SpaceX has just given Foxconn a major order for NVIDIA GB300 AI server racks, significantly upping its training- and inference-related compute footprint.

According to a report by a Taiwan-based publication, SpaceX has just awarded Foxconn a $52 billion order for around 13,000 racks of NVIDIA GB300 AI servers , with shipments expected to begin in Q4 2026 and continuing until Q1 2027, and where each rack costs around $4 million. Do note that Foxconn expects to corner around 40 percent of the global AI server market by the end of 2026.

As we reported recently, as of May 2026, SpaceX's Colossus 1 data center supported over 220,000 NVIDIA GPUs , divided between H100, H200, and around 30,000 units of the GB200 AI accelerators.

Meanwhile, the Colossus 2 data center currently boasts of over 550,000 GPUs , divided between GB200 and GB300 accelerators.

Do note that, in the runup to its IPO, SpaceX locked a Cloud Service Agreement with Google, furnishing it with a compute capacity equivalent to "110,000 NVIDIA GPUs, CPUs, memory, and other related components," all for a consideration of $920 million per month .

Also, the parent company of Starlink recently inked a similar agreement with Anthropic, providing access to "220,000 NVIDIA GPUs (including H100S, H200S, GB200s, and others)," all for a consideration of $1.25 billion per month or $15 billion per year.

While the Anthropic deal is locked to SpaceX's Colossus 1 data center, where a mish-mash of GPUs renders the data center less attractive for training future iterations of Grok AI, and where xAI consumes just 11 percent of the total available FLOPs, the Google deal is primarily tied to the Colossus 2 data center .

Back in June 2026, SpaceX signed another $6.3 billion compute deal for the Colossus 2 data center, this time with Pentagon-linked Reflection.

Follow Wccftech on Google to get more of our news coverage in your feeds.