Microsoft is reportedly planning to release the company’s first chip designed for artificial intelligence (AI) at its annual developers’ conference ‘Ignite’ in Seattle next month, reported The Information, citing a person with direct knowledge of the matter.
Microsoft To Debut Its First AI Chip Next Month
The move by the Redmond giant is seen as an effort to reduce its reliance on NVIDIA’s graphics processing unit (GPU) chips, which have faced supply shortages due to increased demand. Besides the limited availability, the dependence on NVIDIA chips has also proven challenging due to high costs.
In April, The Information reported that Microsoft was secretly working on a new in-house chip internally codenamed “Athena” since 2019, with about 300 employees dedicated to the project.
This chip is designed to be on par with NVIDIA’s H100 GPU, one of the most valuable resources in the current technology field, as they are essential for training large language models (LLMs) such as ChatGPT.
The company has publicly stated that it will continue to purchase NVIDIA GPUs. It also added that it doesn’t see its own in-house chips as a “broad replacement for NVIDIA’s products” but instead looks at them as an enhancement to Microsoft’s existing infrastructure.
Microsoft is currently using NVIDIA’s H100 GPUs to power its data center servers that take care of its AI services like Bing Chat AI chatbot, the Bing Image Creator art generator, and the recently released AI feature, Copilot.
The in-house Microsoft chips are aimed for use in data center servers and training of large language models for cloud customers, as well as AI capabilities in the company’s productivity apps, and streamline operations while cutting costs.
If Microsoft launches its own AI chip, it will put the company in the same ranks of well-known tech behemoths, such as Amazon, Alphabet (Google), and Meta (Facebook), who have built their own in-house chips for their AI projects.
Keep watching this space for more updates, as this is a developing story!
