Intel Shares Touch Nasdaq Record on Apple Chip Deal; Foundry Pipeline Faces Its Sharpest Commercial Test Yet

A reported Apple foundry agreement and a confirmed Google TPU deal have lifted Intel to a fresh Nasdaq record, but converting advanced-node ambition into commercial yield remains the company's defining challenge.

Intel shares touched a fresh record on the Nasdaq on June 23, 2026, buoyed by reports that the company is advancing toward a chip manufacturing agreement with Apple — a development that, if confirmed, would rank as the most significant commercial validation of Intel Foundry Services to date. Neither company has officially confirmed the deal, and it should be treated as a reported development rather than a signed contract; but even the prospect of Apple custom silicon flowing through Intel fabs crystallized a narrative that has been building since April: that Intel's restructured foundry operation, anchored on its 18A-P process node, is attracting hyperscaler and consumer-technology clients who once seemed committed exclusively to TSMC.

The Apple headline arrived two days after Intel shares surged on a separately confirmed agreement to manufacture Google's Tensor Processing Units, a deal that positions Intel as a domestic U.S. supplier of AI accelerator silicon at a moment when supply-chain resilience has become a geopolitical imperative. Reports as early as June 8 had flagged that both Google and Nvidia were evaluating Intel as a backup chipmaker; the Google TPU contract converted that speculation into revenue-visible reality. ASML's commentary on June 22 placed Intel's 18A-P progress alongside the MATCH Act's tightening restrictions on China's access to advanced lithography equipment — a pairing that underscores why U.S. policymakers and technology companies share an interest in Intel's foundry success that goes beyond ordinary commercial calculus. A $14.2 billion fab deal reported in early April, alongside reports of a collaboration with Elon Musk's Terafab project, had already begun reframing Intel's foundry pipeline as substantive rather than aspirational.

At the product level, two developments from the past week stand out. Intel and AMD jointly ratified the x86 AI Compute, or ACE, instruction set on June 21-22 as a shared ecosystem standard, retiring Intel's proprietary AMX architecture in the process. The cooperation is commercially pragmatic: a unified AI acceleration specification reduces developer fragmentation across x86 platforms and concentrates competitive pressure against the CUDA-and-Arm axis rather than within x86 itself. At Computex 2026, Intel simultaneously previewed its Xeon 6-plus server processor and described the forthcoming Crescent Island Xe3P GPU — reportedly scaling to 480 GB of cost-optimized LPDDR5X — as a lower-cost, more thermally efficient alternative to Nvidia's Rubin and AMD's MI450X. Company executives stated those characteristics publicly at Computex; production validation at commercial scale will determine whether the claims hold. The reported cancellation of discrete gaming variants from the Xe3P Arc Celestial lineup suggests Intel is concentrating its GPU roadmap on data-center workloads, exiting a consumer gaming segment where competitive returns have proved elusive.

The financial arc of the past decade provides indispensable context. Intel's annual capital expenditure climbed from roughly $4.5 billion in fiscal 2009 to a peak of $25.75 billion in fiscal 2023 — the high-water mark of Pat Gelsinger's IDM 2.0 drive to simultaneously reclaim process leadership and build an external foundry business. Execution struggles, most visibly the multiyear delays on 10-nanometer and the consequent ceding of process leadership to TSMC, made that capital cycle expensive and dilutive for shareholders. Under Lip-Bu Tan, who took the helm in 2025 and has since pursued aggressive operational restructuring, Intel cut annual capex to $14.65 billion in fiscal 2025, a 43 percent reduction from the 2023 peak. Q1 2026 capex of $3.64 billion, measured against FY2025 revenue of $52.85 billion, implies a capital intensity of roughly 28 percent — elevated by semiconductor-industry norms but sharply reduced from the 2022-2023 buildout. Intel's Q1 2026 earnings were broadly read as signaling recovery, and the 8 percent share-price jump in April on fab-deal news underscored how quickly investor sentiment has shifted; the full-year FY2026 capex trajectory remains the number that will tell markets whether foundry pipeline is converting into committed wafer starts.

The risks deserve equal weight. Intel's foundry track record is difficult to separate from a history of process delays: TSMC's advanced nodes continue to command the majority of leading-edge foundry demand, and its 2-nanometer roadmap is advancing on schedule. On the product side, Nvidia's RTX Spark SoC entry into the PC market adds competitive pressure on Intel's core client-computing franchise, while Nvidia's Vera data-center CPU — deployed alongside Intel Xeon 6 as the host processor for DGX Rubin NVL8 systems — hints at an architecture that could over time reduce hyperscaler dependency on third-party CPUs. The ACE cooperation with AMD, while strategically sensible, also implicitly acknowledges that neither company alone controls the x86 AI compute agenda. Three signals are worth monitoring over the next twelve months: whether the Apple foundry agreement advances to a signed volume contract specifying a process node and production timeline; whether Intel's 18A-P process achieves independently verified yield milestones sufficient for commercial qualification rather than benchmark-level demonstration; and whether full-year FY2026 capex rises materially above the $14.65 billion FY2025 floor — a sustained uptick would be the clearest indicator that foundry orders are converting from reported pipeline to actual wafer starts, and that Intel's manufacturing recovery is moving from aspiration toward execution.