Gelsinger: Intel Is Prioritizing AI Sales Enablement For Partners

‘Clearly, this is a priority for us, particularly our enterprise and cloud sales teams. Their number one priority is our AI offerings,’ CEO Pat Gelsinger tells CRN about Intel’s strategy to compete with rival Nvidia through channel partners. In a follow-up interview, Intel’s global channel chief and chief commercial officer detail how the company is enabling partners to sell AI solutions.


Intel CEO Pat Gelsinger said enabling channel partners to sell the chipmaker’s portfolio of AI hardware and software solutions is a “priority” as it moves to take market share away from rival Nvidia.

More than two years into his tenure as Intel’s CEO, the semiconductor veteran made the remarks to CRN in the middle of the company’s significant charm offensive to convince partners, developers and customers to build solutions on its AI products at the third annual Intel Innovation event this week.

[Related: 6 Big Announcements At Intel Innovation 2023: From 288-Core CPU To AI Supercomputer]

Sponsored post

“Clearly, this is a priority for us, particularly our enterprise and cloud sales teams. Their number one priority is our AI offerings,” he said Tuesday in a question-and-answer session with journalists.

Those offerings include Intel’s Xeon Scalable data center CPUs with AI capabilities, which Gelsinger said are “becoming a larger portion” of Xeon sales due to their “very differentiated capabilities” for enterprise deployments running inference on AI models for live applications.

Intel is also making a lot of noise about its Gaudi2 deep learning processors and Max Series GPUs that are built for AI workloads with larger data sets in data centers, such as training large language models.

The company plans to make a big push for AI with PCs too, thanks to its upcoming Core Ultra processors that can handle a diverse range of AI workloads thanks to their CPU, GPU and neural processing unit.

“That’s where channel programs are helpful, [deploying] field resources, working with the channel partners, directly engaging with the end customer,” Gelsinger said.

The CEO also said Intel’s growing set of software development tools such as the Intel Developer Cloud and oneAPI programming model will serve as “scale mechanisms” for partners selling Intel AI solutions.

While Nvidia remains the market leader for AI computing with its GPUs, systems and software, Gelsinger said “there’s a lot of industry interest in a good alternative.”

That’s good news for Intel’s prospects, he added, because the chipmaker has “traditional confidence from the market” that it’s the “open alternative that we can trust” due to its focus on enabling a “broader ecosystem” through open APIs.

“I think that’s the play here in face of a very proprietary alternative,” Gelsinger said, referring to Nvidia. “We’ll get a lot of good momentum globally through channel partners, through an ecosystem of partners really covering that whole space.”

An executive at a global IT services firm told CRN that the range of AI hardware and software announcements made at Intel Innovation this week shows that the chipmaker “has pretty much gotten its act together” as it seeks to take market share away from Nvidia.

“There’s a significant amount of mindshare and market share that has already been taken by Intel’s competitors. But as long as the ingredients are there, competing becomes easier, and it will be a level playing field for almost all of them. It’s definitely a very strong a set of products that Intel has now to compete,” said Suresh T. Kumar, a former Intel employee who is now vice president and global head of the ecosystem business unit at India-based HCL Technologies.

Intel Is ‘Enabling A Partner Ecosystem To Bring AI Everywhere’

In an interview with CRN, Intel Global Channel Chief John Kalvin said the totality of the chipmaker’s investments in its AI strategy is about “enabling a partner ecosystem to bring AI everywhere.”

“It’s going be on the AI PC. You’re going to have AI at the edge. You have AI that will take place in the data center. Some of it will be inference. Some of it will be deep learning,” said Kalvin, whose official title is vice president and general manager of global partners and support.

“But I don’t really distinguish the company’s partner investment separate from the whole investment we’re making to really bring AI everywhere,” he added.

Intel provides a wealth of AI-related resources to partners under the Intel Partner Alliance program, according to Kalvin. These resources include an AI accelerator program, training as well as engineering and support “that can help customers accelerate their AI solutions to market,” he said.

Intel Developer Cloud Can Help Partners With AI Deals

In the same interview, Intel Chief Commercial Officer Christoph Schell said the company’s recently launched Intel Developer Cloud platform will play an important role in how partners build and test Intel-based AI solutions for their joint customers.

Schell said when partners are in early development for AI projects, they should take those projects to Intel Developer Cloud, where they can get access to the latest Xeon CPUs, data center GPUs and Gaudi deep learning processors running on cloud infrastructure built by the company.

The platform also includes AI frameworks and tooling such as Intel’s OpenVINO toolkit and Intel optimizations of PyTorch and TensorFlow. In addition, it provides access to Intel toolkits and libraries such as the Intel oneAPI Base Toolkit as well as popular AI foundation models like Llama 2 and Stable Diffusion.

“Before we do anything, let’s test the model, let them experience our software stack. Let’s see if we can help them on the specific workloads and then have the discussion of where we take it. And so having that additional step is, I think, a relief for a lot of customers and channel partners because they don’t have to build it themselves,” Schell said.

Intel Developer Cloud can help partners determine whether they need a powerful processor such as a GPU or if they can use something less expensive like a CPU.

“Not every workload needs a GPU. As a matter of fact, if you look at our funnel [of AI customer deals], the majority of our funnel is enterprise proprietary data, which is important from a data integrity point of view,” Schell said.

“But these model sizes are not the largest ones. So also giving a customer and partner flexibility to work across both CPU and GPU has a huge impact on [total cost of ownership],” he added.

While Intel is trying to compete against Nvidia’s GPUs with its own graphics chips and Gaudi processors, Schell said he hasn’t heard from customers who want to know how they compare. Instead, customers want to know how Intel’s software stack differs from rivals,” he added.

“You need to take your customer by the hand and educate them and learn together actually how their data will operate on a software stack. And so this is not an off-the-shelf type of value proposition. It’s one customer at a time, one set of data at a time. And I think the best investment that we’ve made in this is our Developer Cloud,” Schell said.

The executive, who joined Intel from HP Inc. last year, said the results have already been promising.

“I think being able to tell a partner, ‘Bring your customer onto this cloud, and let’s stop talking theoretically but really actually look into how we can train a model, how we can inference that, and then have a business discussion,’ that has been very helpful in the last couple of weeks,” he said.