• Tesla plans to use Intel 14A chips for Terafab.
• Terafab targets large-scale AI chip production.

Tesla plans to use Intel’s next-generation 14A manufacturing process to produce chips for its proposed Terafab facility in Austin, according to Reuters, citing comments by Elon Musk. The site is intended to serve as an advanced AI chip complex supporting the company’s robotics and data infrastructure efforts.

Intel 14A follows the company’s 18A process as part of its angstrom-era manufacturing roadmap. The node builds on RibbonFET, a gate-all-around transistor design, and extends backside power delivery using PowerVia-derived technology, which routes power through the rear of the wafer to improve routing efficiency.

Intel is currently ramping production using its 18A process at its Arizona facilities as it prepares for future nodes like 14A.

The arrangement would represent the first major external customer for Intel’s 14A process. The company has previously stated it was in discussions with potential clients but had not identified any. Intel declined to comment on Musk’s remarks.

Musk said the company selected the process based on its expected maturity. “Given that by the time Terafab scales up, 14A will be probably fairly mature or ready for prime time,” he said, adding that “14A seems like the right move.”

Foundry competition and costs

Intel’s foundry business remains smaller than TSMC, which leads the market in third-party chip manufacturing by revenue and customer base. In the fourth quarter of 2025, TSMC reported US$33.7 billion in foundry revenue, compared with US$4.5 billion for Intel Foundry, according to industry estimates.

Intel Foundry primarily manufactures chips for its own products, including Xeon server processors, while also providing packaging services for external customers. The company is expanding its contract manufacturing operations and is seeking external clients for advanced nodes like 14A, which are designed for both internal and third-party chip production.

Its external customer base remains limited, with most production still tied to internal product demand.

Leading-edge chip manufacturing is currently concentrated among three companies – Intel, TSMC, and Samsung – with the ability to produce advanced logic chips. TSMC chief executive C. C. Wei said during a recent earnings call that the company views Intel as a “formidable competitor” in the foundry market.

Intel’s 14A process is also designed to incorporate High-NA extreme ultraviolet (EUV) lithography tools from ASML, which offer higher resolution patterning compared with current EUV systems.

ASML’s High-NA EUV systems are more expensive than current-generation tools, with unit costs estimated at around $380 million compared with about $235 million for existing EUV systems, contributing to higher per-wafer production costs.

The 14A node is expected to deliver improvements in performance per watt compared with 18A, while also reducing power consumption. Intel has said the higher cost of the process makes securing external customers necessary to justify development and production.

Shares in Intel rose 3.6% in extended trading following the announcement. Tesla’s stock moved slightly lower after hours despite the company increasing its capital expenditure plans tied to future projects.

Terafab scale and infrastructure

Intel joined Tesla and SpaceX earlier in April in relation to the Terafab project. The project is intended to manufacture processors for autonomous vehicles, humanoid robotics, and data centre systems.

Musk has outlined plans for two fabrication facilities in the Terafab complex. One is expected to support chips for vehicles and robotics, while the second is intended for space-based data centre infrastructure.

Key operational details remain undisclosed, including funding arrangements, equipment sourcing, and timelines for deployment. Musk has stated that the facility could eventually produce one terawatt of computing capacity annually, compared with roughly half a terawatt currently estimated in the US.

In this context, a terawatt of compute refers to overall processing capacity not electrical output. Achieving this level of compute would require large-scale semiconductor fabrication, advanced chip packaging, and deployment in data centre infrastructure.

Advanced packaging is used to integrate multiple chip components in a single system.

Leading-edge semiconductor fabrication facilities typically cost between $20 billion and $30 billion each, depending on capacity and tooling requirements. Analysts have estimated that building capacity at that scale would require between $5 trillion and $13 trillion in capital expenditure, based on figures from Bernstein.

Analyst views

Industry analysts said the agreement could support Intel’s process development. Ben Bajarin of Creative Strategies said early design partners can help refine leading-edge manufacturing technologies.

Bajarin said Intel’s 14A technology could “turn out to be a bigger deal for Intel than folks thought,” adding that having multiple early design partners helps refine leading-edge processes.

Seaport Research Partners analyst Jay Goldberg said securing a customer is more than project timing. “Having a customer is more important than the timing,” Goldberg said, adding that even supplying chips for Tesla’s current operations could represent meaningful production volumes. “It’s not equivalent to Apple or Nvidia… But it’s a real customer. It can be real volumes.”

Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and is co-located with other leading technology events, click here for more information.

AI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here.