Intel Unveils Crescent Island AI Chip to Reignite Data Center Ambitions

Intel has announced its newest artificial intelligence chip, Crescent Island, set to launch next year as part of the company’s renewed push into the competitive AI chip market. Revealed by Intel Chief Technology Officer Sachin Katti at the Open Compute Summit, the new GPU focuses on energy efficiency and cost-effective performance for running AI inference applications.

Katti emphasized that Crescent Island is designed to deliver the best “performance per dollar” while optimizing for AI workloads. The launch represents Intel’s latest effort to regain ground in the booming AI sector dominated by Nvidia and AMD. Both competitors have seen massive revenue growth driven by demand for GPUs powering generative AI systems like ChatGPT.

Crescent Island will include 160 gigabytes of slower memory compared to the high-bandwidth memory (HBM) found in Nvidia and AMD’s chips. Based on Intel’s existing consumer GPU architecture, the chip underscores the company’s strategy of balancing cost and performance for data center clients. Intel did not disclose details about its manufacturing process.

Under CEO Lip-Bu Tan, Intel aims to revitalize its AI roadmap after pausing projects such as the Gaudi and Falcon Shores chips. The company now plans to release new data center AI chips annually, aligning with AMD, Nvidia, and major cloud providers’ release cycles.

Intel’s modular “mix and match” approach allows customers to integrate chips from multiple vendors, giving data centers more flexibility. The move comes amid surging demand for GPUs, leading to global shortages and skyrocketing prices.

In a strategic partnership, Nvidia recently invested $5 billion in Intel, taking a 4% stake to co-develop future PC and data center chips. According to Katti, this collaboration will help ensure Intel CPUs remain a core component in every AI-driven system sold, strengthening its role in the evolving AI infrastructure market.