After SpaceX completed its historic initial public offering (IPO) on Nasdaq, its stock price soared by more than 19%, and the company's market value immediately rose to approximately US$2.11 trillion, making founder Elon Musk the first entrepreneur in the world to have a net worth of over US$1 trillion. Amid the enthusiastic pursuit of the capital market, Musk publicly stated that he hopes to elevate the "exciting partnership" with NVIDIA to a "new level," indicating that the two parties' bonding in the fields of space and artificial intelligence is expected to further deepen.
Hours after SpaceX shares began to trade freely, Nvidia took the lead in posting a post on social platform Musk later forwarded and responded, expressing his expectation to push cooperation with Nvidia to the "next-level." This interaction was regarded by the outside world as a signal that the two technology giants may launch a new round of in-depth cooperation.
In its congratulatory message, NVIDIA also reviewed the nearly ten years of cooperation between the two parties, from the personal delivery of the world's first NVIDIA DGX‑1 supercomputer in 2016 to the recent delivery of the DGX Spark system customized for SpaceX, showing that its AI computing power has long been deeply embedded in Musk's space business layout. Now that SpaceX has completed its IPO and its financing capabilities have significantly increased, it is generally expected that the scale and scope of cooperation between the two parties will expand significantly.
This trend is also closely related to SpaceX’s latest AI-specific satellite plan. The company recently officially announced the first-generation dedicated AI computing satellite "AI1". The satellite is designed to support a peak computing load of up to 150 kilowatts and an average computing load of 120 kilowatts. It is equipped with a series of high-power and high-density computing support solutions such as liquid-cooled radiators, meteoroid protective structures, centralized computing modules and deployable solar arrays. These AI1 satellites will be mass-produced by SpaceX’s “Gigasat” super factory in Texas, USA, and are positioned as an infrastructure platform for orbit-level AI computing power.
In this context, outsiders speculate that Musk’s claim of “raising the cooperation with NVIDIA to a new level” is likely to be related to the computing platform used on the AI1 satellite, for which NVIDIA is expected to provide GPU and related AI hardware solutions. If the two parties engage in deep binding on orbital AI computing power, it means that NVIDIA's high-end GPU and system-level solutions will further extend from ground data centers to space scenes, while SpaceX will receive dedicated computing power support optimized for its constellation-based AI network.
In addition to the satellite itself, SpaceX is also accelerating the construction of its large-scale manufacturing facility called "Terafab". Musk has previously said that the total area of the facility will reach about 100 million square feet, which is about 10 times the size of Tesla's Giga Texas factory (Giga Texas). The design capacity can achieve an output of 1 terawatt per year. This ultra-large-scale production line will become a key production platform to support AI satellites and related energy and infrastructure projects, and also reflects Musk’s long-term plan for energy and computing infrastructure.
In terms of cloud computing power, SpaceX has reached a cloud service agreement with Google on the eve of its IPO. The latter will provide SpaceX with computing resources equivalent to "110,000 NVIDIA GPUs, CPUs, memory and other related components". The contract value is approximately US$920 million per month. This cooperation not only provides powerful computing power support for SpaceX's ground data processing and AI training, but also further consolidates its resource layout in the field of global cloud computing and AI infrastructure.
At the same time, as the parent company of Starlink and xAI, SpaceX recently signed another computing power agreement with the AI company Anthropic to provide it with approximately 220,000 Nvidia GPUs (including H100, H200, GB200 and other high-end products). The contract value is approximately US$1.25 billion per month and approximately US$15 billion per year. Through large long-term orders with partners such as Google and Anthropic, Musk's companies have locked in Nvidia's high-end GPU production capacity on the one hand, and on the other hand have strengthened their bargaining power and industrial influence in the broader AI ecosystem.
Taken together, when SpaceX successfully IPOed and Musk became a "trillionaire", its cooperation with Nvidia in the fields of space and AI not only did not stop, but showed a strong willingness to further upgrade. As the AI1 satellite, Terafab super factory and multi-party cloud and computing power cooperation are implemented one after another, the "next level" partnership described by Musk is likely to mean that a broad "space-AI computing power network" covering the entire chain of orbiting satellites, ground factories, cloud data centers and AI applications is taking shape.