- Intel introduced its latest AI chip, Gaudi 3, surpassing Nvidia’s H100 GPU in power efficiency and speed.
- The chip underwent testing on various models and is set to compete with Nvidia in the third quarter.
- Additionally, Intel plans to collaborate with major industry players to develop open software solutions, while also expanding its manufacturing capabilities in Ohio.
On Tuesday, Intel unveiled its latest AI chip, Gaudi 3, amidst a rush among chipmakers to develop semiconductors for training and deploying large AI models, such as those powering OpenAI’s ChatGPT. Intel claims that the new Gaudi 3 chip is more than twice as power-efficient and runs AI models 1.5 times faster than Nvidia’s H100 GPU. It offers various configurations, including a bundle of eight chips on one motherboard or a card compatible with existing systems.
Also read: Intel discloses US$7 billion operating loss for chipmaking unit
Also read: Biden to give $8.5billion to Intel to build more chip factories
Intel asserts that its chips consume less power than Nvidia’s
Intel tested the chip on models like Meta’s open-source Llama and the Abu Dhabi-backed Falcon, stating that Gaudi 3 can assist in training or deploying models, including Stable Diffusion or OpenAI’s Whisper model for speech recognition. Intel asserts that its chips consume less power than Nvidia’s.
Despite Nvidia’s dominance in the AI chip market, accounting for approximately 80%, Intel announced that Gaudi 3 chips would be available to customers in the third quarter, with Dell, HP, and Supermicro among the companies planning to incorporate them into their systems. Das Kamhout, Intel’s vice president of Xeon software, expressed confidence in Gaudi 3’s competitiveness against Nvidia’s latest offerings, citing Intel’s competitive pricing and distinctive open integrated network on chip.
The data centre AI market is poised for growth
The data centre AI market is poised for growth as cloud providers and businesses invest in AI software infrastructure, presenting opportunities for competitors. With Nvidia’s stock tripling in the past year, and Intel’s up by 18%, AMD is also expanding its AI chip offerings, introducing the MI300X data centre GPU. Nvidia recently unveiled its B100 and B200 GPUs, successors to the H100, promising performance gains and scheduled for release later this year.
Nvidia’s success is attributed to CUDA
Nvidia’s success is attributed to CUDA, its proprietary software enabling comprehensive access to GPU hardware features for AI scientists. Intel is collaborating with industry giants like Google, Qualcomm, and Arm to develop open software solutions, providing flexibility for software companies in selecting chip providers. Gaudi 3 is built on a five-nanometer process, indicating Intel’s adoption of cutting-edge manufacturing techniques.
Additionally, Intel plans to manufacture AI chips, potentially for external companies, at a new Ohio factory expected to open in 2027 or 2028, as revealed by CEO Patrick Gelsinger.
