OpenAI, known for its groundbreaking AI models, is reportedly making moves to enter the semiconductor manufacturing space. The company has been in talks with top chipmakers, including TSMC and Samsung, aiming to address its growing demand for AI chips. OpenAI CEO Sam Altman previously proposed a bold $7 trillion plan to build 36 wafer fabs and data centers to support AI development. However, TSMC viewed the proposal as overly ambitious and risky due to the immense costs involved.

Building wafer fabs is a costly and complex endeavor, with expenses running into billions of dollars. The construction, equipment, research, and operational costs associated with modern fabs are significant. For instance, Intel's Arizona fabs each cost around $15 billion, while Samsung’s Texas fab is projected at $25 billion. Additionally, regional factors, such as supply chains and talent availability, impact fab costs, with regions like Asia offering relatively lower costs compared to Europe and the U.S.

Despite scaling back its original plans, OpenAI is still pursuing custom AI silicon production. Reports indicate that OpenAI has secured production capacity at TSMC for its upcoming A16 process node, a cutting-edge 1.6-nanometer manufacturing process slated for mass production in 2026. This node promises enhanced speed, reduced power consumption, and increased chip density compared to current technologies.

While OpenAI initially explored building its own fabs, it has now opted for a more feasible approach by partnering with established players like Broadcom and Marvell to design custom machine-learning ASICs. These chips will be manufactured by TSMC on its 3nm and A16 process nodes, aligning OpenAI with a strategy commonly used by tech giants like Google, which partnered with Broadcom to design its TPUs.

OpenAI’s ambitions are backed by strong financial growth and usage metrics. The company is reportedly working on a funding deal that would value it at $100 billion. With more than 200 million weekly active users of ChatGPT and widespread adoption of its AI products by 92% of Fortune 500 companies, OpenAI's future in the semiconductor space seems promising—though still fraught with challenges.

While the AI-driven demand for chips surges, OpenAI is balancing its lofty goals with more grounded strategies, partnering with industry leaders to navigate the complex world of semiconductor manufacturing.