Intel introduces ‘Falcon Shores’ AI chip

Intel Corp (INTC.O) plans to introduce a new chip for artificial intelligence (AI) in 2025, as it shifts its strategy to compete against Nvidia Corp (NVDA.O) and Advanced Micro Devices Inc (AMD.O).

At a supercomputing conference, held in Germany on Monday, Intel provided details of its forthcoming “Falcon Shores” chip.

The chip is expected to have 288 gigabytes of memory and support 8-bit floating-point computation. Such technical specifications are making the chip suitable for large AI models like ChatGPT.

The release of the “Falcon Shores” chip is Intel’s strategy shift to compete with Nvidia, which led the market in chips for AI, and AMD which competed with Nvidia’s dominance with a chip called the MI300.

Previously, Intel struggled to gain market shares due to numerous delays in its Ponte Vecchio chip which was intended to compete with Nvidia.

Recently, Intel stated that it nearly completed shipments of the Ponte Vecchio-based Aurora supercomputer to Argonne National Lab. Intel claimed that Ponte Vecchio had better performance than Nvidia’s latest AI chip, the H100.

However, Intel’s forthcoming Falcon Shores chip won’t be available in the market until 2025, which gave Nvidia ample time to have another chip of its own.

Jeff McVeigh, interim head of Intel’s accelerated computing systems and graphics group, mentioned that the company is taking time to rework the chip after renouncing its previous strategy of combining graphics processing units (GPUs) with its central processing units (CPUs).

Jeff McVeigh discussed that Intel aimed to have the best CPU and the best GPU in the market, but it may not be feasible for one vendor to have the best combination of both.

He added, “If you have discrete offerings, that allows you at the platform level to choose both between the ratio as well as the vendors.” So, by offering discrete CPU and GPU options, Intel allows customers to choose the ratio and vendors that best suit their needs.