In a significant technological stride, Microsoft has unveiled its own bespoke computing chips designed specifically for artificial intelligence (AI). This move sees the company joining the ranks of other tech behemoths in internalizing key technologies, a strategy driven by the escalating costs associated with delivering AI services.
The newly announced chip, christened “Athena,” is anticipated to expedite AI computing tasks and underpin Microsoft’s suite of services. This development comes in the wake of chip shortages that have inflated the prices of the highly coveted Nvidia AI GPUs.
The Athena project, which has reportedly been in the works since 2019, will see the initial version of the chip utilize Taiwan Semiconductor’s (TSMC) 5-nanometer process. The chips are engineered to train large-language models (LLMs) and similar software, including inferring from the data these models gather during training.
READ :UK Supreme Court Blocks Rwanda Asylum Deal
Currently, the Athena chips are being utilized by a select group of employees at Microsoft and OpenAI. While there are plans to potentially broaden the availability of these chips “as early as next year,” it remains uncertain whether they will be accessible to Azure customers.
The introduction of Athena could slash the cost per chip by a third compared to Nvidia’s offerings, potentially translating into substantial savings for the companny given the current estimated operational costs of AI services.
Microsoft’s foray into the development of its own AI chips mirrors similar initiatives undertaken by other tech giants such as Amazon, Google, and Meta. Despite collaborating with Nvidia on AI supercomputing and the Nvidia Omniverse cloud, Microsoft reportedly does not view Athena as a “broad replacement for Nvidia’s products”.