Microsoft is set to introduce its first custom AI chip at the upcoming Ignite 2023 event. This live event is scheduled to take place from November 14 to 17 in Seattle, USA.
The company has announced a comprehensive agenda for the event, which includes various sessions including a live session showcasing its latest AI breakthroughs.
This chip, codenamed “Athena”, is designed to reduce the company’s reliance on NVIDIA GPUs for data center servers used in the training and running of large language models.
This move is aimed at reducing costs and strengthening Microsoft’s Azure cloud computing platform. Microsoft has been secretly developing these chips since 2019, with the potential to support OpenAI in their GPU needs.
OpenAI is also researching developing its own AI chip or acquiring a chipmaker to make custom chips. This development is due to the global chip shortage and the high cost of GPUs. Also at this year’s Ignite event, NVIDIA plans to launch the GH200 chip in 2024 with three times the memory of the H100.
Since forging its partnership with OpenAI through a significant multi-billion dollar investment, Microsoft has focused on AI technology and its continued advancements. A recent report from The Information now reveals that Microsoft could launch its first dedicated artificial intelligence chip in Ignite 2023.
As AI continues to emerge, more and more organizations are adopting the technology, leading to greater reliance on NVIDIA for the supply of GPUs, which are used in the development and training of large language models. like Bing Chat and ChatGPT.
However, NVIDIA is currently struggling with overwhelming demand for GPUs, leading to supply shortages. Microsoft is well aware of this challenge and has been diligently working to develop its AI chip for many years.
NVIDIA’s inability to meet soaring GPU demand could hinder Microsoft’s progress in AI, and could limit the company’s potential revenue from the technology.
Information from The Information also emphasizes that Microsoft intends to maintain its procurement of NVIDIA GPUs, while its AI chips will be used in servers in data centers, to help enhance AI capabilities. throughout your suite of productivity apps.
Incorporating its custom AI chip could help boost Microsoft’s profits in AI technology, allowing the company to compete more effectively with industry giants like Amazon and Google, which have developed develop its own AI chip.