Skip to main content

Meta's Custom AI Chip Iris Set for September Mass Production with Broadcom and TSMC

Meta is making a bold move in the AI hardware race. The company is collaborating with Broadcom and TSMC to develop a custom-designed processor for its data centers, codenamed "Iris." According to Reuters and internal memos, the chip is slated to enter mass production in September 2026.

Iris is part of Meta's MTIA (Meta Training and Inference Accelerator) project. It passed initial vulnerability testing within just six weeks. Its main job? To supercharge the core content ranking, recommendation systems, and generative algorithms that power Facebook and Instagram. By handling these tasks more efficiently, Meta hopes to slash data center costs and ease its dependence on expensive third-party GPUs from companies like NVIDIA and AMD.

Image

But Meta isn't stopping at chip design. With global hardware shortages and rising costs—what some call "chip inflation"—the company has locked in long-term supply deals with Samsung Electronics, SanDisk, and Sumitomo Electric. These agreements are meant to keep its supply chain stable and secure.

Looking ahead, Meta has ambitious plans for computing power. By the end of 2026, it aims to deploy about 7 gigawatts of computing capacity. And by 2027, that number is expected to double to 14 gigawatts. To support this growth, Meta's infrastructure capital expenditure for 2026 is set at a whopping $14.5 billion.

The launch of Iris marks a turning point for Meta in the trillion-dollar computing power arms race among tech giants. For now, Iris will complement—not replace—existing GPUs from NVIDIA and AMD. But Meta's aggressive self-development cycle, with new chip iterations every six months, gives it more control over its hardware deployment. On an industry level, this move helps Meta avoid supply chain bottlenecks and build an independent AI ecosystem that isn't tied to external software or hardware constraints.

Key Points

  • Meta partners with Broadcom and TSMC to produce its custom AI chip Iris, mass production starts September 2026.
  • Iris optimizes content ranking and recommendations on Facebook and Instagram, reducing data center costs and reliance on third-party GPUs.
  • Meta secures long-term supply contracts with Samsung, SanDisk, and Sumitomo Electric to ensure supply chain stability.
  • Meta plans to deploy 7 gigawatts of computing capacity by end of 2026, doubling to 14 gigawatts by 2027, with $14.5 billion in infrastructure spending.
  • Iris is a complement to existing GPUs, but Meta's rapid self-development cycle enhances hardware autonomy and supports an independent AI ecosystem.