eWEEK content and product recommendations are editorially independent. We may make money when you click on links to our partners. Learn More.
Silicon Valley startup d-Matrix has received $160 million in funding, including an investment from Microsoft’s venture capital arm MSFT.O, for a new AI chip it expects to deliver by next year. Based in Santa Clara, CA, and backed by the tech giant Microsoft, d-Matrix has invited customers to test its new AI inference hardware, Corsair, which is focused on generative models’ interactions with end users. The chip is designed to meet the requirements of AI end-user service delivery, such as chatbots and video generators.
Nvidia has been dominating the AI chips market in recent years. Rather than competing, d-Matrix plans to complement the tech giant with its focus on the requirements of system end users. Nvidia’s chips are used to train AI systems that handle huge amounts of data. d-Matrix designed its chips to meet the demands and requests of users seeking to generate more responses from AI. Designed to meet high-volume user requests, Corsair is ideal in artificial intelligence applications with multiple users requesting simultaneous system responses—especially with customized outputs.
The Future is Video
d-Matrix raised $110 million last year at a time when chip companies struggled to attract new investments. This year, co-founders Sid Sheth and Sudeep Bhoja led the company to promote its latest chip, which offers high performance, efficiency, and scalability.
“We are getting a lot of interest in video use cases where we have customers coming and saying, ‘Hey, look, we want to generate videos, and we want a collection of users, all interacting with their own respective video,’” CEO Sid Sheth said.
Corsair promises to address energy efficiency, speed in token generation, costs, and scalability challenges of AI inference. It offers the potential to revolutionize how generative AI handles data in complex workloads and requirements. d-Matrix was one of the first companies to use a chiplet-based architecture in its chip, highlighting flexibility and scalability while lowering costs, and it can be installed and integrated into data center servers, making generative AI commercially viable.