Intel’s AI Chip Overhaul: New CEO, Foundry Pivot, and Geopolitical Strategy

Intel Corporation was once the undisputed champion of semiconductors. Now, the company is fighting to get back in the ring. It’s making an aggressive push into artificial intelligence chips, a market where rivals like Nvidia and AMD are already throwing heavy punches. The competition is fierce, and Intel is playing catch-up.

Starting in August 2025, a new CEO is calling the shots. Lip-Bu Tan has a plan to overhaul Intel’s AI strategy. The focus is shifting to integrated systems, edge computing, and foundry services. In simple terms, Intel wants to be the “systems foundry for the AI era.” This means they’ll not only design their own chips but also manufacture chips for other companies. It’s a bold move that challenges Nvidia’s dominance while also banking on U.S. government support to bring manufacturing back home.

A Shake-Up at the Top

To understand Intel’s new direction, you have to look at the leadership changes. The company brought in Lip-Bu Tan, a semiconductor industry veteran, as CEO in March 2025. He replaced Pat Gelsinger amid concerns about the company’s direction and missed opportunities in AI. Tan is pushing for an “engineering-first mindset” and a culture that focuses on customers and collaboration.

The changes have been swift and deep.

• Executive Moves: Three senior manufacturing executives retired in August 2025 as part of a restructuring. In June, new vice presidents were hired to oversee customer satisfaction and foundry success. The heads of the data center and PC units now report directly to Tan.
• Job Cuts: Intel is planning to cut 25,000 jobs in 2025, which is about 15% of its workforce. The goal is to slash costs by $10 billion. This points to serious structural issues but also a tighter focus on AI and its foundry business.
• Political Pressure: The political climate adds another layer of complexity. President Trump called for Tan’s resignation in early August over alleged ties to China. By mid-August, the tone changed. The administration began supportive talks about a potential government stake in the company. This shows how important Intel is to national strategy.

These moves are meant to stabilize the company. But with five CEOs in just ten years, the frequent changes have created a sense of uncertainty.

The New Game Plan

Intel’s 2025 roadmap is all about weaving AI into everything. This includes PCs, devices at the network’s edge, and data centers. The company is moving away from scattered acquisitions and focusing on building new technology in-house.

Here are the key parts of the plan:

• AI Chips and Accelerators: The Gaudi3 is Intel’s latest AI accelerator, designed for training and inference. But it’s far behind Nvidia’s offerings. Intel projects $500 million in Gaudi3 sales for 2024, while Nvidia is making billions. Another high-end AI GPU, Falcon Shores, was delayed from late 2025, showing that the roadmap is still being adjusted. The company is now fine-tuning products for trends like robotics and AI agents.
• Manufacturing and Products: The advanced 18A manufacturing process is on track for high-volume production in the second half of 2025. This will power upcoming mobile CPUs (Panther Lake) and server chips (Clearwater Forest). A desktop chip called Nova Lake is set for 2026. Intel’s AI PCs, running on Core Ultra processors, feature specialized neural processing units (NPUs) to handle local AI tasks efficiently.
• Foundry and Systems Focus: The bigger picture is for Intel to be a one-stop shop. They want to offer chips, software, and full systems for AI, especially for “physical AI” in edge devices and data centers.

This strategy makes sense for enterprise and edge AI. However, Intel still doesn’t have the powerful ecosystem that makes Nvidia the king of large-scale AI training.

Losing Ground in the Market

Intel still leads in the x86 CPU market, but its grip is slipping, particularly on the desktop side.

Finding the Right Customers

Intel’s AI customer base looks different from Nvidia’s. Instead of dominating massive AI training centers, Intel is finding its footing with businesses and edge applications.

• Enterprise and Edge: Companies like Bosch, Siemens, and specialized firms like EdgeRunner AI are using Intel’s Xeon CPUs and Habana Gaudi chips for robotics, distributed AI, and inference tasks.
• Government and Defense: U.S. agencies are key customers, seeking secure, domestically produced AI computing power.
• PC Makers: Major manufacturers like Dell and HP are using Intel’s chips with NPUs to build the next generation of AI PCs.

The biggest cloud providers, known as hyperscalers, are potential customers for Intel’s foundry services. But they rarely use Intel’s own chips for their main AI work.

Hyperscalers: Build or Buy?

Cloud giants like AWS, Google, and Microsoft have a different strategy. They are increasingly building their own custom chips to optimize performance and cut costs. This reduces their reliance on vendors like Intel or Nvidia.

• The Custom Chip Trend: AWS has its Trainium and Inferentia chips. Google has its TPUs. Microsoft has Maia. Amazon is also investing heavily in its own silicon for AI. These companies see custom chips as a way to escape the high costs and limitations of general-purpose GPUs. By 2030, the custom AI chip market could be worth over $400 billion.
• A Chance for Collaboration: Intel’s foundry business (IFS) could still win over these giants. With U.S. government incentives, hyperscalers might choose Intel to manufacture their custom designs. CEO Lip-Bu Tan is working to improve the foundry’s customer service. But for now, these companies still prefer TSMC for the most advanced manufacturing.

For the most part, hyperscalers will likely keep building their own chips. They may work with Intel selectively, especially if U.S.-based production becomes a priority due to geopolitics.

A Pillar of National Strategy

Intel plays a crucial role in the Trump administration’s “America’s AI Action Plan.” The plan aims to “win the AI race” by cutting regulations, building infrastructure, and securing a domestic supply of semiconductors.

• Domestic Manufacturing: The plan uses CHIPS Act grants to bring chip production back to the U.S. It also simplifies regulations and puts export controls on China. As the only major U.S.-based foundry, Intel is central to reducing reliance on Taiwan and China.
• Government Stake: As of August 18, 2025, the administration is in talks to take a roughly 10% stake in Intel. It would do this by converting CHIPS Act grants into equity. The government views Intel as “too big to fail” for national security. This could provide a major capital injection and help counter China’s progress in AI.
• Broader Impact: Intel’s role supports the administration’s goals of promoting U.S. AI exports, creating jobs, and integrating technology into national defense. This makes the company a key piece in a global chess match.

The Next 5 Years: A High-Stakes Bet

Intel’s outlook for the next five years is a toss-up. Its success hinges on its 18A manufacturing process, the adoption of its AI products, and continued government support. If the foundry business takes off, revenue could top $70 billion by 2030. Analysts see the stock price sitting at $22-25 in 2025. In a best-case scenario, it could rise above $30 by 2030. But the company faces huge challenges, including projected losses of over $10 billion in 2025. Still, AI PCs and edge computing could fuel 10-15% annual growth.

The Bull Case:

• Foundry and AI Win: The 18A process succeeds and attracts big customers like Microsoft. AI PCs capture 50% of the market by 2028. A government stake provides over $10 billion in funding to boost research.
• Market Comeback: Intel reclaims 75% of the desktop market and grows its server business with Xeon chips for AI inference. The stock doubles to over $40 by 2030 as the turnaround succeeds.
• Geopolitical Advantage: U.S. policies continue to favor domestic chipmakers, making Intel a national champion.

The Bear Case:

• Execution Fails: Delays in the 18A process kill confidence. AI chips like Gaudi fail to gain traction, with sales making up less than 1% of Nvidia’s. The desktop market share falls below 60%.
• Financial Strain: The company continues to lose money. Its second-quarter revenue was $11.9 billion with break-even earnings per share. More layoffs and debt problems follow.
• Competition and Politics: Hyperscalers stick with their own custom chips. Political scrutiny from Trump or failed stake negotiations could sink the stock to $10.

In short, Intel’s push into AI is a high-risk gamble. It’s betting big on its foundry and its ability to win at the edge, all with the backing of U.S. strategic interests. Government involvement could be a game-changer, but deep internal challenges remain. Success will depend entirely on execution.