Urgent.News

What's breaking now, across thousands of outlets.

Tech

Diamonds, Crescents and Wildcats: Intel shows off its hardware for the next generation of agentic AI workloads

Intel laid out its whole agentic AI hardware stack at Hot Chips 2026, but two of the three chips are still a year or more from shipping.

Diamonds, Crescents and Wildcats: Intel shows off its hardware for the next generation of agentic AI workloads

Intel showcased three agentic AI hardware designs during its Hot Chips 2026 presentation: the Diamond Rapids Xeon CPU, Crescent Island GPU, and Wildcat Lake client SoC. The Diamond Rapids, also known as Xeon 7, boasts up to 256 cores and 1.28 GB of fast L3 cache. In contrast, Crescent Island, an inference-optimized GPU, is built on the Xe3P architecture and features 32 cores, 256 XMX engines, and 32 MB of L2 cache.

Interestingly, the GPU utilizes LPDDR5X memory rather than HBM, allowing for greater capacity at a lower cost and power consumption. However, Intel's Wildcat Lake SoC is currently the only design available for purchase, with a 17 TOPS NPU that falls short of the 40+ TOPS required for Microsoft's Copilot+ certification. While Diamond Rapids aims to match AMD's EPYC Venice in terms of core count and bandwidth, the company's upcoming version, Xeon 7, will have 50% fewer threads upon release in 2027.

Crescent Island's memory capacity is also limited to 160 GB for the Intel-branded version, despite supporting up to 480 GB via third-party partner configurations. Intel's packaging shift from Foveros to an organic multi-chip design resulted in a 38% reduction in compute die area and a 15% decrease in I/O tile area, contributing to cost savings and yield improvements.

Written by urgent.news from TechRadar's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

Read the original at techradar.com →

More in Tech

After Building Bulk Messaging, I Became More Cautious About It

Bulk messaging sounds like a straightforward productivity feature: write once, send to many people, and save a large amount of repetitive work.

  • Bulk messaging, initially seen as simple, introduced complex considerations for MSG.AI.
  • Responsible messaging is a better product strategy, focusing on control, relevance, and usefulness.

Why I Built a Browser Extension for WhatsApp Web

I did not start building MSG.AI because I had discovered a grand new AI opportunity. It began with a much smaller problem—one that kept repeating every day. A message arrives in another language.

  • MSG.AI extension integrates directly into WhatsApp Web to streamline translation and messaging.
  • Prioritizes data privacy and security through local storage and authorized AI processing.

More from Saturday 29 August →