Intel CFO predicts imminent major foundry wins, sees 2027 breakeven realistic with 18A,

Intel’s foundry outlook has shifted from a long‑term hope‑list to a near‑term revenue driver, as CFO David Zinsner told investors at the Morgan Stanley Technology, Media and Telecom conference that the company is “close to closing some deals that are in the billions of dollars per year” (The Register). The deals Zinsner referenced hinge on Intel’s advanced‑packaging portfolio—EMIB, Foveros 2.5 D and 3D technologies—that have already been used to stitch together in‑house silicon with TSMC‑fabricated dies. That capability, he said, is attracting “really good engagement” from customers seeking to assemble multi‑die AI accelerators, a segment that has exploded as generative‑AI workloads demand ever‑larger memory bandwidth and compute density (The Register). If those contracts materialize, they would inject “billions” of annual revenue into a division that has been a quarterly loss center for Intel, according to the same source.

The financial calculus behind the optimism rests on the performance of Intel’s 18A‑P and 14A process nodes, which Zinsner described as “effective solutions for external customers.” Wccftech reported that the 18A node has already met its yield targets, with early production runs showing the improvements the company promised when it launched the technology. Those yields, combined with the anticipated adoption of 14A, are expected to make Intel’s foundry services competitive against the industry’s dominant players, particularly for customers who need the tight integration that Intel’s packaging can provide. Zinsner linked the success of these nodes directly to the 2027 breakeven projection, noting that the combined revenue from 18A, 14A, and advanced packaging could lift the division to profitability by that year (Wccftech).

CEO Lip‑Bu Tan’s strategic pivot toward external customers is also a key factor. Under his leadership Intel has accelerated the ramp‑up of the Panther Lake platform, a testbed for the 18A process that now serves both internal and external workloads. Tan has publicly emphasized that the company’s packaging technology is not limited to chips built in Intel’s own fabs, a point reinforced by The Register’s observation that Intel already uses its EMIB and Foveros solutions to combine in‑house silicon with TSMC‑produced dies. That flexibility opens the door to a broader addressable market, especially among AI chipmakers that require heterogeneous integration but lack the capacity to develop their own packaging stacks.

Analysts remain cautious, however, about the timing and execution risk. The Register warned that Intel’s historical track record on delivering foundry commitments “is not the greatest,” and any delay could erode the confidence of prospective customers. Moreover, the “big Foundry wins” Zinsner described are still in negotiation, and the company has not disclosed specific customer names or contract values. The lack of concrete commitments means the projected “billions per year” revenue stream is still contingent on closing those deals and meeting the aggressive production schedules that AI customers demand.

If Intel can convert the packaging engagements into signed contracts and sustain the yield improvements on 18A and 14A, the financial impact could be transformative. The combined revenue from those process nodes and the advanced‑packaging business would not only offset the division’s current quarterly losses but also provide a steady cash flow that supports Intel’s broader capital‑intensive roadmap, including its upcoming 12A node and future AI‑centric products. In that scenario, the 2027 breakeven target—once viewed as a distant aspiration—becomes a realistic milestone, anchored by the very technologies that Intel has been betting on to stay relevant in an AI‑driven market (Wccftech, The Register).