After a bruising few years defined by lost ground in AI, mounting competitive pressure, and the fading importance of some of its core products, Intel entered 2025 in recovery mode.
Twelve months on, the chipmaker is betting on a new wave of AI PCs and a sharper strategic focus to restore its relevance in a new era of computing.
AI has been the guiding force for the entire technology sector, with very few companies managing to stick to their principles without some influence from AI. Intel, more than most, has been forced to come to terms with its own limitations in this area and pivot its operations to be a player in the AI market.
The fully-integrated AI chipset
The company has aggressively pushed the AI PC category, building on the integrated neural processing unit (NPU) Core Ultra architecture first introduced in 2023 and expanding its message that AI workloads should run locally on everyday desktops and laptops.
This was the key focus at CES 2026, where Intel unveiled the Core Ultra Series 3 built on its new Panther Lake architecture and Intel 18A node manufacturing process. The product brings several Intel projects together.
Panther Lake is an attempt to make its processors scalable to a wider range of laptops, potentially being an option for everything from lightweight laptops to high-performance desktops.
At the same time, the 18A node was part of ex-CEO Pat Gelsinger’s plan to spend heavily on foundries and R&D to make Intel the best-in-class in chip manufacturing again.
Intel reports high demand at CES 2026
It seems to have been somewhat successful, with Intel reporting high demand for its AI chips and announcing at CES that they have already been integrated into 200 laptop designs.
Early tests and reviews of the Core Ultra Series 3 are highly positive, thanks to the chips’ impressive efficiency and improved integration with Microsoft Copilot and other AI models.
Intel sees the AI PC market as likely to consume the regular, predicting that half of all PCs sold in 2026 will feature built-in AI processing in the form of NPUs.
Intel remains the dominant supplier of integrated chipsets for laptops and desktops, so most of these AI-enabled PCs will run its silicon.
Partnerships instead of acquisitions: what the SambaNove investment signals
Another sign of Intel changing its strategy is its recent partnership with SambaNova, which was reportedly an acquisition target of Intel in late 2025.
The two companies announced a strategic, multiyear collaboration, alongside an Intel investment of between $100 and $150 million in SambaNova as part of a $350 million funding round.
The collaboration will involve SambaNova integrating Intel’s silicon into its AI data center architecture, providing Intel with a fast-growing startup in the field.
“AI is no longer a contest to build the biggest model,” said Rodrigo Liang, co-founder and CEO of SambaNova, announcing the deal.
“The real race is about who can light up entire data centers with AI agents that answer instantly, never stall, and do it at a cost that turns AI from an experiment into the most profitable engine in the cloud.”
SambaNova is a direct competitor to Intel rival NVIDIA in the AI hardware space, particularly in inference workloads and enterprise solutions.
The startup’s Reconfigurable Dataflow Unit and dataflow-optimized architecture are excellent for the day-to-day operations of an AI chatbot or image generator, as they have been shown to perform more efficiently than standard GPUs on routine tasks.
The company, which generates a microscopic amount of revenue compared to Nvidia, was valued at $5 billion by SoftBank in 2021.