Microsoft’s custom AI chips could challenge Nvidia, which reached a $1trn valuation earlier this year thanks to its focus on this surging market.
Microsoft has unveiled its own custom silicon chips, as the company moves deeper into the growing AI market.
The tech giant unveiled two chips at the Microsoft Ignite event – the Azure Maia AI Accelerator and the Azure Cobalt CPU. The company said that the Azure Maia in particular is designed for AI tasks and generative AI.
Microsoft said these chips were designed using a silicon-testing method that the company has been secretly refining for years. The company also said these chips will meet the needs of customers that seek to take advantage of the latest cloud and AI breakthroughs.
“Microsoft is building the infrastructure to support AI innovation and we are reimagining every aspect of our data centres to meet the needs of our customers,” said Scott Guthrie, executive VP of Microsoft’s Cloud and AI group.
“At the scale we operate, it’s important for us to optimise and integrate every layer of the infrastructure stack to maximise performance, diversify our supply chain and give customers infrastructure choice.”
Microsoft said these chips will roll out in early 2024, initially going to the company’s data centres to boost its own AI services such as Microsoft CoPilot. The company also said it aims to create flexible chips that can be optimised for power, performance, sustainability or cost.
A new challenge for Nvidia
Microsoft has been a primary force in the recent surge in AI, being a major investor in ChatGPT-creator OpenAI. The tech giant has also integrated OpenAI technology into some of its own AI offerings.
But other companies have benefitted from the focus on AI, particularly Nvidia which has a dominant hold in the AI chips market. This focus on specialised chips helped Nvidia become the first chipmaker to be valued at $1trn in May.
In August, Nvidia reported record sales of $13.51bn in its second fiscal quarter of 2023. This was driven largely by its data centre business, which is responsible for the Nvidia A100 and H100 chips that are crucial for generative AI.
This is not the first attempt by Microsoft to break into this Nvidia-dominated market however. In September, Microsoft contributed to the $110m funding round of AI chip start-up d-Matrix, which aims to launch its own AI chip in 2024.
10 things you need to know direct to your inbox every weekday. Sign up for the Daily Brief, Silicon Republic’s digest of essential sci-tech news.