Key facts
- Meta plans to begin production of its custom AI chip, 'Iris', in September.
- The chip is designed to improve AI powering Facebook and Instagram.
- Meta aims to double its computing capacity to 14 gigawatts by 2027.
- The company is working with Broadcom on design and TSMC on manufacturing.
- Meta expects to spend up to $145 billion on AI infrastructure this year.
Meta Platforms intends to commence manufacturing of its internally developed artificial intelligence chip, codenamed 'Iris', in September. This initiative is a key component of the company's strategy to significantly expand its overall computing power to 14 gigawatts by next year, as revealed in an internal memo reviewed by Reuters.
The 'Iris' chip is part of Meta's four-generation Training and Inference Accelerators (MTIA) project, with the company handling the chip's design in-house. The goal is to leverage custom-built silicon to enhance the artificial intelligence that underpins its social media platforms, Facebook and Instagram.
Testing of the chip concluded rapidly in six weeks, encountering no substantial issues, which suggests positive progress for an in-house effort that has faced challenges over the past five years. Meta is partnering with Broadcom for the chip's design and Taiwan Semiconductor Manufacturing Co for its production. This approach is expected to help Meta reduce its substantial computing expenses and increase its autonomy from major chip suppliers like Nvidia and AMD.
Meta unveiled 'Iris' in March, alongside three other AI processors, and plans to release new chips approximately every six months through 2027, a more frequent cycle than the typical industry release schedule of a year or more.
For the current year, Meta plans to deploy seven gigawatts of computing infrastructure, with a projection to double this capacity in 2027. The company anticipates investing as much as $145 billion in AI infrastructure this year, representing a considerable portion of the estimated $700 billion Big Tech collectively plans to spend on AI technology.
To support its data center expansion, Meta has secured multi-year supply agreements with companies including Samsung Electronics for memory chips, Sandisk for flash storage, and Sumitomo Electric for fiber-optic equipment. These long-term agreements are crucial amid a global memory chip shortage that has impacted other tech giants like Apple.
The surge in demand for components like memory and AI chips is driven by the race among tech companies to scale up data centers to meet the growing computational demands of artificial intelligence. Analysts at Morgan Stanley have noted that the rapid and significant rise in memory and chip prices has become a notable macroeconomic concern, termed 'chipflation'.