthropic can generate substantially higher revenue through token production, creating profitable arbitrage opportunities. Serving inference tokens proves unbelievably profitable for frontier model companies, particularly when operating at realistic performance levels that optimize tokens per second per GPU. This revenue potential justifies the enormous capital expenditure required for buildout.

Orbital Compute Strategy Extends Beyond Earth

SpaceX plans to extend its compute infrastructure beyond terrestrial data centers through an ambitious satellite network. The upcoming Starmind AI1 satellite, equipped with the Nvidia Space-1 Vera Rubin module, promises 25 times more AI compute per GPU compared to the H100 architecture. An FCC filing requests authorization for up to 1 million satellites in sun-synchronous low Earth orbit, attempting to solve latency and memory bottlenecks that plague ground-based clusters. This orbital differentiation represents a unique strategic advantage in the competitive AI infrastructure landscape.

The 2027 buildout target of 6-8 GW in a single year, with potential exceeding 10 GW, represents capital expenditure of approximately $300-500 billion at $50 billion per GW. This spending level matches expectations for AWS and Google, creating an unprecedented scenario where SpaceX competes directly with established hyperscalers despite significantly lower profitability. Industry experts believe the number is achievable given SpaceX’s execution track record. Energy infrastructure suppliers tracking 30+ turbine, engine, and fuel cell providers confirm sufficient gas generation equipment exists to support the timeline, validating the feasibility of SpaceX’s aggressive schedule.

Cellular Market Entry Adds Strategic Complexity

Beyond AI infrastructure, SpaceX signaled intentions to disrupt the US telecommunications industry during its earnings call. The company revealed plans to create small, low-power femtocell-type cellular base stations on existing Starlink dishes, avoiding massive upfront capital expenditure for cell towers and terrestrial spectrum acquisition. This strategy sent shockwaves through the industry, though T-Mobile’s CEO dismissed the threat as “exaggerated,” questioning which consumer problems satellite companies would solve by entering the market.

“The threat posed to mobile operators from satellite companies such as Starlink has been exaggerated and I struggle to see which consumer problems they would solve by entering the market,” the T-Mobile CEO stated.

SpaceX recently acquired approximately 65 MHz of nationwide, exclusive-use, contiguous spectrum for $19.6 billion, divided into three distinct bands. The acquisition includes around 15 MHz of unpaired nationwide Advanced Wireless Services-3 (AWS-3) spectrum, already used by major carriers for LTE/5G networks and supported by most premium and mid-range smartphones. Additionally, SpaceX secured around 40 MHz of nationwide AWS-4 spectrum, with propagation characteristics ideal for space-to-ground communication, though less widely supported by existing cellphones. The package includes approximately 10 MHz of nationwide H-Block spectrum, providing up to 20 times the data throughput compared to low-band frequencies previously leased from T-Mobile.

Strategic Assets Position SpaceX for Disruption

Despite T-Mobile’s dismissive stance, SpaceX possesses formidable strategic advantages. The company commands well-endowed financial resources, a vast operational satellite constellation, and Elon Musk’s relentless ambition driving execution. Large-scale, near-term compute capacity remains remarkably scarce, commanding premium pricing up to $50 billion per GW annually. SpaceX’s strategy involves developing anything available and bringing infrastructure online as rapidly as possible, capitalizing on this scarcity premium. The combination of AI infrastructure dominance and potential cellular market disruption creates multiple revenue streams that could fundamentally reshape technology and telecommunications landscapes over the next two years.