Intel has introduced a 160 GB, energy-efficient data center GPU as part of a new annual release cycle that reflects its evolving AI hardware strategy focused on open systems and software integration. The new GPU, code-named Crescent Island, is optimized for inference workloads running on air-cooled enterprise servers and represents Intel’s commitment to developing scalable, cost-effective solutions for the data center AI market. The chip features Intel’s Xe3P microarchitecture, designed to enhance performance-per-watt, and supports a variety of data types alongside 160 GB of LPDDR5X memory.
According to Intel, Crescent Island is engineered to deliver optimal power and cost efficiency for enterprise-level inference tasks, with sampling for customers expected to begin in the second half of 2026. Until its release, Intel will continue refining its open and unified software stack for heterogeneous AI systems, leveraging its Arc Pro B-Series GPUs to test optimizations and improve compatibility across diverse hardware environments. The initiative forms a key component of Intel’s broader effort to strengthen its AI infrastructure presence after years of limited success in the accelerator chip segment.
Crescent Island was showcased at the 2025 OCP Global Summit, marking the official start of Intel’s new yearly cadence of GPU launches. This move follows similar strategies adopted by competitors Nvidia and AMD, as major semiconductor players increasingly focus on frequent hardware updates to meet the growing computational needs of AI workloads. Intel’s announcement also aligns with its recent press campaign surrounding its upcoming Panther Lake and Clearwater Forest CPUs, further demonstrating its intent to position itself as a competitive force in the AI computing space.
During a recent media briefing, Intel’s Chief AI and Technology Officer, Sachin Katti, described Crescent Island as an ideal fit for token cloud and enterprise-level inference tasks, highlighting its enhanced memory bandwidth and large capacity. Katti emphasized that the company’s future strategy revolves around building scalable, heterogeneous systems that deliver the best performance-per-dollar for AI workloads. He explained that Intel’s open architecture approach will provide customers and partners with greater choice at both the systems and hardware layers, enabling collaboration among multiple vendors and technologies.
Katti, appointed earlier this year by Intel CEO Lip-Bu Tan, outlined the company’s renewed focus on openness and interoperability within the AI hardware ecosystem. He stated that this architecture will allow Intel to integrate emerging and disruptive technologies as they evolve, ensuring that its hardware and software systems remain adaptable to future requirements. While Intel did not provide updates on Jaguar Shores, its next-generation GPU designed for rack-scale platforms, the announcement of Crescent Island reaffirms the company’s intent to establish a consistent, transparent product roadmap and regain momentum in the data center acceleration market dominated by Nvidia.
With the Crescent Island GPU, Intel is positioning itself as a strong alternative for enterprises seeking energy-efficient, high-performance inference hardware built on open principles. The move underscores Intel’s shift toward creating a more flexible and collaborative AI ecosystem that balances cost, performance, and sustainability across modern data infrastructure.
Follow the SPIN IDG WhatsApp Channel for updates across the Smart Pakistan Insights Network covering all of Pakistan’s technology ecosystem.
