Graphics chipmaker Nvidia (NASDAQ: NVDA) has unveiled a new AI processor designed to power the next generation of artificial intelligence systems, as competition heats up with rivals in the burgeoning generative AI sector.
Nvidia revealed its new chip builds on the existing GH200 Grace Hopper architecture and integrates the company’s groundbreaking HBM3e memory technology. Tailored for generative AI and accelerated computing applications, the new chip provides three times more memory capacity and bandwidth than prior chips.
“To meet surging demand for generative AI, data centers require accelerated computing platforms with specialized needs,” said Nvidia CEO Jensen Huang. “The new GH200 Grace Hopper Superchip delivers this with exceptional memory technology and bandwidth to improve throughput.”
Operating in a dual-chip configuration, the chip can achieve up to 1.2TB of high-speed memory by linking additional units via Nvidia’s proprietary NVLink technology. Huang stated this will enable developers to “deploy the giant models used for generative AI.”
Designed to “scale out across data centers,” Nvidia claims the chip significantly reduces the cost of running large language models. Sampling is expected by year’s end, with full availability in Q2 2024.
The new chip highlights Nvidia’s aggressive investments in the booming generative AI sector, which has also seen the company debut an AI supercomputer to aid developers in building their own models. Surging AI demand recently propelled Nvidia’s market cap above $1 trillion.
However, Nvidia faces mounting competition, controlling 80% of the AI chip market. Rival AMD recently launched a new AI chip seen as a potential challenger to Nvidia’s dominance. The rise of AI has also led some former crypto mining firms to rebrand towards generative AI.
With AI adoption exploding, the race is on between semiconductor firms to power the next iteration of AI systems. Nvidia aims to maintain its lead, but competitors are angling to carve out share in this high-growth arena.
