Intel’s third quarter earnings call provided some fascinating insights.
Intel’s third quarter earnings call provided some fascinating insights.
As Windows 10 approaches its end, Intel anticipates significant growth in the PC industry — the highest since 2021, when the covid-19 pandemic triggered a substantial uptick in demand. However, the struggling Intel, which has recently received support from Nvidia, Softbank, and the US government, isn’t quite positioned to fully capitalize on it and is focusing on AI instead.
During the company’s Q3 2025 earnings call, where Intel achieved its inaugural profit in nearly two years primarily thanks to these supports, CEO Lip-Bu Tan and CFO David Zinsner elaborated on the chip shortages the company is currently facing, anticipating that these shortages will peak in the first quarter of the following year — in the interim, leadership indicates a shift in focus towards AI server chips over certain consumer processors as they navigate supply and demand challenges.
“We project a slight decrease in CCG [Intel’s consumer chips] while DCAI [Intel’s server chips] will see significant growth as we prioritize our capacity for server shipments over entry-level client components,” states Intel. Tan disclosed today that Intel intends to introduce new AI GPUs annually, following Nvidia and AMD in altering their traditional schedules to meet the burgeoning demand for AI server solutions. It remains uncertain what this means for those seeking additional Intel gaming GPUs.
With attention fixed on Intel’s exciting new Panther Lake and its 18A process aiming to demonstrate its capability in delivering top-tier consumer PC chips in-house, the company reaffirmed that it will only roll out one Panther Lake SKU this year with additional releases planned for 2026. Another potential reason for this is Zinsner’s mention today that the initial offering of Panther Lake will be “quite expensive,” prompting Intel to push existing Lunar Lake chips instead “at least during the first half of the year.”
Despite Intel’s continuous rebuttal regarding the notion of poor yields with its 18A process, the company conceded to investors and analysts today that it’s not primed for significant financial success just yet: yields are “sufficient to meet supply needs but not where we require them to be to achieve the desired margins,” remarks Zinsner, suggesting an “acceptable yield level” may not be realized until 2026 or potentially even 2027.
At present, Intel will be “collaborating closely with customers to optimize our available output, including altering pricing and product mix, to steer demand towards items where we have supply and they have the demand” — implying adjustments in pricing for PC manufacturers to favor Intel components in their systems while steering them towards Lunar Lake chips instead of newer models. Tan reaffirmed today that further investments in capacity won’t happen unless there is “committed external demand,” while Zinsner noted that investment in capacity next year isn’t expected to “substantially alter expectations.”
Intel asserted that the 18A will be a “long-lasting node” powering “at least the next three generations of client and server offerings.” If you were anticipating a return to the “tick-tock” approach where Intel alternates between miniaturizing its chips and launching new architectures each generation, that seems unlikely to occur.
However, this doesn’t imply Intel will cancel its next node, Intel 14A, as it had previously hinted. Tan indicated today that clients have intervened to bolster 14A, and Intel is “pleased and more assured” about it, while Zinsner remarked that it’s not merely “off to a good beginning,” but has surpassed 18A at this point “in terms of performance and yields.”
