TERAFAB

A Vertically Integrated Chip Empire — Under One Roof

Terafab is not just a foundry. Elon Musk officially unveiled it on March 21, 2026 at the defunct Seaholm Power Plant in Austin, Texas, calling it "the most epic chip building exercise in history by far." It is a joint venture between Tesla, SpaceX, and xAI — the AI company SpaceX recently acquired in an all-stock deal — designed to bring every stage of semiconductor production in-house: chip design, lithography, fabrication, memory production, advanced packaging, and testing.

At full capacity, Terafab would scale to roughly 70% of TSMC's total global wafer output — from a single facility. The stated goal is one terawatt of computing power annually, split between two chip families: Tesla's AI5 inference chips for vehicles and Optimus robots, and D3 chips custom-designed for SpaceXAI's orbital satellite constellation.

A novel manufacturing claim accompanies the announcement: rather than traditional ultra-cleanroom protocols, Tesla says it will use a wafer isolation method — enclosing individual wafers in protective enclosures throughout the fabrication process, theoretically enabling processing in non-sterile ambient conditions. This is unproven at any leading-edge node and should be treated as speculative until demonstrated at scale.

A Supply Chain Ceiling — and a Dojo-Shaped Gap

• External suppliers are hitting a ceiling. At Tesla's Q4 2025 earnings call, Musk warned that capacity from TSMC, Samsung, and Micron would become inadequate within three to four years. "There's a maximum rate at which they're comfortable expanding," he said. The chip supply constraint facing Tesla is not hypothetical — it mirrors what every major AI compute buyer is navigating right now.
• Autonomous vehicles and robotics demand is scaling exponentially. Both the Cybercab robotaxi and Optimus humanoid robot require AI inference chips at a volume Tesla's current supply chain cannot sustainably support. Without a proprietary supply, both programs are at the mercy of TSMC and Samsung allocation queues.
• The Dojo program was shut down, leaving a strategic gap. In August 2025, Tesla disbanded its entire Dojo supercomputer team. Musk called it "an evolutionary dead end" and pivoted Tesla's AI training compute to a new supercomputer called Cortex, built on third-party Nvidia chips at the Austin headquarters. Terafab is the manufacturing answer to what Dojo was supposed to solve — but it's a different problem: where do the chips themselves come from?

Giga Texas North Campus, Austin

Terafab is planned for the North Campus of Giga Texas, co-locating it with Tesla's existing EV manufacturing operations and Cortex supercomputer. The announcement event was held at the Seaholm Power Plant — a decommissioned facility whose energy-infrastructure heritage was a pointed architectural choice given the power demands of leading-edge chip fabrication.

The joint venture structure brings together three Musk entities for the first time under a single capital project: Tesla (vehicles, Optimus), SpaceX (orbital infrastructure), and xAI (AI model training and inference). Each entity contributes a distinct demand signal; the logic is that consolidated volume justifies the fab's fixed cost structure.

Two Chip Families, Three Markets — 80% of Output Going to Space

• AI5 — Autonomous Vehicles & Optimus Robots (Internal). Tesla claims the AI5 delivers a 50x total improvement over the AI4. Small-batch production is targeted for 2026 with volume production in 2027. The chip powers full self-driving inference for the Cybercab robotaxi and provides the AI backbone for Optimus. Tesla also renamed its onboard inference chip from "FSD Computer" to "AI Computer" in software update 2026.2.9 — a signal the chip's identity is broadening beyond autonomous driving into a general AI compute platform.
• D3 Chips — Orbital AI Satellites (SpaceXAI). The most unexpected market: 80% of Terafab's compute output is earmarked for space-based orbital AI satellites. Musk's argument is physics-based — solar irradiance in space is roughly 5x greater than at Earth's surface, and heat rejection in vacuum removes a key thermal scaling constraint. The SpaceXAI thesis is that within two to three years, running AI workloads in orbit will be cheaper than doing so on the ground. This market does not yet exist commercially.
• Potential Third-Party Foundry Services (Future / Speculative). If Terafab achieves anything close to its stated 1 million WSPM capacity ambition, Tesla would be one of only a handful of entities capable of producing frontier AI silicon at scale. This opens a path to competing with TSMC and Samsung as a merchant foundry. Musk has not officially stated this as a goal, but the capacity math makes it a logical downstream opportunity.

From Shareholder Meeting to Groundbreaking in Four Months

The Strategic Logic Is Sound — If Execution Follows

• Supply Chain Independence. A domestic fab eliminates existential TSMC concentration risk for Tesla's most critical compute systems. This is a structurally sound argument — TSMC's Taiwan exposure is a legitimate geopolitical risk that every major AI company is currently trying to hedge.
• Vertical Integration Margin Capture. If Terafab produces chips at competitive cost, Tesla eliminates chipmaker markup from its cost structure — meaningful when autonomous systems require continuous silicon iteration at increasing volumes.
• CHIPS Act Alignment / National Security Relevance. Domestic 2nm fabrication is precisely what U.S. industrial policy is incentivizing. Terafab is eligible in principle for CHIPS Act subsidies, though Tesla has not confirmed this pathway. National security computing applications (via SpaceX's Space Force relationships) could also open government contract channels.
• Orbital AI Compute as a First-Mover Market. If space-based AI inference becomes economically viable, SpaceXAI — backed by Terafab silicon — would be the only vertically integrated orbital AI infrastructure provider on Earth. The market doesn't exist yet, but the physics argument is not trivial.
• Merchant Foundry Optionality. Scale creates optionality. A Terafab operating at even half its stated capacity ambition could serve external customers — potentially positioning Tesla as a competitor to TSMC for AI chip production in the 2030s.