Tesla Targets March 21 for Terafab Launch, Its $25 Billion Bid to Build the World's Largest AI Chip Factory
Tesla plans to formally launch its Terafab semiconductor facility on March 21, aiming to produce up to 200 billion custom AI chips annually at 2nm process technology.
Overview
Tesla is preparing to formally launch its Terafab project on March 21, a vertically integrated semiconductor fabrication facility that CEO Elon Musk describes as the company’s answer to a looming AI chip supply crisis. The facility, estimated to cost approximately $25 billion, would combine logic processing, memory fabrication, and advanced packaging under one roof at a scale that no private company outside Taiwan and South Korea currently operates.
Musk announced the timeline on March 14 in a post on X, writing that the “Terafab Project launches in 7 days.” The announcement follows Tesla’s January 28 earnings call, where Musk first disclosed the initiative, telling investors that “in order to remove the probable constraint in three or four years, we are going to have to build a Tesla TeraFab” — describing it as “a very big fab that includes logic, memory and packaging, domestically.”
What We Know
Terafab is designed to produce between 100 billion and 200 billion custom AI and memory chips per year, targeting an initial output of 100,000 wafer starts per month with a stated ambition to scale toward one million — roughly 70 percent of Taiwan Semiconductor Manufacturing Company’s total current output, concentrated in a single U.S. facility.
Tesla is targeting 2-nanometer process technology, the most advanced node currently entering commercial production. The company’s fifth-generation AI chip, AI5, is among the first products the facility is designed to produce, with small-batch production expected in 2026 and volume production projected for 2027. Tesla is also in early stages of developing the AI6 chip.
The AI5 chip reportedly delivers 40 to 50 times more compute performance and nine times more memory than its predecessor, AI4. These chips are intended to power Tesla’s Full Self-Driving software, the Cybercab robotaxi program, and the Optimus humanoid robot line. Musk has also indicated that Terafab would supply chips for xAI’s Dojo supercomputer and Grok model training infrastructure.
All indications point to the North Campus of Giga Texas in Austin as the likely location, based on construction activity observed adjacent to the existing facility. The $25 billion estimated cost forms part of Tesla’s record capital expenditure plan for 2026, which exceeds $20 billion.
What We Don’t Know
The March 21 date almost certainly does not mean a fully operational chip fabrication facility opens its doors. Semiconductor fabs of this scale take years to construct and commission. Industry observers expect the launch to signal a formal project announcement with location and timeline details, a groundbreaking ceremony, or the start of construction on the first phase.
Tesla has not disclosed how it plans to source the extreme ultraviolet lithography equipment required for 2nm production. ASML, the sole manufacturer of such machines, has a multi-year backlog, and securing sufficient tools would represent a significant supply chain challenge. Tesla has not released detailed timelines for when the facility would reach its stated production targets.
The relationship between Terafab and Tesla’s existing chip manufacturing partnerships also remains unclear. Samsung Electronics is already building a plant in Taylor, Texas, under a $16.5 billion deal to produce Tesla’s AI6 chips, and TSMC continues to fabricate current-generation Tesla silicon. Whether Terafab is intended to replace or supplement these arrangements has not been specified.
Analysis
Terafab represents the most ambitious vertical integration play in the automotive industry’s history. No automaker has attempted to operate its own leading-edge semiconductor fabrication facility. The project would place Tesla in direct competition with dedicated foundries like TSMC and Samsung for talent, equipment, and raw materials.
The scale of the undertaking invites scrutiny. Building a greenfield 2nm fab is a multi-year, multi-billion-dollar effort that even established chipmakers approach with caution. Intel’s own foundry ambitions have encountered repeated delays and cost overruns. Tesla would need to recruit thousands of semiconductor engineers and process technicians in a labor market where such talent is already scarce.
Yet Musk’s track record of pursuing vertical integration — from battery cells to insurance to charging networks — suggests the announcement is more than aspirational. Tesla’s AI chip team, led by former Apple silicon architect Pete Bannon, has already delivered four generations of custom inference hardware. The question is whether that design expertise can scale to encompass fabrication as well.
If Tesla succeeds, it would fundamentally alter the economics of its AI-dependent product lines by eliminating third-party fabrication margins and securing guaranteed chip supply. If it stumbles, the $25 billion capital commitment could weigh on a company that is simultaneously funding robotaxi deployment, humanoid robot production, and energy storage expansion.