Elon Musk said his AI operation will be running on at least 13 gigawatts of compute by the end of 2027 — up sharply from the 1.7 GW currently in operation.

For scale, one gigawatt is roughly the output of a large nuclear reactor unit, so 13 GW corresponds to the generating capacity of roughly a dozen such reactors dedicated to computing.

Musk laid out the underlying figures in more detail on SpaceX's Q3 2026 earnings call on October 20, and the numbers describe SpaceX's own AI buildout. The plan amounts to a roughly sevenfold increase in computing capacity over about 15 months. Sizing buildouts in gigawatts rather than chip counts has become the industry norm, as electricity and cooling — which take years to procure and construct — increasingly determine how quickly AI data centers can come online.

The math behind 13 gigawatts

On the call, Musk pointed to a potential increase of approximately 10 GW in 2027, which would lift total compute capacity to around 13 GW by year-end.

SpaceX's AI compute capacity stood at 1.7 GW as of September 30, 2026, up from 1.4 GW in Q2. The near-term target is more than 2 GW by the end of 2026. Reaching 13 GW a year later means the buildout must accelerate dramatically.

From Q2 to Q3, capacity grew by about 0.3 GW. The 2027 plan calls for roughly 10 GW to be added within a single year, which works out to about 2.5 GW per quarter if spread evenly.

Nvidia chips by hundreds of thousands

Musk said SpaceX is installing about 200,000 Nvidia GB300 GPUs every month for the remainder of 2026. After that, deployments are expected to shift toward Nvidia's upcoming Vera Rubin architecture. The GB300 sits in Nvidia's Blackwell family, and Vera Rubin is the next-generation architecture Nvidia has announced as its successor.

The headline project is a large-scale training cluster. Musk described a system of 440,000 GPUs that is expected to be operational by mid-2027.

SpaceX's longer-term plans call for power and cooling infrastructure reaching 15 to 20 GW by the end of 2027, a range that sits above the 13 GW compute target. The spread between the two figures reflects the volume of supporting electricity and cooling capacity that GPU facilities of this size require alongside the compute itself.

The revenue engine paying for it

SpaceX's Super Intelligence segment, which includes xAI cloud services and other offerings, generated $7.56 billion in revenue in Q3 2026. That figure was up 830% year-over-year, making it SpaceX's largest business area. That growth anchors the funding side of a capital program measured in gigawatts and hundreds of thousands of GPUs.

What this means

For Nvidia, a customer installing roughly 200,000 GB300 GPUs per month before transitioning to Vera Rubin represents a large and sustained order book. For companies that supply power, cooling, and data center construction, the 15 to 20 GW power and cooling target may matter even more than the chip count. The targets also land in a market where other major AI builders — including OpenAI, Meta, Google, and Amazon — have announced multi-gigawatt data center programs of their own, putting access to large blocks of electricity at the center of expansion planning across the industry.

The 440,000-GPU cluster is the first milestone to watch. It is expected to come online by mid-2027, so its status will indicate whether the broader buildout is on schedule well before year-end.

The year-end 2026 figure is the other near-term checkpoint. Clearing more than 2 GW would show the current deployment pace is holding, while falling short would raise questions about the far steeper climb that follows.