Microsoft is preparing to launch its first-ever artificial intelligence (AI) chip. This move is seen as an attempt to lessen the company’s reliance on Nvidia’s high-demand, limited-supply graphics processing unit (GPU) chips.
The new AI chip, named Athena, is specifically designed for data center servers. It is expected to compete directly with Nvidia’s flagship H100 GPU is currently used by Microsoft and other cloud service providers to power large language models (LLMs) and various other AI applications.
The Information first reported about Athena in April of this year, and the chip is set to make its official debut at Microsoft’s Ignite conference, scheduled for November 14-17.
This development comes amidst rising demand for AI chips, particularly due to the substantial computing power required for training and operating large language models (LLMs). This increased demand has led to a shortage of AI chips, resulting in escalating prices.
Furthermore, OpenAI, backed by Microsoft, is reportedly exploring the possibility of creating its own AI chips. This strategic move aims to reduce dependence on Nvidia and other chip manufacturers.
Beyond reducing dependency, the introduction of Athena also holds the potential to enhance cost efficiency and improve the performance of Microsoft’s cloud services.
This move by Microsoft aligns with similar efforts by Google and Amazon, who are also actively pursuing the development of their proprietary AI chips, indicating a rapid expansion in the AI chip market on the horizon.