Intel goes after AI PC value of $50B by targeting software

Just days ago, Nvidia showed the world how valuable the AI market is to silicon and solution providers by announcing a new set of hardware and services during its annual GPU Technology Conference in San Jose. While that news was focused on data center AI compute in the form of multi-million dollar servers powered by some of the largest processors and chips ever produced, the world of client AI, where AI runs on individual laptops and PCs, presents an equally appealing, and profitable, arena for technology giants to take advantage of.

The generative AI battle is on track by 2027 to be worth $50B-$150B according to guidance provided by chip companies and becoming the default platform of choice to run that AI compute locally, on consumer laptops and smartphones rather than in the cloud, is the primary fight we’ll see take place in the devices segments in the next 2 years.

Those kinds of dollars are exactly why we are seeing Intel, AMD, and Qualcomm position themselves as the best options for consumers and commercial enterprises looking to take advantage of the AI transition.

Of those three primary market competitors, it is likely that Intel will actually find itself in a raw performance deficit when compared to the others, at least for the next couple of quarters. It’s new Intel Core Ultra processor family integrates an AI processor called an NPU (neural processing unit) but offers slightly lower performance than the solutions AMD is shipping in its Ryzen mobile family. And the upcoming Snapdragon X Elite platform from Qualcomm will likely quadruple the theoretical peak performance when it comes out this June.

But Intel has a couple of advantages that neither of the others have: scale and resources. Intel continues to claim it will ship more than 100 million units of its Intel Core Ultra family of processors that include an NPU by the end of 2025. That could be 5x what AMD and Qualcomm ship, combined.

And it is that scale that enables Intel to put its second advantage, resources, to work.

If Intel won’t be the pure performance leader for AI applications on the PC, then it plans to win by making sure it has software developers on its side, ensuring that these new AI workloads and applications work best on Intel platforms.

Late last year Intel started its AI PC Acceleration Program, a combination of technical assets and marketing dollars to help woo software developers into focusing first on enabling the Intel NPU hardware. This includes mainstay software like that from Adobe or Autodesk, but also startups and indie app developers, all of which get some combination of hands-on time from Intel engineers for guidance and the potential for dollars to market these tools in partnership with the biggest IT chip vendor in the consumer space.

Newly announced this week is an expansion of that program to provide hardware, small mini-PC enabled with the latest Intel chips and AI accelerators, to software developers at a free or highly subsidized price. More hardware in the hands of developers means a better chance of Intel’s computing platform being the default option for the wave of AI tools coming this year. 

It's a pattern that Intel has used time and again, turning its massive market share and marketing funds into real-world benefits for software developers and partners. This then serves two purposes: getting more AI applications out into the world that have compelling use cases to drive the adoption and sales of AI PCs AND making sure that those applications run out of the box, on day one, on Intel hardware for a great consumer experience.

In some ways this movement is trying to replicate what Nvidia was able to accomplish with its CUDA programming platform for AI in the data center. Intel has its own software layer, called OpenVINO, that it hopes more developers will integrate to ensure long term enablement of Intel NPU hardware for consumer and commercial systems. Intel is even working to get developer kits (computers that have the NPU hardware included) to universities to make sure that that next wave of genius’ building the tools of the future have access to Intel hardware to test, learn, and deploy.

Success of a program like this isn’t a guarantee as software developers are traditionally stuck in their ways. But for the AI PC segment, we are paving new roads as we go, without a clear winner on the software side (unlike in the data center segment with CUDA). Intel wants to take advantage of that opening to create a leadership position that will eventually turn into a financial benefit.

AMD and Qualcomm might have competitive, or more than competitive, solutions on the hardware side, but neither of those chip companies has the history and experience of working with the entirety of the PC ecosystem like Intel does. From consumer marketing campaigns and partner funding programs, to driving the enterprise PC segment with security and manageability tools, Intel uses its massive 120,000+ person organization to be everywhere, and work with everyone.

The question then is how does Intel turn this outreach and things like the AI PC Acceleration Program into revenue? In a discussion this week with Todd Lewellen, VP of the PC ecosystem at Intel, and Carla Rodriquez, VP of client software, both discussed this as a long play, not something that they expect to see immediate returns on. Software development takes time and the industry settling on a specific software model will happen over the course of quarters, not weeks.

Intel also mentioned the possibility of this software acceleration program reaping benefits from a paid services standpoint. If Intel is able to work with a developer in early enough stages and build a software-as-a-service business model that could augment the sale of PC chips, just as Nvidia is attempting to do with its inference microservice announcements this month, then it turns Intel’s position as a one-off chip sale to a company that can benefit from the AI PC race over the long haul.

Every week it seems there are new layers being added to the complexity around how this AI PC race, and how AI for end consumers in general, will play out. Across the gamut of software and hardware companies, from industry giants like Microsoft to indie devs looking to making a name for themselves with an innovative idea, Intel has a unique opportunity. It has the scale and history to be a leader in this enormous segment, but it needs to show that it isn’t stuck in the old ways of doing things. It’s future as a critical piece of the PC ecosystem is at stake.