Intel has made a smart move positioning its new Intel Gaudi 3 AI accelerator and Intel Xeon 6 processor as part of its new open systems strategy for accelerating the development of Enterprise AI. While the major generative AI players are scouring the darkest corners of the internet for the last dregs of “publicly available” data to feed their models, the semiconductor giant is showing that it is focused on enabling businesses to unlock the value of their proprietary internal data safe from external predators.
The performance numbers touted by Intel for the Gaudi 3 AI accelerator look reasonably competitive, delivering on average 50% better inference and 40% better power efficiency than the Nvidia H100 at a significantly lower cost.
The availability of Gaudi 3 systems from major original equipment manufacturers (OEMs) like Dell Technologies, Hewlett Packard Enterprise, Lenovo, and Supermicro is just as significant because of their strong presence in the global enterprise market.
Intel’s plans to create an open platform for enterprise AI with SAP, RedHat, VMware, and others together with the leadership role it is taking in the Ultra Ethernet Consortium (UEC) to promote open Ethernet networking for AI fabrics should further strengthen the supporting ecosystem for generative AI system deployments.
Even as Nvidia continues to dominate the overall AI server market, Intel has an excellent opportunity to carve out a strong position for itself in the enterprise space with the Gaudi 3 AI accelerator and open systems strategy. After its recent struggles, the company is now moving in the right direction and beginning to build momentum.
Long time technology industry fan here in Taiwan.
