Intel foundry aims for breakeven by 2027

Intel Foundry, still burning “billions every quarter” while adding capacity and developing new technologies, now has a clear profitability target. Speaking at J.P. Morgan’s Global Technology, Media & Communications Conference, CFO Zinsner reiterated that the operation should reach breakeven “sometime in 2027,” a window the company first floated earlier this year.

The first commercial 18A product—a client CPU codenamed Panther Lake—is scheduled to ship late 2025. Intel will also migrate the Clearwater Forest Xeon line and selected third-party designs to the same node. Management describes 18A as a proof of concept meant to convince customers to adopt its successors, 18A-P and 14A.

For now, external demand is modest. Many prospective customers are still running test chips, and “committed volume is not significant,” Zinsner conceded. To offset that, Intel expects its products to consume most early 18A wafers, improving fab utilisation and margins.

The breakeven model requires only “low- to mid-single-digit billions” of annual revenue from external wafers, supplemented by advanced packaging, mature-node output such as Intel 16, and joint production with partners like UMC and Tower. Adoption of High-NA EUV for 14A will raise costs initially, but Intel believes the density and performance gains will justify the expense.

Intel also plans to keep more of its high-end client silicon in-house: the upcoming Nova Lake CPU, built on 18A-P, will join Panther Lake on internal lines. Zinsner says forcing Foundry to “compete” for these wafers under the company’s smart-capital model should sharpen cost discipline while filling capacity, improving margins even before large external orders arrive.