Intel Unveils Latest AI Chip Amid Growing Competition with Nvidia
Intel has just announced its newest artificial intelligence chip, dubbed Gaudi 3, as semiconductor manufacturers race to develop chips capable of training and deploying large AI models, such as the one powering OpenAI’s ChatGPT.
According to Intel, the new Gaudi 3 chip boasts over twice the power efficiency and can run AI models one-and-a-half times faster than Nvidia’s H100 GPU. It comes in various configurations, including a bundle of eight Gaudi 3 chips on a single motherboard or a card that can be inserted into existing systems.
During testing, Intel evaluated the chip’s performance on models like Meta’s open-source Llama and Falcon, backed by Abu Dhabi. The Gaudi 3 chip proves beneficial for training or deploying models like Stable Diffusion and OpenAI’s Whisper model for speech recognition.
One of Intel’s key selling points is that its chips consume less power compared to Nvidia’s offerings.
While Nvidia currently dominates the AI chip market with its GPUs, which have been the preferred choice for AI developers, Intel aims to challenge this dominance with the Gaudi 3. The chip is set to be available to customers in the third quarter, with major companies like Dell, Hewlett Packard Enterprise, and Supermicro manufacturing systems equipped with the Gaudi 3 chips.
Das Kamhout, vice president of Xeon software at Intel, expressed confidence in the competitiveness of the Gaudi 3 chips compared to Nvidia’s latest offerings, emphasizing Intel’s competitive pricing and open integrated network on chip.
Despite Nvidia’s stronghold on the market, there is still room for other competitors in the data center AI market as cloud providers and businesses continue to invest in AI infrastructure. The high costs associated with running generative AI and purchasing Nvidia GPUs have led companies to seek alternative suppliers to reduce expenses.
Intel’s move to expand its presence in the AI chip market aligns with similar efforts by AMD, which recently introduced the MI300X data center GPU. Additionally, Nvidia has announced its B100 and B200 GPUs as successors to the H100, further intensifying competition in the AI chip space.
To compete with Nvidia’s proprietary software CUDA, Intel is collaborating with chip and software giants like Google, Qualcomm, and Arm to develop open software solutions that offer flexibility and interoperability across different hardware platforms.
With the Gaudi 3 chip built on a cutting-edge five nanometer process, Intel demonstrates its commitment to leveraging advanced manufacturing techniques to deliver high-performance AI chips. Moreover, Intel’s plans to manufacture AI chips at a new Ohio factory signal its intent to expand its manufacturing capabilities and potentially cater to external clients in the future.
In summary, Intel’s introduction of the Gaudi 3 chip marks a significant development in the AI chip market, signaling increased competition with Nvidia and offering customers more options for AI hardware solutions.
