Intel Joins Musk Terafab AI Chip Project in Texas

AUSTIN, TEXAS — Intel said on Tuesday it is joining Elon Musk's Terafab AI chip complex, adding its semiconductor manufacturing expertise to a project already backed by Tesla, SpaceX, and xAI. The chipmaker's shares rose on the announcement, which pairs one of America's most established foundries with Musk's vertically integrated vision for AI hardware.

What Intel Brings to the Table | 18A Process and 3D Packaging

The Terafab project, unveiled on March 21 at the historic Seaholm Power Plant in Austin, Texas, is a joint venture estimated to cost between $20 billion and $25 billion. It aims to consolidate chip design, lithography, fabrication, memory production, advanced packaging, and testing under one roof, with a stated goal of producing one terawatt of computing power per year. ObjectWire covered the original Terafab announcement when Musk first set the March 21 launch date.

Intel's entry gives the venture something it lacked: proven process technology and advanced packaging at scale. The company operates Fab 9 in New Mexico, the only U.S. factory producing high-volume 3D packaging using Intel's Foveros technology. Foveros allows chiplets to be stacked vertically rather than laid side by side, increasing transistor density and reducing latency between logic and memory dies. For an AI chip complex targeting terawatt-scale output, that capability is not optional, it is a prerequisite.

Intel also brings its 18A process node, which entered high-volume manufacturing earlier this year. At roughly 1.8 nanometers, 18A uses backside power delivery and gate-all-around transistors, two features that improve power efficiency and performance density simultaneously. No other American foundry currently offers both in production. TSMC and Samsung are working on comparable nodes, but neither operates a facility on U.S. soil capable of volume manufacturing at this level.

Terafab Venture Structure | Who Owns What in the $25B Complex

The financial terms of Intel's participation have not been publicly disclosed. What is known is that Terafab was originally structured as a joint venture between Tesla, SpaceX, and xAI, with Musk providing the site and the capital coordination. The venture's stated purpose is to build a single integrated campus that handles every step from chip design to finished, packaged silicon.

Partner	Contribution
Tesla	Capital coordination, Dojo chip architecture, FSD inference demand
SpaceX	Radiation-hardened chip design expertise, satellite compute demand
xAI	Grok model training, high-bandwidth memory requirements
Intel	18A process node, Foveros 3D packaging, Fab 9 (NM) expertise

Terafab joint venture partner contributions as of April 2026

Intel's involvement shifts the dynamic. Previously, analysts questioned whether Musk's companies had the deep manufacturing experience needed to operate a leading-edge fab from scratch. Building a chip is one challenge; running a fab at competitive yields on a bleeding-edge node is a different discipline entirely. Intel has been doing it for five decades.

Intel Stock Reaction | Shares Rise on Terafab Confirmation

Intel's stock moved higher on the news, reflecting investor confidence that the partnership validates the company's foundry pivot. Under CEO Lip-Bu Tan, Intel has aggressively courted external customers for its Intel Foundry Services (IFS) division, but landmark wins have been difficult to secure. Intel's broader transformation has hinged on proving that IFS can compete with TSMC for high-profile clients. Landing a role inside Terafab, alongside three Musk-backed entities with enormous compute demand, is exactly the kind of anchor customer IFS has been pursuing.

The market also sees strategic upside. If Terafab achieves its production targets, Intel would gain a high-profile reference customer, access to next-generation AI packaging requirements, and a co-development relationship with xAI, whose Grok large language model already consumes substantial GPU and HBM capacity.

Why Musk Needs Intel | Filling the Foundry Gap

Musk's existing companies design chips but do not manufacture them. Tesla's Full Self-Driving (FSD) hardware and Dojo training chips are fabbed by Samsung and TSMC. SpaceX contracts out its radiation-hardened processors. xAI relies on Nvidia GPUs for Grok training and inference. None of these companies have in-house lithography, etching, deposition, or packaging lines.

That dependency is precisely what Terafab is designed to eliminate. But eliminating it requires a partner that already knows how to run EUV lithography tools, manage defect density at sub-2nm geometries, and ramp production yield to commercially viable levels. Intel is, along with TSMC and Samsung, one of only three companies on earth with that capability. It is the only one headquartered in the United States.

Geopolitical Context | U.S. Semiconductor Sovereignty at Stake

The partnership arrives against a backdrop of intensifying U.S. efforts to reshore semiconductor manufacturing. The CHIPS and Science Act allocated $52.7 billion in subsidies to domestic fab construction, with Intel receiving $8.5 billion in direct funding. TSMC has committed to building fabs in Arizona, and Samsung is expanding in Texas, but both are foreign-headquartered companies subject to their home governments' export controls.

A Musk-Intel venture producing advanced AI chips entirely on American soil, from design through packaging, would represent a step beyond what any single CHIPS Act project has achieved. It would also create a domestic supply chain for military and intelligence applications that currently depend on Taiwanese and South Korean foundries.

What Comes Next | Timelines and Open Questions

Neither Intel nor the Terafab venture has announced production timelines tied to Intel's involvement. The original Terafab roadmap called for initial wafer starts by late 2027, with volume production in 2028. Intel's participation could accelerate that schedule, given that the company already has a trained workforce and operational packaging lines that could be replicated or expanded in Austin.

Open questions remain. How will intellectual property be shared between Intel's foundry customers and the Musk-affiliated entities? Will Intel manufacture Terafab-designed chips exclusively, or will it also produce its own products at the Austin campus? And how will the venture interact with Intel's existing CHIPS Act obligations, which tie federal subsidies to specific facilities and production commitments?

For now, the signal is clear: Musk is assembling the manufacturing capability to match his AI ambitions, and Intel is betting that its foundry future runs through partnerships exactly like this one.