OpenAI to Partner with TSMC for its In-house 1.6nm AI Chip Development

Taiwan Semiconductor Manufacturing Company to Manufacture OpenAI's In-house AI Chip

OpenAI, a leader in the artificial intelligence industry, is reported to be manufacturing its first custom silicon chips with the Taiwan Semiconductor Manufacturing Company (TSMC), a trusted partner of technology giants like Nvidia, Apple, AMD, and Intel.

OpenAI Gears Up for Mass Production by 2026

TSMC's A16 process node will be used for the manufacturing, aiming to reach mass production by the second half of 2026. The advanced 1.6-nanometer manufacturing process is known to offer significant improvements in speed, power efficiency, and chip density compared to the 2nm-class node.

OpenAI and Apple: An emerging partnership

Apple, having been the first major customer to reserve A16 production capacity, announced a partnership with OpenAI. A new service named Apple Intelligence will see the integration of ChatGPT to iOS 18, iPadOS 18, and macOS Sequoia. Although Apple plans to deploy Private Cloud Compute on Apple Silicon, there's no current indication of OpenAI's silicon usage.

OpenAI Transitions from Chip Fabs to ASIC Production

Following reports of potential investment in chip fabs, OpenAI has opted for a more feasible route. They now aim to fabricate machine-learning accelerating ASICs with the assistance of Broadcom and Marvell, using a TSMC node. This practical approach mirrors successful collaborations, such as Google and Broadcom's partnership on TPUs.

Growth and Value of OpenAI

OpenAI is reportedly working towards a funding deal that could evaluate the company at $100 billion. It claims that majority of the Fortune 500 companies use OpenAI products, and ChatGPT usage has doubled since last November. Despite significant investments and pledges from Microsoft, OpenAI is likely to experience a financial dip due to substantial costs related to neural net training and staffing.