--- title: "Intel leads with 480 GB of memory for its new AI GPU; its own card carries 160 GB and the stock fell 3.2%" description: "Intel named its first data-center AI GPU since 2021 at the Open Compute Project summit on October 12 and said customer sampling has begun. The 480 GB memory figure it leads with belongs to partner-built cards; Intel's own carries 160 GB, and the shares fell 3.2% on the day." publisher: "The Inference" section: "Silicon" published: 2026-10-13T01:13:33.221Z modified: 2026-10-13T01:13:33.221Z canonical: https://theinference.org/article/intel-leads-with-480-gb-of-memory-for-its-new-ai-gpu-its-own-card-carries-160-gb-and-the-stock-fell-3-2 language: en keywords: "Intel, Nvidia, AMD, GPUs, AI Chips, Data Centers" --- # Intel leads with 480 GB of memory for its new AI GPU; its own card carries 160 GB and the stock fell 3.2% > Intel named its first data-center AI GPU since 2021 at the Open Compute Project summit on October 12 and said customer sampling has begun. The 480 GB memory figure it leads with belongs to partner-built cards; Intel's own carries 160 GB, and the shares fell 3.2% on the day. ## A name, at last, for Intel's AI inference card At the Open Compute Project Global Summit in San Jose on October 12, [Intel](https://theinference.org/markets/companies/intel) named its first commercial data-center AI GPU since Ponte Vecchio in 2021: Intel GPU D-Series, [formerly code-named Crescent Island](https://www.servethehome.com/intels-crescent-island-ai-server-gpu-gets-a-name-and-specs-meet-the-intel-gpu-d-series/) [1][3]. The lead card, the D270, is a 350-watt, air-cooled PCIe card with 1.3 petaflops of FP8 compute and 1.5 TB/s of memory bandwidth, sized for cheap, sustained inference: models up to 300 billion parameters on one card, past a trillion on eight [1]. Customer sampling is underway. There is no price and no launch date [1][3]. ## The 480 GB headline is a partner build The memory number Intel leads with is not on the card Intel sells. [The release's own footnote](https://finance.yahoo.com/technology/ai/articles/intel-accelerates-open-infrastructure-agentic-193200979.html): the Intel-branded PCIe card carries 160 GB of LPDDR5X, and the design "enables partners to build ODM branded cards with flexible options up to 480 GB" [1]. ServeTheHome's spec table shows both configurations; [TechPowerUp](https://www.techpowerup.com/353647/intel-crescent-island-gpu-gets-d-series-name-supports-up-to-480-gb-of-lpddr5x) confirms 160 GB is exclusive to Intel-branded cards [3][4]. The 480 GB figure, the one that beats [AMD](https://theinference.org/markets/companies/amd)'s 432 GB MI450X and [Nvidia](https://theinference.org/markets/companies/nvidia) Rubin's claimed 288 GB on paper, exists only if original design manufacturers build it [5]. *Chart: **Intel's own card carries 160 GB; the 480 GB headline is a partner build** Bar chart of memory capacity per accelerator card: Intel D270 ODM build 480 GB, AMD MI450X 432 GB, Nvidia Rubin 288 GB, Intel D270 Intel-branded card 160 GB* *Memory per accelerator card, GB. Intel figures per the October 12, 2026, release: 160 GB on the Intel-branded card, 480 GB as an ODM design option. AMD MI450X (432 GB HBM4) and Nvidia Rubin (288 GB) as reported by TweakTown for unreleased parts. [1][3][4][5]* ## The rally is CPU money, and the GPU is a fourth attempt Two successors to Ponte Vecchio were canceled on the way here, Rialto Bridge in 2023 and Falcon Shores in 2025, and Crescent Island was unveiled at last year's OCP as an interim, cost-optimized part [3]. ServeTheHome expects server vendors to wait until 2027 for volume [3]. Meanwhile [the stock's 183% run this year](https://theinference.org/article/ai-s-price-wave-reaches-the-processor-intel-up-to-20-qualcomm-10-tsmc-next) rests on the other product line: [revenue hit $16.1 billion in the June 2026 quarter, up 25.4% year over year](https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0000050863), and [Yahoo Finance](https://finance.yahoo.com/technology/article/intel-aims-to-boost-ai-turnaround-with-new-gpu-chips-to-take-on-nvidia-192916083.html) attributes the rally to CPU demand from the AI build-out, not GPUs [8][2]. On launch day the shares fell 3.2% to $101.35 [2][6]. *Chart: **June 2026 revenue of $16.1 billion, up 25.4% year over year, is the rally's engine** Column chart of Intel's quarterly revenue from the September 2023 quarter to the June 2026 quarter, rising from $14.2 billion to a series-high $16.1 billion in the June 2026 quarter, which is highlighted* *Intel revenue by calendar quarter, USD billions, from quarterly figures in its SEC filings. The June 2026 quarter's $16.1 billion is up 25.4% year over year and is the largest of the twelve quarters shown. [8]* ## The open lane against Nvidia's racks The counterargument is real. High-bandwidth memory is scarce and expensive; LPDDR5X, the cheap memory in laptops, buys capacity at a fraction of the power [5]. A 350-watt air-cooled card drops into servers already on the floor, while Nvidia sells inference as [the GB300 NVL72: a fully liquid-cooled rack of 72 Blackwell Ultra GPUs and 36 Grace CPUs](https://www.nvidia.com/en-us/data-center/gb300-nvl72/), moving up to 576 TB/s across the rack, roughly 8 TB/s per chip against the D270's 1.5 [7]. AMD's PCIe rival, the MI350P, peaks at 2.3 petaflops of dense FP8 but at 600 watts [3]. If the metric is tokens per watt in ordinary servers, Intel has a lane, and ASUS, HPE, Dell, Supermicro and several neoclouds are building systems for it [2]. Intel is also driving an OCP workgroup to standardize the sub-rack accelerator card, the form factor Nvidia's racks do not use [1]. ## No price, no launch date, no benchmarks The card needs a price, a launch date and independent benchmarks, none of which exist yet [3][4]. Until one arrives, the GPU leg of the turnaround is a spec sheet with sampling status attached, and holders sitting on a 183% run are underwriting it. Next readings: Intel's third-quarter results in the coming weeks, and any D-Series announcement that carries a number. ## Takeaway Intel's first data-center AI GPU since 2021 is real and sampling, but the 480 GB headline is an ODM option and the card Intel sells carries 160 GB; until a price, a launch date and benchmarks arrive, a 183% stock run still rests on CPUs. ## Sources 1. [Intel press release via Business Wire (Yahoo Finance), 'Intel Accelerates Open Infrastructure for the Agentic AI Era', 12 October 2026](https://finance.yahoo.com/technology/ai/articles/intel-accelerates-open-infrastructure-agentic-193200979.html) 2. [Yahoo Finance, Daniel Howley, 'Intel aims to boost AI turnaround with new GPU chips to take on Nvidia', 12 October 2026](https://finance.yahoo.com/technology/article/intel-aims-to-boost-ai-turnaround-with-new-gpu-chips-to-take-on-nvidia-192916083.html) 3. [ServeTheHome, Ryan Smith, 'Intel's Crescent Island AI Server GPU Gets a Name and Specs: Meet the Intel GPU D-Series', 12 October 2026](https://www.servethehome.com/intels-crescent-island-ai-server-gpu-gets-a-name-and-specs-meet-the-intel-gpu-d-series/) 4. [TechPowerUp, Aleksandar Kostadinov, 'Intel "Crescent Island" GPU Gets D-Series Name, Supports Up to 480 GB of LPDDR5X', 12 October 2026](https://www.techpowerup.com/353647/intel-crescent-island-gpu-gets-d-series-name-supports-up-to-480-gb-of-lpddr5x) 5. [TweakTown, Hassam Nasir, 'Intel "Crescent Island" GPUs set to have up to 480 GB LPDDR5X memory with 32 Xe3P cores', 25 August 2026](https://www.tweaktown.com/news/113252/intel-crescent-island-gpus-set-to-have-up-to-480-gb-lpddr5x-memory-with-32-xe3p-cores/index.html) 6. [MoneyCheck, Oliver Dale, 'Intel (INTC) Shares Drop 3.20% Despite Unveiling Advanced GPU D-Series and Xeon 6+ Infrastructure', 12 October 2026](https://moneycheck.com/intel-intc-shares-drop-3-20-despite-unveiling-advanced-gpu-d-series-and-xeon-6-infrastructure/) 7. [NVIDIA, 'NVIDIA GB300 NVL72' product page, accessed 13 October 2026](https://www.nvidia.com/en-us/data-center/gb300-nvl72/) 8. [Intel, quarterly figures from its SEC filings, retrieved 13 October 2026](https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0000050863) --- Intel leads with 480 GB of memory for its new AI GPU; its own card carries 160 GB and the stock fell 3.2% — The Inference. Canonical: https://theinference.org/article/intel-leads-with-480-gb-of-memory-for-its-new-ai-gpu-its-own-card-carries-160-gb-and-the-stock-fell-3-2