SpaceX Plans Self-Manufacturing of GPUs

At a crucial juncture in preparing for an IPO valued at $1.75 trillion, SpaceX has disclosed plans to manufacture its own GPUs, simultaneously alerting potential investors to the risks of chip supply shortages and high capital expenditures, bringing its ambitions and potential bottlenecks in AI and space business布局 to the forefront.

Core Disclosure on the Eve of IPO: In-House GPU Manufacturing as a Significant Capital Expenditure

Excerpts from the S-1 registration document obtained by the media show that SpaceX is expected to go public by this summer, with a current valuation of about $1.75 trillion. In its submission to the U.S. SEC, the company clearly lists "in-house manufacturing of GPUs" as a major capital expenditure project underway. In the risk factors section, it lists insufficient chip supply as a potential growth bottleneck, with the specific scale of this expenditure yet to be disclosed.

Tied to Musk's Terafab Joint Manufacturing Project

SpaceX's GPU manufacturing plan is deeply tied to the Terafab project, which is jointly advanced by several of Musk's companies. The project, created by SpaceX, xAI, and Tesla, is located in Austin, Texas, and is positioned as a comprehensive facility for advanced AI chip manufacturing, with Musk leading the planning.

Musk has previously stated that Terafab's production target covers chips needed for automobiles, humanoid robots, and space data centers, but the specific types of chips, including GPUs, and core details such as manufacturing technology have not been clearly disclosed. The main body responsible for the technological implementation of the manufacturing process is also temporarily undecided. Musk mentioned at Tesla's analysts' conference on Wednesday that by the expansion phase of Terafab, Intel's next-generation 14A process technology may be ready for mass production, stating that this "seems to be the right choice".

The High Industry Threshold Becomes a Key Variable in IPO Valuation

GPU manufacturing is a high-threshold area recognized in the chip industry, where the mainstream mode is highly dependent on professional foundries such as TSMC. Industry leader Nvidia dominates the AI computing power market with general-purpose GPUs, while Google takes a specialized route through self-developed TPU, both having accumulated technology and production capacity over many years.

The disclosure by SpaceX on the eve of the IPO has highly focused market attention on the feasibility and implementation timeline of its AI chip strategy, adding new core variables to potential investors' valuation judgments. The difficulty of technological implementation and high capital expenditures have become one of the most watched risk points for the company before its listing.