Intel has announced a major expansion of its Arc Pro lineup at Computex 2025, directly targeting the Artificial Intelligence and professional visualisation sectors that have been dominated by NVIDIA and AMD. The new Arc Pro B60, built on the Battle Range architecture, introduces a scalable multi-GPU platform tailored for high-performance Artificial Intelligence tasks and enterprise-grade inference workstations, with a strong focus on edge computing. This strategic move signifies Intel’s ambition to solidify its presence in the GPU market for professional and enterprise workloads.
Key innovations within the Arc Pro B60 include 24GB of VRAM per GPU, with support for memory pooling across GPUs to deliver up to 192GB of VRAM—a capacity capable of handling Artificial Intelligence models with up to 70 billion parameters. Intel’s proprietary xMEX engine streamlines memory management, minimising bottlenecks during inference. Utilizing PCIe 5.0 and the latest Xeon platforms, Intel claims that B60 delivers up to 50% higher Artificial Intelligence inference performance than NVIDIA’s RTX A1000, offering tangible value for enterprise users. Intel also diverges from NVIDIA’s more rigid approach by working with board partners to offer the B60 in multiple form factors, enabling versatile deployment tailored to specific enterprise needs.
Intel introduced an Artificial Intelligence-driven debugging architecture as part of its Inference Workstations. By harnessing agentic structures—featuring triage, debugging, design, and auto-testing agents—Intel slashes debugging and deployment times from days to hours. These solutions, paired with containerised software like IPEX and OneAPI, offer an end-to-end stack optimised for large language models and scalable Artificial Intelligence applications, reducing latency, bolstering data security, and cutting operational overhead. The Arc Pro software roadmap includes phased driver releases and certifications starting in Q2 2025, with system rollouts by board partners projected for Q3. Intel’s enterprise push is amplified through collaborations with seven major system integrators currently sampling Arc Pro B50 and B60 GPUs.
Intel’s aggressive expansion aims at capturing demand for on-premises Artificial Intelligence processing, sidestepping cloud dependency and meeting enterprise requirements for speed, privacy, and cost-effectiveness. With edge-optimised variants planned and a comprehensive ecosystem strategy, Intel is positioned to compete more directly with NVIDIA and AMD in the market for Artificial Intelligence-driven inference and workstation solutions, signaling a potential shift in the enterprise GPU landscape.
