Intel recently compared its Gaudi 3 performance against Nvidia's H100 accelerator in a press release and white paper. Intel claimed that Gaudi 3's better throughput and inference are due to its large high-bandwidth memory, more efficient architecture, and higher HBM capacity.

However, Nvidia is soon set to release the Blackwell platform, which is four times faster than the H100. The Blackwell platform is said to run trillion-parameter large language models at significantly reduced costs and energy needs compared to the H100.

Nvidia's advantage in the AI race is further solidified by its CUDA software platform, offering developer tools and libraries for AI application development. This advantage poses a challenge for competitors like Intel, aiming to match CUDA's capabilities with an open-source platform.

Despite Intel's cost advantage, the generative AI market is flooded with investments seeking sustainable AI advancements. In this scenario, the focus is on computing capabilities, speed, and capacity rather than pricing among Nvidia, Intel, and Advanced Micro Devices.

Dethroning Nvidia in the AI industry will be a tough feat for Intel. Nvidia has been investing in AI technology for years, offers the CUDA software platform, and dominates the revenue in this sector. While Intel may attract revenue with Gaudi 3, a complete shift in AI leadership from Nvidia to Intel is unlikely. Nvidia investors should remain confident in Nvidia's position in the market.