Intel’s Resurgence: A Story of Supply and Demand

After a year of rebuilding its reputation with buyers of central processing units (CPUs), Intel, the semiconductor giant, has faced a new challenge: it cannot produce enough chips to meet the growing demand. The shortage is starting to show up in prices, with Intel reportedly raising list prices on its Xeon server chips and Core Ultra laptop chips heading into the third quarter.

According to Wedbush Securities analyst Matt Bryson, the ongoing shortages of server CPUs give Intel room to raise prices without hurting demand. Bryson noted that the real question is not whether Intel can raise prices, but where those increases are landing. This could be either in official list prices or in smaller retail and distribution prices.

Server Chips and the Rule of 45

Intel’s server chip revenue grew in the 20% to 25% range last quarter, driven primarily by average selling price increases rather than unit growth. This is a significant development, as server chips account for a much larger share of Intel’s business. If the increases are showing up broadly, it suggests that Intel has pricing power for the first time in years, a sign that supply is tight.

Chief Financial Officer David Zinsner explained that as core counts per chip increase, prices naturally rise too. Notably, Intel is also seeing like-for-like price gains on a per-core basis, something that had been sliding for years. Zinsner added that Intel is locking in longer-term agreements with customers that fix both price and volume, giving the company better visibility into how much capacity to build.

Intel has set a longer-term financial goal, internally referred to as the Rule of 45, that combines revenue growth and operating margin into a single target above 45%. Zinsner called it a multiyear goal rather than something achievable immediately, but said improving costs, steadier pricing, and rising demand all point in the right direction.

The Shift to CPU-Heavy Workloads

The demand itself is being reshaped by artificial intelligence (AI). Intel CEO Lip-Bu Tan told investors that the ratio of central processing units (CPUs) to graphics processing units (GPUs) in AI systems has shifted dramatically. Training-era workloads leaned heavily on GPUs, often with one CPU per eight GPUs. However, with agentic AI, where software agents plan, use tools, and complete multi-step tasks on their own, customers are telling Intel the ratio is closer to 1:1, and in some cases, four CPUs for every GPU.

At Intel’s Computex keynote, Corporate Vice President Kevork Kechichian showed a live demo comparing traditional AI inference against an agentic AI workflow. The traditional setup was GPU-heavy, running close to 7-to-1 in favor of graphics chips. The agentic version flipped that balance toward CPU-heavy work, since agents spend more time on tasks like fetching data, running code, and checking rules, jobs that CPUs handle best.

Intel’s ability to meet the growing demand for its CPUs is crucial to its success. The company has pulled forward its manufacturing yield targets for its advanced 18A chip-making process by at least a quarter, which should help produce chips faster and ease some of the current bottleneck.

While Intel’s price hikes are a positive development, it’s essential to note that none of this guarantees that Intel will keep its pricing power once new capacity comes online. However, for now, a company that struggled for years to sell its story to Wall Street has a much simpler pitch: demand is outrunning supply, and customers are paying up rather than walking away.

With 11 analysts recommending ‘Strong Buy’ and an average price target of $102.87, Intel’s stock is poised for growth. However, investors should be cautious and keep an eye on the company’s ability to maintain its pricing power in the long term.