Intel Latest News: GPU D270, Applied Materials AI Partnership, and 14A Foundry Timeline, October 2026

Intel's October 2026 news centres on the GPU D270 AI inference chip delivering 1.3 PFLOPS and 480 GB of memory, a deepened Applied Materials collaboration at the EPIC Center, and a reaffirmed 14A node role in Elon Musk's Terafab complex, while FY2025 capital expenditure fell 38.8% year-over-year to $14.65 billion.

The most concrete development to emerge from Intel's October 2026 news cycle is the commercial debut of its Crescent Island AI accelerator under the name GPU D270, offering 1.3 PFLOPS of FP8 compute and up to 480 GB of memory. The product's arrival in the GPU D-Series line marks Intel's most direct bid yet to contest the AI inference market, where memory bandwidth and on-chip capacity have become primary competitive axes. Alongside the launch, Intel and Applied Materials announced an expanded collaboration at Applied's EPIC Center, targeting three-dimensional chip stacking and next-generation AI chipmaking processes — a partnership corroborated across multiple outlets this month, even as Applied Materials shares dipped approximately 1.5%, with investors weighing when the arrangement would translate into manufacturing-side returns.

Intel's foundry ambitions are crystallising around two distinct tracks. The first is its own 14A process node: CEO Lip-Bu Tan has reportedly moved up risk production to the first quarter of 2027, with defect density said to be approaching targets, and Intel has publicly claimed the node will perform within 5% of TSMC's A14. September reporting indicated Nvidia was actively testing 14A for AI chip production — a potentially significant customer qualification if it proceeds. The second track is Terafab, Elon Musk's AI chip complex that one report described as exceeding the combined footprint of the Pentagon, Apple Park, and a major American shopping mall. Lip-Bu Tan confirmed Intel's continued participation in Terafab's cleanroom project and its 14A node arrangement there, even as Musk reaffirmed his own control of the venture and confirmed that talks with TSMC over a potential sublease are underway. According to available reporting, Intel remains the only other publicly named partner in Terafab.

The surrounding industry texture adds further complexity. Reports indicate AI server demand is outrunning available Xeon capacity, a supply squeeze that, if sustained, could support pricing. One outlet reported that both Intel and AMD raised chip prices by approximately 10% in response to rising TSMC manufacturing costs — though this figure comes from a single source and should be treated with appropriate caution. On the security-architecture front, Intel announced plans to phase out SGX after Xeon Diamond Rapids in favour of TDX, consolidating its trusted-execution strategy for cloud and enterprise workloads. The question of Solidigm, Intel's memory subsidiary, is also drawing renewed attention: September reporting put a potential IPO valuation at $150 billion, while more recent coverage indicates SK Hynix is still evaluating its funding options — figures that, given their single-outlet provenance, warrant continued monitoring rather than treatment as settled fact.

The financial backdrop tells a story of deliberate retrenchment after years of aggressive spending. Slicast's own compilation of SEC XBRL data shows Intel's capital expenditure fell 38.8% year-over-year, from $23.94 billion in FY2024 to $14.65 billion in FY2025, against revenue of $52.85 billion — a capex intensity of 28%, still placing Intel second among 14 chip companies tracked by absolute outlay (behind Micron) and second of 15 by intensity (behind TSMC). The stock has had a turbulent month: Intel closed down 5.3% in the 8 October 2026 session at $107.08, though no coverage item offers an explicit explanation for that day's move. The structured data price of $98.08, down 4.6%, indicates the stock has extended its losses since that session, again without a specific catalyst identified in available reporting. A separate data point illustrating the competitive backdrop: Chinese DRAM maker CXMT reportedly surpassed Intel's market capitalisation following a strong Shanghai debut — a symbolic marker, even though the two companies compete primarily in different product segments. Intel's CEO, alongside Broadcom's, was also cited this month as having warned about the sustainability of AI funding flows, contributing to a broader chip-sector pullback.

The near-term picture for Intel is therefore one of genuine but contested momentum. On the opportunity side, the GPU D270 enters a fast-growing inference market; the Applied Materials collaboration deepens Intel's capabilities in EMIB and Foveros advanced packaging — technologies Intel pioneered among large chipmakers between 2017 and 2019; a reported $5 billion Nvidia investment and Nvidia's adoption of Xeon 6, combined with the reported 14A testing, suggest at least one hyperscaler-adjacent player is treating Intel's foundry ambitions seriously; and the Terafab role, however its TSMC-sublease dynamics resolve, keeps Intel embedded in one of the most high-profile AI infrastructure projects currently under construction. On the risk side, the sharp capex compression raises legitimate questions about whether Intel can sustain the fab investment required to close the remaining process gap; CEO-level caution about AI funding sustainability is a signal that demand assumptions may need recalibrating; Nvidia's N1X PC chip and its Vera CPU push into data-centre servers — the latter reported in September — open competitive flanks in both client and server CPU markets; and TSMC's growing Terafab role, if confirmed, could dilute the foundry narrative Intel has built around that project.

Three signals merit close tracking in the months ahead. First, whether 14A risk production proceeds on the reportedly accelerated Q1 2027 schedule, and whether customer-level defect-density data corroborates Intel's own claim of parity within 5% of TSMC's A14. Second, the Terafab manufacturing structure: how the TSMC sublease discussions conclude will determine how much node volume Intel actually captures from that project, materially affecting its foundry revenue trajectory. Third, the Solidigm capital path — whether SK Hynix pursues an IPO, a funding round, or a strategic restructuring will carry direct implications for Intel's balance sheet and its residual exposure to DRAM economics as high-bandwidth memory continues to reshape capacity allocation across the industry.