AI capable PCs are set to dominate enterprise fleets within the next few years as chipmakers embed dedicated artificial intelligence acceleration directly into business laptops.
“By the end of 2026 nearly all of the products we ship into the enterprise space will have, or will be, an AI PC,” said Phil van de Mortel, Intel GM of global accounts for Europe, Middle East, and Africa (EMEA), during the Acer EMEA press conference this week in Milan, Italy.
The shift toward AI capable PCs is closely tied to Intel’s progress in semiconductor manufacturing, he said. Acer is already translating that silicon roadmap into commercial hardware, with the new TravelMate Copilot+ range among the first business laptops to use Intel’s Core Ultra Series 3 processors.
“A few years ago, we talked about five nodes in four years and how we were going to drive forward with technology development. And this year, we actually have moved into high-volume production on 18A.”
In semiconductor manufacturing, a node refers to a generation of chip production technology that determines how small and densely packed transistors can be. Each new node typically allows processors to deliver higher performance and improved energy efficiency.
“It uses the RibbonFET transistor, which is an all-round transistor technology. It uses PowerVia technology, which is a back-end power delivery. And this is enabling us to bring a new range of products to market which are cornerstone of Acer’s products this year.”
Those manufacturing advances feed directly into Intel’s latest processor lineup. Core Ultra Series 3 chips are designed to run AI workloads locally while maintaining the efficiency required for mobile devices.
“When I look across the Core Ultra series product range, also code-named Panther Lake, this actually brings more designs, more price points, more experiences. What I mean by that is over 200 designs in the market, more price points across mainstream and premium, and more experiences.”
By combining a hybrid architecture of performance and efficiency cores with the new 18A manufacturing process, Intel says it has achieved a 30% increase in chip density and a 15% improvement in performance per watt. The approach allows Core Ultra Series 3 processors to run demanding AI workloads more efficiently while extending battery life.
Intel is expanding graphics capabilities through the company’s latest integrated GPU architecture. Van de Mortel said the Xe3 core inside the products this year will deliver laptop gaming into the next level.
The architectural changes are designed to improve several aspects of system performance simultaneously, including graphics, AI processing and battery life.
“It’s going to improve battery life up to 27 hours, up to 77% faster graphics, faster CPU performance, and faster AI performance.”
Van de Mortel addressed perceptions around power efficiency in x86 processors.
“The new products that we’ve developed with this low-power technology island allows us to deliver 2.8x reduction in power from previous generations. Across the whole fleet of products, we can deliver low power, long battery life, [and] that experience that consumers, customers, [and] commercial customers want to experience.”
In practice, this means some parts of the processor can handle lighter tasks using far less energy, helping extend battery life without sacrificing performance when heavier workloads are required.
Building an AI software ecosystem
Intel’s strategy extends beyond silicon to the software ecosystem required for AI computing.
“But it’s not just about the processor,” said van de Mortel. “What we’ve invested heavily in is not just the hardware, it’s also the operating system, the framework and tools, AI models and applications.”
The company’s next-generation processors are designed to run AI workloads across multiple processing engines, combining CPU, GPU and NPU capabilities. Intel’s Panther Lake platform can deliver up to 180 TOPS of total AI performance, including an NPU capable of around 50 TOPS. The lower-power Wildcat Lake platform offers roughly 40 TOPS of combined AI compute, positioning it for entry-level AI PC systems.
“We’re also seeing up to 4.3x faster LLMs. A lot of these models were taken straight away as they’re released and actually optimising them on a PC.”
Intel is working closely with Microsoft to support the operating system environment for AI capable PCs.
“On the operating system side of things. Deep relationship with Microsoft. We’re now getting the bits, the drivers, all of their models well ahead of schedule for our next generation products to ensure that the out-of-the-box experience with Windows and Intel through partners like Acer is world class.”
Developer frameworks are another focus area.
“We’re making sure that we’re investing across the whole stack. Whether it’s PyTorch, whether it’s Llama, OpenVINO, making sure there’s choice for developers to utilise Intel technology and making sure that it’s open sourced as much as possible.”
Thousands of engineers are working to optimise AI models so they can run efficiently on x86 systems used in devices from partners such as Acer.
“One of the things that I found fascinating with Acer is that already they have applications that are optimised on the latest products, whether it’s AI image generator, display lenses, QR capture, the video speed controller. These are pre-loaded Acer applications that deliver AI applications into the market from day one.”
With devices already reaching the market, Intel’s AI PC strategy is moving from roadmap to enterprise deployment, with partners such as Acer integrating the latest technology into commercial laptops.