Intel Announces “Crescent Island” Inference-Optimized Xe3P Graphics Card With 160GB vRAM

Back during the Intel Tech Tour in Arizona, Intel teased a new inference-optimized enterprise GPU would be announced soon. This new product would feature enhanced memory, bandwidth, and enterprise-level AI inference capabilities. Today the embargo expires on talking about this new GPU offering.

When Intel was teasing this new inference-optimized GPU a few weeks back in Arizona it sounded like Intel may have had an unexpected trick up its sleeves. What’s being announced today is indeed a new enterprise GPU for AI that is interesting from a technology perspective, but it’s not shipping until at least H2’2026. So while there was hope that perhaps Intel had managed to innovate some interesting Battlemage / BMG-G31 part for AI or the like with lots of vRAM, what’s being announced is a next-gen part but one that is at least one year away still.

This new graphics card is codenamed Crescent Island and is built on their next-gen Xe3P Celestial micro-architecture. Xe3P will be optimized around performance-per-Watt and Crescent Island will feature 160GB of LPDDR5x memory to allow for plenty of space for large language models (LLMs).

Intel’s embargoed announcement also notes that Crescent Island will feature support for a variety of different data types and be an “ideal” solution for tokens-as-a-service providers and inference use cases.

In addition to being optimized around performance-per-Watt, Crescent Island will also be air cooled and cost-optimized. Intel is currently working on refining their open-source software stack for Crescent Island via using current-generation Arc Pro B-Series GPUs.

Intel’s announcement notes that customer sampling of this new data center GPU will begin in the second half of 2026. No official release timeframe was provided if they also hope to squeeze it out next year or if (more than likely) it will actually ship more broadly in 2027 but just noting their customer sampling for H2’2026 in the embargoed news release. No slides or prototype images or anything else to share today on Intel’s Crescent Island.

Long story short, Intel is announcing Crescent Island today as a Xe3LP + 160GB LPDDR5X offering for H2’2026 or later that will be AI inference optimized around power efficiency and cost. It sounds interesting but technical details beyond those basics were light and it’s going to be a long while before we see Crescent Island. Given the timing this will be going up against the AMD Instinct MI450 series and NVIDIA Vera Rubin. It seems like Intel wanted to have something to announce now given the ongoing AI rush albeit not many details today and no short term AI solution.

At least this does lead to more weight for the ongoing Project Battlematrix Linux driver improvements and other ongoing Intel Compute Runtime and Intel Xe Linux driver enhancements that are currently ongoing for the Arc Pro B-Series. With confirming Crescent Island now it also opens the door to them beginning to push open-source hardware enablement patches without otherwise spilling the beans on this forthcoming enterprise AI product.

Intel is using the OCP Global Summit to announce some additional Gaudi 3 rack-scale reference designs. These new Gaudi 3 rack-scale reference designs will allow up to 64 accelerators per rack with liquid cooling and 8.2TB of high bandwidth memory. Intel hasn’t aggressively promoted Gaudi 3 in recent quarters after its launch last year. Gaudi 3 has enjoyed some reprieve since Intel canceled their Falcon Shores AI accelerator chip but still appears to be the end of the road for Gaudi especially with Jaguar Shores still expected and now Crescent Island too. The Gaudi 3 software support has been neglected over the past year with losing multiple rounds of the Habana Labs Linux driver maintainers and only recently seeing new activity to return to working on this AI accelerator Linux driver albeit as of writing for Linux 6.18 there still is no mainline kernel driver support for Gaudi 3.