OpenAI to Build In-House Chips Using TSMC’s Most Advanced A16 Process:
OpenAI to Build In-House Chips Using TSMC’s Most Advanced A16 Process


OpenAI is taking a significant step towards developing its own custom AI chips using Taiwan Semiconductor Manufacturing Company's (TSMC) upcoming A16 process, a cutting-edge 1.6nm technology scheduled to begin mass production in 2026. The decision is part of OpenAI's strategy to reduce its reliance on third-party hardware, such as Nvidia's expensive AI servers, by developing customized in-house solutions  to optimize its AI models. OpenAI plans to work with US companies Broadcom and Marvell to leverage TSMC's advanced semiconductor technology to improve performance, power efficiency and chip density. The effort is seen as a long-term investment aimed at improving the capabilities of AI applications, including potential improvements in video generation using OpenAI's Sora model. In addition, there is also speculation about a possible partnership between OpenAI and Apple, as both companies are early customers for TSMC's A16 process. This could lead to further integration of AI technology into Apple devices, although no formal partnership has yet been confirmed. The project highlights the growing importance of custom hardware in the AI industry and underscores OpenAI's commitment to remaining at the forefront of technological advancements. Sources: - Fusion Chat Source All About AI Cryptopolitan.