Meta Begins Production of Iris AI Chips Amid Component Shortage

• •Meta to produce Iris AI chips to enhance computing capacity.
• •Chips designed for evolving AI workloads using modular architecture.
• •Production will be managed by TSMC, with significant investment in AI infrastructure.

Meta Platforms is set to start manufacturing its Iris AI chips in September 2026 as part of its Meta Training and Inference Accelerator (MTIA) program. This initiative aims to enhance computing capacity and reduce reliance on external GPU providers like Nvidia and AMD. The Iris chips utilize a modular architecture designed to adapt to evolving AI workloads, with production handled by Taiwan Semiconductor Manufacturing Company (TSMC). Meta has been investing heavily in its AI infrastructure, with capital expenditures projected between $125 billion and $145 billion for the year, focusing on securing compute capacity for its AI models. The company has previously deployed hundreds of thousands of MTIA chips and aims to double its compute capacity with the new Iris chips. The chips are expected to play a crucial role in powering Meta's AI-driven services, including content recommendations and generative AI applications.