The Instinct MI300X, AMD’s newest AI chip, will be used, according to Meta, OpenAI, and Microsoft during a Wednesday AMD investor event. As far as technology businesses are concerned, this is the most significant indication that they are looking for alternatives to the pricey Nvidia graphics processors that have been necessary for developing and implementing artificial intelligence applications like OpenAI’s ChatGPT.
If AMD’s most recent high-end processor meets the needs of cloud service providers and tech companies for creating and delivering AI models when it begins shipping early in the next year, it may reduce the cost of creating AI models and exert competitive pressure on Nvidia’s rapidly increasing sales of AI chips.
“huge iron and huge GPUs for the cloud are all the rage,” AMD CEO Lisa Su stated on Wednesday.
The MI300X, according to AMD, is built on a new architecture, which frequently results in notable performance increases. Its 192GB of HBM3, a state-of-the-art, high-performance memory type that transports data more quickly and can accommodate larger AI models, is what makes it stand out the most.
Su explicitly contrasted Nvidia’s flagship AI GPU, the H100, with the MI300X and the systems constructed around it.
“This performance simply translates into an improved user experience,” Su stated. “You want a model to respond to your questions as quickly as possible, especially when the responses become more intricate.”
AMD is primarily concerned with whether businesses that have been developing with Nvidia will devote the necessary time and resources to adding a different GPU supplier. Adopting AMD requires work, according to Su.
On Wednesday, AMD informed partners and investors that it had made improvements to its ROCm software suite, which is now competitive with Nvidia’s industry-standard CUDA software. This improvement addressed a significant flaw that was one of the main reasons AI developers now favour Nvidia.
Cost will also be a factor. Su informed reporters that AMD’s processor would need to be less expensive to buy and run than Nvidia’s in order to entice customers to buy it. AMD did not disclose price for the MI300X on Wednesday, but Nvidia’s chips may run up to $40,000.
Source (CNBC)
