The goal is platform neutrality, although Intel still expects to sell its HPC solutions.
Earlier this month Intel partnered with investment firm DigitalBridge to form a software spin-off specializing in the deployment of large language models (LLMs) and other machine learning technologies. The new company, called Articul8, creates an independent software services firm using Intel’s AI IP library. The Articul8 platform can be deployed on-prem, in the cloud, or in a hybrid scenario.
Of course, the new company will push Intel’s AI hardware offerings, such as its Xeon Scalable processors, GPU Max, and Gaudi AI accelerators, but Articul8 will also “support a range of hybrid infrastructure alternatives,” including Nvidia GPUs.
This, Intel claims, will make it easier for businesses to securely train and run models based on their enterprise data.
“The Articul8 launch allows Intel to continue focusing on its core business and long-term strategy of delivering best-in-class compute options to its customers, and as an independent company supported by industry investment, Articul8 will be able to accelerate its go-to-market strategy and scale its product offerings for the broader GenAI ecosystem,” the company said in a statement.
Addison Snell, CEO of Intersect360 Research, which specializes in the HPC market, said the move helps position Intel in terms of leadership in AI software without getting into the complete vertical integration game, where they would be competing with their own customers.
“Intel, to the extent that it’s enabling AI, to me is stating that they want to do it more as a supplier or partner, and not as a vertically integrated competitor to the server vendors or the cloud or the hyperscale companies,” he told me.
He added that Intel’s investments in AI are more about what’s open and portable for the broader environment.
Intel lands major talent
The comings and goings of industry executives isn’t always significant, but this move is. Intel has hired Justin Hotard, a rising star at Hewlett Packard Enterprise, to lead the company’s data center and AI group.
Hotard will become executive vice president and general manager of the group effective February 1. He will succeed Sandra Rivera, who took over Intel’s programmable solutions group after it was spun out as a new, standalone business.
Hotard was executive vice president and general manager of high-performance computing, AI and labs at HPE. That included responsibility for the Cray supercomputer division. So he knows a thing or two about HPC in supercomputers.
More than that, though, it puts Hotard into contention for company leadership after Gelsinger departs. Clearly he has to deliver, but he’s in the place to do it, because the data center and AI group is the most important group at Intel.
Snell said HPE has a really strong talent bench, so it’s not too great of a negative on the HPE side. “But I think Intel did need to show some commitment in HPC and AI, and Justin has been right at the center of all of these major supercomputing deployments from HPE with its Cray architecture,” he said.
