Executives linked to the joint venture between Tesla and SpaceX have already looked for chip equipment suppliers, including Applied Materials, Tokyo Electron and Lam Research, to make the so-called Terafab viable. These are initial steps in an ambitious – and likely difficult – attempt to enter cutting-edge chip production.
Teams working on the Tesla–SpaceX partnership have sought price quotes and lead times for a range of semiconductor manufacturing equipment, people with knowledge of the matter said. In recent weeks, they have made contact with manufacturers of photomasks, substrates, recorders (“etchers“), deposition machines, cleaning equipment, testers and other tools, according to these sources, who requested anonymity because the discussions are private.
The Terafab team also asked Samsung, Musk’s current chip manufacturing partner, for support. The South Korean company, however, responded by proposing to reserve more capacity for Tesla at its planned factory in Taylor, Texas, the people said.
Continues after advertising
This move indicates that Musk is moving forward with Terafab despite skepticism from the semiconductor industry. The project, as conceived, aims to reshape the chip manufacturing sector and put the billionaire on the same playing field as giants like TSMC. Intel stated that it will join the initiative, and its CEO, Lip Bu Tan, even published a photo of Musk’s recent visit to the manufacturer’s office in Santa Clara.
Musk’s representatives have been charging very quick price estimates, offering little information about the products that would be manufactured. In one case, they asked for a quote on a holiday Friday to be delivered the following Monday, one of the people reported. Musk wants to move at “the speed of light,” this source said.
Shares of Tokyo Electron closed up 5.3% in Tokyo on Thursday, while news of the talks also helped boost shares of chip equipment companies such as Advantest, Screen Holdings and Disco. Shares of Applied Materials and Lam Research rose more than 2% in U.S. premarket trading.
The Terafab project — which aims to achieve the impressive goal of providing 1 terawatt of computing capacity per year — is Musk’s latest large-scale bet. Although Tesla designs its own autonomous driving (FSD) chips, the billionaire’s companies have never manufactured semiconductors. Still, he proposes producing chips at a scale that would surpass current global capacity, starting with a pilot line in Austin that should take advantage of Tesla’s existing electric vehicle factory and infrastructure.
The idea is that these chips will underpin the xAI artificial intelligence business, a line of humanoid robots and data centers in space — plans that many in the semiconductor industry don’t take entirely seriously. The final scale of the project, and whether it will be concentrated in a mega-complex or distributed across multiple units beyond Texas, has not yet been defined.
According to sources, the project has offered to pay an amount considerably above initial quotes if suppliers give priority to Terafab. There are no firm orders yet because it is not clear what technology will be used or where the chips will actually be manufactured. The first step, however, is to build a pilot line with the capacity to process 3,000 wafers per month, the people said. The goal is to start silicon manufacturing by 2029 and, from then on, scale production.
Continues after advertising
Bernstein analysts estimate the project would require somewhere between $5 trillion and $13 trillion in capital investment. Spokespeople for Applied Materials, Tokyo Electron, Lam Research and Samsung declined to comment. Representatives for Tesla, SpaceX and xAI also did not respond to requests for comment.
Musk — echoing concerns in part of Silicon Valley — has been saying that the semiconductor industry is not increasing supply quickly enough to meet demand for AI chips. Amazon.com Inc., Alphabet Inc. and other major cloud providers are expected to invest about $650 billion this year alone to expand data center infrastructure. This movement is already creating strong pressure on the memory chip market and is starting to spread to AI accelerators.
In March, Musk presented his plans for a large-scale operation focused on producing next-generation chips for artificial intelligence, robotics and space applications. With this, he intends to compete with TSMC, considered the best chipmaker in the world. An undertaking of this size involves hundreds of steps that cross multiple engineering disciplines and require collaboration from companies ranging from industrial gas suppliers to test equipment manufacturers.
Continues after advertising
CC Wei, CEO of TSMC, sent a message to Musk during an earnings call. “It takes two to three years to build a new factory — there is no shortcut — and it takes another year or two to increase production,” he said. “There are no shortcuts.”
Applied Materials, Tokyo Electron and Lam Research are key players in the global semiconductor market, currently valued at around US$1 trillion. They provide the equipment that TSMC and competitors use in advanced processes such as etching and deposition. The Dutch ASML Holding NV is perhaps the most critical link in this chain: it is the only manufacturer of extreme ultraviolet (EUV) lithography machines, essential for any company that wants to produce cutting-edge chips on a large scale. It is not yet known whether Musk’s team has already made contact with ASML. The company’s shares rose 3.1% in Amsterdam.
Tesla advisor Ira Ehrenpreis accompanied Musk on a visit to Intel headquarters this month. In addition to the expertise of the American chip manufacturer, Musk has been recruiting professionals with knowledge in various aspects of operating semiconductor factories — from chip design to energy management, construction and purchasing. In manufacturing, job offers were made to engineers at companies including Applied Materials, Samsung and TSMC, the people said.
Continues after advertising
Terafab aims to produce AI chips for devices such as robotaxis and Optimus humanoid robots. The plan also includes manufacturing high-power semiconductors for use in space by SpaceX and xAI. Musk said he expects xAI to consume most of these components.
Some analysts doubt Musk can achieve anything close to the scale he described for Terafab. Others remember that the businessman has already carried out projects that seemed impossible, such as creating a commercially viable rocket business with SpaceX and bringing electric vehicles to the mass market with Tesla.
Still, it is possible that Musk ends up pursuing more modest goals in the chip industry, given the costs and complexity involved. Hamburg-based bank Berenberg has not yet incorporated Terafab into its financial models for ASML, whose high-numerical aperture (High-NA) EUV lithography equipment would likely be needed on a project of this size, Tammy Qiu, head of equity technology research at the institution, told Bloomberg Tech.
Continues after advertising
“The intention is real,” she said, “but it’s not something that’s going to happen in a significant way for at least two years.” According to Qiu, Berenberg “has not yet included Terafab in the numbers”.
© 2026 Bloomberg L.P.
