Elon Musk’s Terafab targets 1 million wafers a month after 2032
Elon Musk's Terafab, a joint venture of Tesla and SpaceX, has informally told equipment makers it is seeking to produce 1 million wafers per month after 2032, according to Financial News. About 70% of this output, or 700,000 wafers, is planned to go to memory, with more than 300,000 going to logic chips. Terafab's procurement team has already visited Japan to ask some local toolmakers for delivery within months. The memory allotment would rival Samsung's estimated DRAM production capacity if it were all of that type.
Samsung's DRAM capacity was estimated last year at 650,000-700,000 wafers a month, while SK hynix is reportedly at about 550,000, with about 1 million a goal by 2030. The comparison has to be taken in context, as Terafab's 700,000 measures all of its expected memory output, which could be NAND or DRAM, while the Samsung and SK hynix figures measure only DRAM. As it is, the memory portion alone would be 3.5 times the reported 200,000 at SK hynix's Wuxi fab in China.
Many in the industry were skeptical when Musk announced the plans in March, but orders are now leading more companies to see a real business opportunity. The timeline is aggressive, as it usually takes three years or more to get from planning to volume at a leading-edge fab, and Terafab is targeting volume production in the second half of 2028. Procurement officials from the company reportedly visited Japan earlier this month to meet local suppliers and to work on details on bringing in equipment, with some suppliers asked to deliver within a few months.
The Elec also reports that Terafab representatives visited booths of about 10 Korean materials, parts, and equipment firms at SEMICON West, with some NDAs signed.
Logic chips are one thing - Jiji Press reports Terafab is working with Intel to adopt its 2nm and 1.4nm processes - but memory is another, and it's not clear what type of memory is on the docket. Technical help from a Taiwanese company is under consideration for memory, but no partner has been named, leaving 70% of the planned wafers without a disclosed process technology.
Memory production tends to split between DRAM and NAND flash, the latter for non-volatile storage. A Tesla job listing for a Terafab Memory Process Integration Engineer, reported in August, covers end-to-end DRAM process integration and DRAM node development, and supports emerging memories such as MRAM, RRAM, and 3D DRAM. Terafab hasn't said how its DRAM would be divided among products like HBM and LPDDR, both useful for AI and both mentioned in its job postings, or whether it expects to make and stack the DRAM itself or package chips it brings in.
Building a cutting-edge logic fab on U.S. soil is its own challenge, and Tesla and SpaceX lack chip-manufacturing experience. Even Musk, on a SpaceX call last month, called Terafab's dates very aggressive" and said he put big error bars" on the timeline. The first SRAM test chip is targeted for around May or June 2027 at the Austin research facility.
How TSMC might fit in remains open. TSMC says any fab with its involvement has to be run by TSMC, while Musk said his own companies would build and operate Terafab, a scenario which could leave TSMC as a space tenant. At least the spending plans are known: $16.8 billion for the first phase in Grimes County, Texas, with total investment of up to $119 billion if later phases proceed. A pilot plant could be completed as early as July 2027.