What If Musk Succeeds? Could His New Technology Upend the Semiconductor Industry for Samsung and SK Hynix? [Chiptalk]

Potential for Particle Accelerator Adoption... A Substitute for ASML Equipment
Revolutionary Cost Reduction and Resolution of 3D DRAM Bottlenecks Expected

If Commercialized, U.S. Memory Companies Like Micron Stand to Benefit
"Concerns

There is increasing speculation that Elon Musk, CEO of Tesla, may introduce a completely different type of lithography (the process of drawing circuits on a wafer) technology in his semiconductor manufacturing project “Terafab.” With the Dutch company ASML having a near-monopoly on the global supply of lithography equipment so far, the industry is closely watching Musk’s interest in novel technologies that could potentially disrupt this dominance. If this technology is commercialized successfully, it is expected to have a significant impact on South Korean memory chip companies.

Elon Musk, CEO of Tesla. Yonhap News Agency

Will Terafab Make Chips Without ASML?

As of August 29, according to semiconductor industry sources and foreign media, Terafab is increasingly likely to adopt a particle accelerator-based “free-electron laser (FEL)” lithography technology. Recently on X (formerly Twitter), CEO Musk gave a positive response to a post suggesting that “the central part of the Terafab building looks like a synchrotron (particle accelerator), possibly for developing FEL-based extreme ultraviolet (EUV) equipment,” replying, “FEL FTW (For The Win).”

Earlier, during Tesla’s earnings announcement in April, Musk had stated, “We plan to build a research semiconductor fab on the Texas campus,” adding, “There is new physics I want to try out.” Although Musk has not officially declared the adoption of FEL technology, the industry views his remarks as a sign that Terafab is considering using an FEL light source instead of the conventional ASML approach.

Terafab is a semiconductor manufacturing project led by Tesla and SpaceX, aiming to handle everything from design to manufacturing and packaging in one facility to produce the vast number of chips needed for vehicles, robots, and space data centers. Until now, Tesla has outsourced production of its in-house designed chips to external foundries such as TSMC or Samsung Electronics, but the goal is to switch to an in-house manufacturing system.

The highlight of the Terafab plan is the focus on a new lithography method. Lithography is a process that uses light to print intricate circuit patterns onto silicon wafers, which are the base material for semiconductors. The finer the circuits, the shorter the wavelength of light needed; thus, state-of-the-art chips require EUV light with a wavelength of 13.5 nanometers (nm). Currently, ASML is the world’s sole supplier of this EUV lithography equipment.

Existing ASML equipment generates light by firing lasers 50,000 times per second at tiny tin droplets, creating a plasma state momentarily. This process is highly challenging for ensuring a reliable light source, and furthermore, each machine costs between 500 billion and 600 billion won, making equipment investment a significant burden for chipmakers.

The FEL method being discussed for Terafab takes a radically different approach. Instead of tin plasma, it harnesses electron beams generated when electrons rapidly circulate inside a particle accelerator. With just one particle accelerator, the amount of light produced could match that of 16 to 20 EUV tools, eliminating the need for individual high-cost light source devices.

‘Double-Edged Sword’ for Korean Memory Companies

Experts believe that if particle accelerator-based FEL lithography technology succeeds, it could fundamentally disrupt the landscape of semiconductor manufacturing processes. First of all, the reduced costs of purchasing and maintaining expensive ASML equipment would ease initial capital investment burdens, and the per-wafer production cost would fall sharply, greatly improving profitability for semiconductor companies.

There are also high expectations that this technology could overcome technical limitations. If it enables an effectively unlimited supply of EUV light and allows free tuning of the wavelength, bottlenecks in the lithography process during the production of next-generation 3D DRAM or advanced logic chips could be resolved. Kyunghui Kwon, a research fellow at the Industrial Research Institute, analyzed, “If commercialized successfully, this will bring about a fundamental change to the process technology of the most crucial logic semiconductors.”

Some concerns have also been raised that this technology could pose risks for domestic memory chip makers such as Samsung Electronics and SK hynix. If particle accelerator-based lithography is commercialized in the United States and enables companies like Micron to massively increase output with drastically lower costs, the cost and yield advantages built up by Korean companies could be undermined. There is even the possibility that the competitive edge of the current EUV-based fab infrastructure — built with investments of trillions of won in areas including Yongin, Gyeonggi Province — may be relatively diminished.

Kwon further emphasized, “There is a strong likelihood that this could completely neutralize the process layout and fab design competitiveness held by Korean companies. We need to pay close attention to paradigm shifts in advanced lithography technology, both in terms of technology development and business planning.”

This content was produced with the assistance of AI translation services.