Intel is making strides in chip innovation by prioritizing power efficiency and cost with its latest designs aimed at artificial intelligence data centers. Recently, the company unveiled its new Xeon 6+ CPU, the Crescent Island GPU, and an Ethernet controller, marking its first major data center releases since Lip-Bu Tan took over as CEO in March 2025.
The technology landscape has prompted Intel to shift its focus. According to Kira Boyko, product line director for Xeon products, customers have expressed concerns over rising costs and energy consumption. As a response, Intel is now emphasizing a balanced approach to cost and performance rather than solely focusing on creating the fastest chips.
The new offerings promise enhanced server consolidation and reduced physical space requirements while improving throughput. Stephen Sopko, an analyst at Hyperframe Research, noted that the explosion of AI applications has strained power, cooling, and operational capacities, making efficiency in these areas increasingly critical.
In line with this focus, Intel has introduced the Crescent Island GPU, specifically designed for the new generation of AI applications, which utilize numerous agents co-ordinating tasks. This GPU operates at 350 watts and omits traditional graphics features to maximize silicon area and AI performance. With 480GB of LPDDR5X memory, chosen for its lower cost and power demand, it aims to sustain the required performance while contending with memory supply constraints.
Intel also launched the Xeon 6+, codenamed Clearwater Forest, utilizing the new 18A manufacturing process to enhance speed and energy efficiency. The design is tailored to support the demands of AI workloads, efficiently distributing tasks among different computing units and reducing the need for high-powered GPUs, which have often been underutilized due to insufficient CPU support.
Additionally, the Xeon 6+ processor features a new Application Energy Telemetry (AET) capability that allows data center operators to monitor energy consumption in real-time and adjust workloads accordingly, potentially offering financial incentives for energy-saving practices.
Another introduction, the E835 Ethernet card, operates at 200Gbps, prioritizing security and power efficiency over raw speed. It’s designed to consume significantly less power than comparable products, aligning with Intel’s goal amidst rising energy demands from AI operations.
Despite these innovations, challenges persist. Intel has struggled to maintain a coherent AI strategy that effectively integrates its offerings across PCs, data centers, and edge technologies. The competition remains fierce, particularly from companies like Nvidia, which dominate the GPU market, and AWS and Google, which are developing their ARM-based CPUs.
Intel will soon announce the ship dates for its new chips as it seeks to regain its leadership in the CPU market while adapting to the changing demands of AI deployments.
For more on Intel’s innovations, check the relevant sources: Intel Official, Trendforce, Counterpoint Research.
