Intel Corp (NASDAQ:INTC, ETR:INL) on Tuesday unveiled a new data center graphics processing unit (GPU), codenamed Crescent Island, and a rack-scale reference design based on its Gaudi 3 systems, marking a further push into the artificial intelligence (AI) market.
The move comes as Intel seeks to catch up with rivals Nvidia Corp (NVDA) and Advanced Micro Devices Inc (AMD) in the AI space.
CEO Lip-Bu Tan has said the company will focus on AI inference and agentic AI, which involves running AI models in real time.
“AI is shifting from static training to real-time, everywhere inference — driven by agentic AI,” Intel chief technology officer Sachin Katti said.
“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.”
Intel did not disclose whether Crescent Island will use its 18A process node but said the GPU will employ the company’s Xe3P microarchitecture and feature up to 160GB of memory.
Crescent Island chips are expected to begin customer sampling in the second half of 2026.
Wedbush analysts noted that while Intel’s AI chips could offer efficiencies when used alongside Nvidia GPUs for certain workloads, it is cautious on the likelihood of Intel generating significant direct revenues from AI accelerators in the near term.
“While it makes sense to us that CSPs or model builders for instance should be able to tailor chip design to their specific workloads creating efficiencies beyond using more standard offerings, we would think it would be more difficult for a third party to come up with a more ubiquitous enhancement tool,” analysts wrote.
Intel shares have risen 84% year to date and 57% over the past 12 months.