Intel is moving away from its long-standing tick-tock model, a strategy that once defined the company’s chip development cycle, and is now focused on prioritizing AI over consumer PC growth. Despite Windows 10 nearing the end of its support life and the broader PC market showing signs of recovery—its first meaningful growth since 2021—Intel is struggling to meet demand, even as it reports its first profit in nearly two years. This turnaround is largely due to recent support from Nvidia, SoftBank, and the U.S. government, which have helped stabilize Intel’s financial position. However, on its Q3 2025 earnings call, CEO Lip-Bu Tan and CFO David Zinsner confirmed that the company still faces significant chip shortages, with supply constraints expected to peak in the first quarter of 2026. In response, Intel is redirecting its production capacity toward AI server chips, particularly in the Data Center and AI (DCAI) segment, which is seeing strong demand. As a result, consumer chip (CCG) shipments are expected to decline modestly. To keep up with the fast pace of the AI market, Intel announced it will now release new AI GPUs annually—aligning with Nvidia and AMD’s updated cadence. This shift suggests a long-term commitment to AI infrastructure, even if it means delaying or limiting consumer-focused product launches. The company’s upcoming Panther Lake processors, built on the new 18A process, are set to launch with only one SKU this year, with additional variants planned for 2026. Intel cited high initial costs for Panther Lake, calling it a “pretty expensive” product to produce. As a result, the company will continue pushing its current Lunar Lake chips through at least the first half of 2026. Despite earlier claims that 18A had strong yields, Zinsner admitted that while yields are sufficient to meet current supply needs, they are not yet at a level that supports healthy margins. He suggested it may take until 2026 or even 2027 to reach acceptable yield levels. Intel is managing the situation by working closely with customers to adjust pricing and product mix, steering demand toward available products. Tan emphasized that the company will not expand manufacturing capacity unless there is firm external demand, and Zinsner noted that planned investments for next year won’t significantly alter supply expectations. Still, Intel remains confident in its next node, 14A. Tan said customer interest has helped secure its future, and Zinsner confirmed it’s off to a better start than 18A was at the same stage, with stronger performance and yield progress. While 18A will serve as a long-lived node for multiple generations of client and server chips, the end of the tick-tock era is now clear. Intel is no longer betting on predictable, alternating architecture and process improvements—it’s betting on AI, even if it means slowing down its consumer roadmap.