BlockBeats News, July 24th, the AMD Advancing AI Conference was held on July 22-23 in San Francisco. At the conference, AMD CEO Lisa Su announced that Helios has entered full production and will soon begin shipping.

OpenAI's infrastructure lead stated that they plan to deploy AMD Helios at scale. OpenAI intends to collaborate with AMD to develop the MI500 series AI chips and future products.

Furthermore, Lisa Su mentioned that the company is partnering with chip design firm Cerebras to provide high-speed inference capabilities through Cerebras' cloud services. The collaborative product between AMD and Cerebras is set to be launched later this year. The AMD and Cerebras system will offer an AI inference solution combining the AMD Helios GPU server rack with Cerebras' wafer-scale chip.

CNBC analysis pointed out that a year ago, Lisa Su predicted the AI accelerator market to reach $500 billion by 2028. According to the latest forecast, by the end of this decade, this scale will roughly match the size of today's entire "semiconductor market." Lisa Su emphasized that GPUs will capture the majority of this market share.