Intel Arc Pro B60 and B50 GPUs launched for AI and creators at Computex 2025 | Tech News

Tech News

Intel announced new Arc Pro B60 and B50 GPUs and Gaudi 3 AI accelerators at Computex 2025. The new hardware is designed for AI inference, workstations, and large-scale data center use. Intel also made its AI Assistant Builder tool publicly available via GitHub.

Intel Computex 2025: Arc Pro B60 B50 GPUs, Gaudi 3 PCIe AI cards, AI Assistant Builder

New Delhi: Ahead of Computex 2025, Intel rolled out a major update to its professional graphics and AI accelerator portfolio, marking its 40th year of partnership with Taiwan’s tech ecosystem. The announcements, made just a day before the industry trade show kicks off, include the new Arc Pro B60 and B50 GPUs and expanded deployment options for the Gaudi 3 AI accelerators.

With AI workloads scaling across industries, Intel is positioning these launches to appeal to developers, engineers, and enterprise AI teams looking for flexible, performance-driven compute solutions.

Arc Pro B60 and B50: Built for AI and workstation performance

Expanding the Arc Pro lineup, Intel introduced the Arc Pro B60 and B50 graphics cards. Both models are based on Intel’s Xe2 architecture and feature XMX AI cores along with advanced ray tracing capabilities. The B60 offers 24 GB of memory, while the B50 comes with 16 GB. These GPUs are targeted at AI inference workloads and professional workstation use cases, including AEC (architecture, engineering, and construction) applications.

“These new Arc Pro GPUs deliver accessibility and scalability to small and medium-sized businesses that have been looking for targeted solutions,” said Vivian Lien, vice president and general manager of Client Graphics at Intel.

The B60 will begin sampling through board partners such as ASRock and Sparkle in June, while the B50 will be available from authorised resellers in July.

Gaudi 3 AI accelerators add PCIe and rack-scale options

Intel also unveiled new deployment models for its Gaudi 3 AI accelerators. The Gaudi 3 PCIe cards support inferencing for models ranging from Llama 3.1 8B to Llama 4 variants, making them suitable for both small businesses and enterprise-scale workloads. These cards are expected to ship in the second half of 2025.

For hyperscale customers, Intel’s new Gaudi 3 rack-scale reference designs support up to 64 accelerators per rack with 8.2 TB of high-bandwidth memory. The open, modular design aims to reduce vendor lock-in and improve energy efficiency using liquid cooling.

Intel AI Assistant Builder now publicly available

In addition to hardware, Intel announced the public beta of the AI Assistant Builder, a lightweight open-source toolkit for creating on-device AI agents. Available on GitHub, the tool is designed to work on Intel-based AI PCs and saw early adoption through partners like Acer and ASUS during the show.

The announcements reflect Intel’s larger strategy to re-establish its relevance in the high-performance AI and GPU markets. CEO Lip-Bu Tan said, “We are renewing our commitment to our partners as we work to build a new Intel for the future.”

Siddharth Shankar brings over a decade of experience in the realm of journalism, film making, and marketing, currently heading the Technology, Science, Gaming, and Automobile sections. He blends his expertise and passion to bring forth cutting-edge stories that resonate with a dynamic audience. Siddharth co-founded BWAutoworld.com for BW Businessworld and was leading Science and Technology domains for TimesNow Digital. For leads Siddharth can be reached on Twitter – @Siddh4rth

Read More

Next Article

Follow us on social media

_{Images are for reference only.Images and contents gathered automatic from google or 3rd party sources.All rights on the images and contents are with their legal original owners.}