Terafab: Inside the $25 Billion Chip Factory That Could Reshape the Space Economy

On March 21, 2026, Elon Musk stood at the north campus of Giga Texas in Austin and unveiled Terafab — what he described as the largest semiconductor fabrication facility ever conceived. The $20–25 billion project is a joint venture between Tesla, SpaceX, and xAI, designed to vertically integrate every stage of chip production: design, lithography, fabrication, memory production, advanced packaging, and testing under one roof.

But the headline number barely tells the story. This isn't just a chip factory for electric cars and robots. Eighty percent of Terafab's planned output is destined for space — specifically, radiation-hardened D3 processors designed to power SpaceX's planned constellation of orbital AI data center satellites. If Terafab delivers even a fraction of its stated ambitions, it will fundamentally alter the semiconductor supply chain for the entire space industry.

What Terafab Actually Is

Terafab is designed to produce two primary chip families. The AI5 is an inference processor for Tesla's Full Self-Driving system, Cybercab robotaxis, and Optimus humanoid robots — essentially the next-generation brain for Tesla's autonomous products. The D3 is a radiation-hardened, high-temperature processor purpose-built for SpaceX's orbital AI satellites, designed to operate in the harsh radiation and thermal environment of low Earth orbit.

The facility targets 2-nanometer process technology — the most advanced node in semiconductor manufacturing — and aims for 1 terawatt of annual AI compute capacity. For context, the entire global advanced chip industry currently produces roughly 20 gigawatts per year. Terafab alone would represent a 50x increase.

Small-batch production of AI5 chips is targeted for late 2026, with volume production in 2027. The D3 space chip timeline has not been publicly disclosed, though it is expected to follow closely behind the AI5 ramp.

Why Musk Says He Has No Choice

Musk framed Terafab not as a strategic expansion but as a survival imperative. He claimed that existing global fab capacity from TSMC, Samsung, and Micron produces only about 2% of what his companies will need within three to four years. 'We either build the Terafab or we don't have the chips,' he said during the announcement.

The bottleneck is driven primarily by SpaceX's orbital ambitions. In January 2026, SpaceX filed with the FCC for authorization to launch up to one million AI data center satellites into low Earth orbit. Each satellite — designated 'AI Sat Mini' — would be approximately 170 meters long, carry 100 kW of onboard computing power, and require custom radiation-hardened processors that no commercial foundry currently produces at scale.

The D3: A Space Chip Unlike Anything on the Market

The D3 processor represents a fundamentally different approach to space-grade semiconductors. Traditional radiation-hardened chips — produced by companies like BAE Systems, Microchip Technology, and Teledyne — are designed for reliability in small volumes. They are expensive, often generations behind commercial process nodes, and produced in quantities measured in thousands, not millions.

The D3 inverts that model. It is designed for mass production at the most advanced process node available, with radiation hardening built into the architecture rather than bolted on as a constraint. It is also designed to run hotter than conventional space processors, reducing the radiator mass and cooling systems needed on each satellite.

If the D3 reaches volume production, it would create an entirely new category of space processor — one optimized for commercial-scale orbital computing rather than the traditional defense and science missions that have defined the radiation-hardened chip market for decades.

What It Means for the Space Industry

Terafab's implications extend well beyond Musk's companies. The radiation-hardened semiconductor market was valued at approximately $1.74 billion in 2026 and growing at about 4.6% annually — a pace set by the relatively slow cadence of government satellite programs. A single facility producing millions of advanced space-grade chips per year would fundamentally alter the supply dynamics, pricing, and competitive landscape of that market.

For space startups, the effects could be transformative. Today, sourcing radiation-hardened processors is one of the most expensive and time-consuming parts of satellite development. If Terafab drives down the cost of space-grade compute, it could lower barriers to entry across the entire satellite industry — from Earth observation to in-orbit servicing to space-based manufacturing.

For incumbent rad-hard chip suppliers like BAE Systems, Microchip, and Teledyne, Terafab represents an existential question: can their low-volume, high-margin model survive if a high-volume competitor enters the market with chips that are both cheaper and more capable?

The Skeptics Have a Case

Musk has never manufactured semiconductors at scale. Analysts at Morgan Stanley have estimated the total cost of building meaningful capacity at $35–45 billion — significantly higher than Musk's stated range — with the earliest chip output likely in mid-2028 rather than late 2026.

Critics have drawn parallels to Tesla's 2020 Battery Day, where Musk promised revolutionary 4680 battery cells that would dramatically cut costs and increase range. In practice, the 4680 program required six to seven design revisions, took years longer than projected, and never achieved the promised cost reductions.

There is also the question of workforce. Advanced semiconductor fabrication requires thousands of specialized engineers and technicians. TSMC's Arizona fab has struggled with recruitment despite years of preparation. Building a team from scratch in Austin for the most advanced process node would be an unprecedented challenge.

The Bigger Picture: Vertical Integration Reaches Orbit

Terafab is best understood as the logical extension of Musk's vertical integration philosophy. Tesla builds its own batteries, motors, and software. SpaceX builds its own engines, avionics, and launch vehicles. Now, the combined enterprise will attempt to build its own chips — and do so at a scale that dwarfs the existing space semiconductor industry.

The strategic logic is clear: if SpaceX intends to deploy hundreds of thousands of AI satellites, each requiring custom radiation-hardened processors, relying on external suppliers who produce those chips in small batches at old process nodes is untenable. The question is whether Musk can execute on semiconductor manufacturing the way he has on rockets and electric vehicles — or whether chips prove to be a fundamentally different challenge.

For the space industry at large, Terafab is worth watching regardless of its ultimate success. The very attempt to mass-produce advanced space-grade semiconductors signals a shift in how the industry thinks about scale, supply chains, and the cost structure of building things for orbit. Even if Terafab delivers at half its stated ambition, it would still represent the largest single investment in space-grade semiconductor capacity in history.

What to Watch

• Whether AI5 small-batch production actually begins in late 2026 as stated, or slips into 2027–2028 as Morgan Stanley projects
• D3 space chip timeline — no production date has been disclosed, and the radiation-hardening requirements add significant complexity
• SpaceX's FCC application progress for the million-satellite orbital data center constellation
• Reaction from incumbent rad-hard chip suppliers (BAE, Microchip, Teledyne) and whether they accelerate their own roadmaps
• Whether any of Terafab's space-grade output becomes available to third-party satellite manufacturers, or remains exclusively for SpaceX

Frequently Asked Questions