Google discusses producing part of its new AI chip at Samsung factories. Is this a threat to TSMC?

Google is in talks with Korean tech giant Samsung to produce part of its next generation of processors for artificial intelligence, The Information reports. According to the publication, Samsung may produce one component of the new Icefish chip using 2nm technology. Google is one of the leaders in manufacturing its own AI processors.

Samsung is being considered as the manufacturer of the component that links the new chip to memory, according to sources The Information quoted by Reuters. Development of Icefish is ongoing, and mass production could conceivably begin as early as 2028, The Information claims. Google plans to outsource the production of the chip itself to Taiwan's TSMC, it said.

Samsung Electronics declined to comment to Reuters. Alphabet, Google's holding company, did not respond to the agency's request for comment. Reuters notes that it could not independently confirm The Information's information.

For Samsung, a possible contract with Google would be one of its biggest successes in growing its semiconductor contract manufacturing business, Reuters notes. The company has been pushing its own 2nm process, which allows it to fit more transistors on a smaller footprint and improve chip performance and energy efficiency.

The Information also reported on June 8 that Google is in talks with Intel to produce more than 3 million processors in 2028. This could indicate the company's desire to expand its range of manufacturing partners and depend less on TSMC, Reuters noted.

In April 2026, Samsung said it was looking to attract more customers for its advanced process products and was considering building a second facility in Texas to expand capacity. In particular, talks on possible chip production were held with Apple, Bloomberg reported.

In July 2025, the Korean company signed a $16.5 billion contract with Tesla to produce AI chips using 2nm technology, recalls Reuters.

Google is actively developing its own AI processors as an alternative to Nvidia's solutions, which currently dominate the AI chip market. In April, the company unveiled two new specialized chips for model training and inference, and sales of AI processors are becoming one of the drivers of growth in Google's cloud business, Reuters concludes.

This article was AI-translated and verified by a human editor