Musk's $25B Terafab Could Require Over 5,000 Plasma and Deposition Tools to Hit 2nm Production

By NineScrolls Team · 2026-03-24 · 6 min read · Industry

Contents

The Terafab Announcement: $25 Billion, One Roof

On March 21, 2026, Elon Musk officially launched Terafab — a joint semiconductor fabrication venture between Tesla, SpaceX, and xAI — with a price tag of $25 billion. The facility, planned for the North Campus of Giga Texas in Austin, aims to consolidate chip design, lithography, wafer fabrication, memory production, advanced packaging, and testing under a single roof. Musk claims it will be the largest chip manufacturing facility ever built.

Terafab targets 2-nanometer process technology with an initial capacity of 100,000 wafer starts per month, scaling to an ambitious 1 million wafer starts per month at full build-out. The fab will produce Tesla's AI5 inference chips for autonomous vehicles and Optimus robots, along with custom D3 processors designed for SpaceX's orbital AI satellite constellation. Small-batch AI5 production is targeted for late 2026, with volume output projected in 2027.

Musk justified the move by stating that current global foundry capacity covers only about 2% of what his companies collectively need. With TSMC's 2nm lines already fully booked through 2026, the supply constraint is real — even if Musk's solution is unprecedented in ambition.

What 2nm Demands: GAA Transistors and the Equipment Stack

Manufacturing at the 2nm node requires a fundamental shift in transistor architecture — from FinFET to Gate-All-Around (GAA) nanosheet structures. GAA transistors wrap the gate material completely around the channel on all four sides, demanding far more deposition and etch steps than previous nodes. The transition from 3nm FinFET to 2nm GAA adds hundreds of additional process steps per wafer, with the majority concentrated in plasma etch and thin film deposition.

Applied Materials reported in February 2026 that its Sym3 Z plasma etch platform has achieved tool-of-record status in 2nm logic manufacturing, with over 250 chambers deployed across leading foundries. Its newer Sym3 Z Magnum system introduces second-generation pulsed voltage technology that eliminates traditional tradeoffs between ion directionality and plasma uniformity — a critical requirement for etching the ultra-narrow nanosheet channels in GAA devices.

On the deposition side, ASM International has seen ALD account for more than half of its equipment revenue, driven directly by GAA adoption. The company estimates its served available market expands by $400 million in the transition from 3nm FinFET to 2nm GAA alone, with roughly 60% of ALD demand at advanced nodes now coming from front-end-of-line transistor fabrication.

Plasma Processing and Thin Film Deposition at Scale

A 2nm fab running 100,000 wafer starts per month would require thousands of individual process chambers. Plasma etch tools — used for patterning gate structures, nanosheet release, contact hole formation, and back-end interconnect definition — represent the single largest equipment category by chamber count. At 2nm, each wafer passes through plasma etch steps dozens of times, with critical applications including atomic layer etching (ALE) for nanosheet channel release and high-aspect-ratio contact etch using advanced pulsed plasma sources.

Thin film deposition is equally intensive. GAA architectures require atomic layer deposition (ALD) for high-k gate dielectrics, work function metal stacking, and liner/barrier layers in interconnects. Chemical vapor deposition (CVD) and plasma-enhanced CVD (PECVD) handle interlayer dielectrics, silicon nitride spacers, and low-k films. Physical vapor deposition (PVD/sputtering) remains essential for metal seed layers and copper interconnect barriers. Each wafer at the 2nm node accumulates over 100 distinct deposition steps.

Scaling to Musk's stated goal of 1 million wafer starts per month would multiply these equipment requirements by an order of magnitude — a procurement challenge that dwarfs anything the industry has seen, including TSMC's aggressive multi-fab 2nm expansion currently underway in Taiwan and Arizona.

Equipment Supply Chain Under Pressure

Even without Terafab, the semiconductor equipment supply chain is already strained. TSMC has begun volume production of 2nm chips and has fully booked its initial two N2 fabs, with three additional facilities under construction. Samsung and Intel are also pushing their own GAA roadmaps. All of this drives intense demand for the same critical tools Terafab would need.

The key equipment suppliers for a 2nm fab include ASML for EUV lithography systems (where it holds 100% market share), Applied Materials for CVD, PVD, etch, and inspection tools (roughly 30% of the overall equipment market), Lam Research for advanced plasma etch and deposition systems, Tokyo Electron for thermal processing and coater/developer tools, and ASM International for ALD systems. KLA Corporation would supply the process control and metrology equipment critical at these geometries.

Lead times for EUV lithography systems alone run 18–24 months. A greenfield 2nm fab of Terafab's proposed scale would likely require 50 or more EUV systems — representing a significant share of ASML's total annual output. Plasma etch and deposition tools have shorter but still substantial lead times, and the sheer volume Terafab would need could create allocation conflicts with existing foundry customers.

Can Musk Actually Build a Leading-Edge Fab?

No company has successfully built a leading-edge foundry from scratch in decades. TSMC, Samsung, and Intel have spent tens of billions over many years refining their 2nm processes, supported by deep institutional knowledge in plasma chemistry, deposition film stress management, and etch profile control. The process know-how required to achieve competitive yields at 2nm is arguably more valuable than the equipment itself.

For the semiconductor equipment industry, however, the signal matters regardless of Terafab's ultimate success. Musk's announcement confirms that demand for advanced manufacturing capacity — and the plasma processing, thin film deposition, and metrology equipment it requires — continues to outstrip supply. Whether Terafab ships its first wafer on schedule or not, the equipment orders it generates will flow through the same supply chain that serves every other advanced fab on the planet: plasma sources, sputtering targets, vacuum pumps, mass flow controllers, RF generators, and process monitoring systems.

The message for the equipment supply chain is clear: the customer base for 2nm-class tools is expanding beyond the traditional big three foundries, and capacity planning needs to account for new entrants with very deep pockets.

Sources