Intel has unveiled a new data center GPU, code-named Crescent Island, specifically designed to handle inferencing workloads.

Unveiled at the 2025 OCP Global Summit, Intel said the AI accelerator will offer high memory capacity and energy-efficient performance.

Crescent Island will be based on the chipmaker’s forthcoming Xe3P architecture – a performance-optimized version of Intel’s Xe3 GPU architecture used in its Panther Lake mobile processors – and offer 160GB of LPDDR5X memory.

The GPU is “power- and cost-optimized” for air-cooled enterprise servers, Intel said, adding that it will offer support for a broad range of data types, making it ideal for “tokens-as-a-service” providers and inference use cases.

“AI is shifting from static training to real-time, everywhere inference - driven by agentic AI,” said Sachin Katti, CTO of Intel. “Scaling these complex workloads requires heterogeneous systems that match the right silicon to the right task, powered by an open software stack. Intel’s Xe architecture data center GPU will provide the efficient headroom customers need -and more value - as token volumes surge.”

Customer samples of the new GPU are expected in the second half of 2026. In the meantime, Intel said it is developing and testing an “open and unified software stack for heterogeneous AI systems” on the company’s Arc Pro B-Series GPUs to “enable early optimizations and iterations.”

Crescent Shores represents a renewed effort from Intel to make headway in the AI accelerator space after its Gaudi offering failed to challenge AMD and Nvidia.

Intel unveiled its Gaudi 3 AI accelerator in April 2024; however, on the company’s Q3 2024 earnings call, former CEO Pat Gelsinger told analysts that uptake of Gaudi has been slower than anticipated and Intel had failed to meet its target of $500 million in revenue for Gaudi in 2024.

Intel is also developing a rack-scale solution, code-named Jaguar Shores, to address the AI data center market. In January, prior to the appointment of CEO Lip-Bu Tan, interim chief executive Michelle Johnston Holthaus, who has since left the company, announced that Intel was canceling Jaguar Shores' predecessor, Falcon Shores, in order to simplify the company’s roadmap and concentrate resources.

At the time, Holthaus said the AI data center market was an attractive proposition for Intel, but admitted the company was not happy with its standing within it.

She continued that while Intel has a “leading position as the host CPU for AI servers,” the company is “not yet participating in the cloud-based AI data center market in a meaningful way.”

Since that announcement, the company has not provided any additional updates on its Jaguar Shores offering.