Le projet THINK

Intel researchers developed EASG-Bench, a new benchmark with more than 1,800 question-answer pairs that tests how well AI models understand what happens in first-person videos by using structured scene graphs instead of narrative video descriptions

When optimizing LLM workloads, hardware is only one piece of the equation. Intel brings decades of experience working with software developers, building a strong ecosystem to optimize software for Intel Xeon processors.

The rise of deepfakes has introduced a new dimension of risk in today’s digital landscape. What started as fringe technology is now a mainstream concern for consumers, businesses, and governments alike. For Leena Elias, Chief Product Officer at Gen Digital, the challenge is clear: protect people from increasingly sophisticated threats while making the technology accessible and understandable. Norton’s latest collaboration with Intel is a step in that direction—bringing together security expertise and AI-enabled hardware to redefine how protection works at the device level.

With the launch of the C4 series, Google Cloud now offers access to Intel® Xeon® 6 processor with P-cores which are well-suited for a variety of workloads, including agentic AI systems.

The United Nations (UN) has taken a bold step toward digital sovereignty by developing an open-source AI infrastructure in collaboration with Intel and the Open Platform for Enterprise AI (OPEA).

The vLLM (Virtualized Large Language Model) framework, optimized for CPU inference, is emerging as a powerful solution for efficiently serving large language models (LLMs).

Developed by Intel, KVCrush can improve LLM inference throughput up to 4x with less than 1% accuracy drop.

In the race to operationalize AI, success depends not on flashy pilots, but on turning experimentation into measurable business value. According to David Ellison, Chief Data Scientist and Director of AI Engineering at Lenovo, the most successful AI projects start with clear business outcomes—not models. From cost savings to new revenue streams, the focus is on impact, supported by infrastructure that can scale and systems that users trust.

In this article, we will cover how to deploy high-performance AI inferencing for media data curation and retrieval-augmented generation (RAG) without requiring discrete GPUs.

When it comes to scaling AI, the conversation isn’t only about the cloud—it’s about the edge. According to Matthew Formica, Senior Director and Head of Edge Product Marketing & AI PC/Edge AI Software Developer Relations at Intel, the edge represents one of the company’s fastest-growing opportunities. With more than 200 million processors shipped into edge devices and over 100,000 deployments worldwide, Intel’s edge business is vast, yet often overlooked. The mission now: demonstrate how AI at the edge is quietly shaping everyday life, from retail checkout to robotics-powered manufacturing.