Intel’s AI pivot banks on CPU demand as Lip-Bu Tan pushes turnaround

Intel has spent years losing ground in leading-edge chipmaking and facing mounting skepticism about whether it could catch up. Since Lip-Bu Tan became CEO effective March 18, 2025, the company has been framed internally and by outside observers as being in turnaround mode. Now, a twist in the AI race is emerging as a potential help, with renewed attention on CPUs, the general-purpose chips that act as the “brains” of computers, and on how they fit into AI systems that go beyond chat.

For years, Intel struggled to maintain a market-leading position as rivals including Nvidia, AMD, and Qualcomm gained momentum. At the same time, the foundry challenge has weighed on the business: Taiwan Semiconductor Manufacturing Company (TSMC) produces the vast majority of the world’s most advanced chips, leaving less room for Intel’s own foundry to compete at the high end. In April 2025, Tan told investors there were “no quick fixes,” indicating how difficult the operational reset would be.

The renewed focus surfaced during Computex, an annual tech trade show in Taipei. Tan delivered a public keynote and, according to CNN’s reporting carried by KEYT, drew a crowd eager to hear more about Intel’s core CPU roadmap, now tied more directly to the AI boom. In the same setting, Nvidia CEO Jensen Huang used a “conductor” and “orchestra” analogy to argue that CPUs and GPUs each have critical roles in AI workloads. Tan told reporters that, over the prior month, many chief executives had been calling asking for more CPUs.

The underlying idea is that AI demand is not only about training large models, which typically leans on GPUs. It is also about inference, the step where a trained model is run to produce outputs. Agentic AI, described in the report as a new generation of AI that can complete tasks on a person’s behalf rather than only answering questions, depends heavily on inference. The reporting also pointed to examples of AI agents to illustrate the category, tying the potential demand shift back to why CPUs may matter more than they did in the earliest wave of AI hype.

Intel, in a statement to CNN, said Tan’s priorities were execution and putting customers back at the center of the company’s efforts. The statement argued that with a stronger balance sheet, a refreshed leadership team, and renewed engineering discipline, Intel is positioned to capture AI-driven growth opportunities. Analysts cited in the report said a CPU demand spike could translate into sales momentum if Intel can ramp production quickly and maintain quality.

Tan’s turnaround plan has also involved visible structural change. The report said he has laid off roughly 34% of Intel’s workforce and paused plans to expand manufacturing facilities in Germany and Poland. It also described steps aimed at cutting bureaucracy and re-centering engineering, including elevating engineers and reducing middle-management layers so leadership could track Intel’s chip and AI efforts more directly. The report further said Tan sold a controlling stake in a subsidiary for liquidity, and hired senior executives from Qualcomm and Arm to lead Intel’s data center and AI business divisions.

Alongside internal restructuring, Intel has received government support that could cushion the company during a fragile ramp period. In an Intel press release, the company said the U.S. government will invest $8.9 billion in Intel common stock as part of an agreement with the Trump Administration, describing it as a 9.9% stake. Intel said the equity stake is passive, with no board representation or other governance or information rights, and that the purchase price implies a discount to the market price at the time of announcement.

Still, the report emphasizes that the story is not fully resolved. Analysts said Intel continues to face major challenges in its chipmaking foundry business, particularly in attracting more clients and improving production quality. Intel did not disclose in the CNN reporting a specific timeline for how quickly new AI-linked CPU demand would translate into higher volumes or improved margins. What to watch next is whether Intel can convert CPU “interest” from executives into signed deployments, and whether it can deliver stable supply and quality as AI agents scale in real-world products.