In a historic step that is set to transform the semiconductor industry, Samsung has entered into a deal with Tesla to manufacture chips worth $16.5 billion. Confirmed by Elon Musk via an X post, the deal will see Samsung Foundry produce Tesla’s next-generation A16 (formerly HW6) chips at its upcoming fab in Taylor, Texas. The chips will play a crucial role in enabling the future of self-driving Teslas, as the AI5 processor is scheduled to launch in 2026. This is strategically huge, Musk said. Samsung is already making AI4. TSMC will produce AI5. But AI6 is ours with Samsung, and the fab is not very far away from my house; I will be there in person to push the development.”
The 10-year contract runs through 2033 and is a major win for Samsung, which has struggled to catch up with Taiwan’s TSMC in the foundry race. By signing this contract, Samsung will not only get a high-profile customer in Tesla but also increase the profile of its $17 billion Taylor fab, where these chips will be manufactured.
The AI6 SoC should be available by 2029, which is consistent with Tesla’s three-to-four-year chip refresh cycle. Although the details of the nodes used are not verified, it is believed that Samsung could use its 2nm-class SF2 or SF2A automotive-grade process to fulfill the performance and reliability demands of Tesla.
Tesla is said to have requested a production yield of 60 to 70 percent, a feasible goal when one has three years to prepare. Remarkably, Samsung gave Tesla engineers access to work on manufacturing optimization, which is not a typical degree of involvement of an automaker. Musk pointed to the practicality of the collaboration, stating that he will physically check the progress on the ground.
Samsung previously produced Tesla’s AI3 (HW3) chip on a 14 nm process and currently makes AI4 on 7 nm technology. TSMC will produce the AI5, most probably with its N3A or N4A node, and will divide production between Taiwan and Arizona. To Samsung Foundry, this is not just a $16.5 billion victory but a significant validation that it can play at the top of the AI and automotive chip design game.