
Intel debuts Arc Pro B60 and B50 GPUs built on Xe2 architecture, tailored for AI, engineering, and creative workloads with up to 24GB of memory
Gaudi 3 AI accelerators now support PCIe and rack-scale deployments, enabling scalable inferencing and efficient performance with liquid cooling
Intel AI Assistant Builder goes public in beta, allowing developers to create and run custom local AI agents on Intel-based systems
Computex 2025 was an absolute sight to behold as Intel took the spotlight, armed not just with ambition but with serious silicon strength. With calculated precision, the tech giant unveiled its latest weapons for the professional battlefield: the Arc Pro B60 and B50 GPUs.These aren’t consumer-grade cards chasing frame rates. They are purpose-built for professionals who live deep in CAD workflows, media production pipelines, and AI model training grounds.
The Arc Pro B60 and B50 deliver ECC memory support, certification for major ISVs, and the kind of horsepower engineers, architects, and data scientists demand.
At the heart of this reveal are the Arc Pro B60 and B50, both anchored in Intel’s new Xe2 architecture. These GPUs bring Xe Matrix Extensions (XMX) AI cores and advanced ray tracing units, purpose-built for professionals who run heavy workloads in fields like architecture, engineering, media, and data science.
Arc Pro B60 packs 24GB of memory, offering room for large datasets and demanding simulations.
Arc Pro B50, slightly leaner with 16GB, still delivers solid horsepower for most workstation tasks.
Both GPUs support multi-GPU scaling, letting users stack performance as needed.
What sets them apart isn’t just raw specs. It’s thoughtful and flexible. They’re compatible with consumer and pro drivers on Windows and feature a containerized stack for Linux, which simplifies deployment for teams working on AI projects or professional design applications.
Also Read: AI and Machine Learning in Blockchain Security: Enhancing or Threatening Decentralization?
Intel has sharpened its focus on real-world workflows. From AEC professionals (architecture, engineering, construction) to creative teams building high-end visual content, the Arc Pro GPUs aim to plug straight into existing pipelines without friction. And with ISV certifications across key applications, these GPUs come with assurance, not guesswork.
Sampling of the B60 begins in June 2025 through partners like ASRock, Gunnir, and Sparkle. The B50 will follow in July, with wide availability by late 2025.
If the Arc Pro series covers the desk, Intel’s Gaudi 3 AI accelerators are meant for the data center floor. At Computex, Intel unveiled two key formats:
Gaudi 3 PCIe cards: Slot into existing server environments. Perfect for businesses wanting high powered inferencing without a complete overhaul.
Gaudi 3 Rack Scale Systems: These beasts support up to 64 accelerators per rack, with a massive 8.2TB bandwidth memory. And yes, they use liquid cooling to keep thermals and costs under control.
Even when running Llama 3.1 models or the ambitious Llama 4 Scout, Intel says its accelerators offer performance with scalability, minus the vendor lock in. Everything is designed around modularity and open standards, appealing to cloud service providers and enterprises tired of rigid infrastructure.
Another star at Intel’s booth is the AI Assistant Builder. First teased at CES 2025, it is now live on GitHub in public beta. Designed for Intel-based PCs, this tool helps developers create tailored AI agents that run locally, without relying on third-party clouds. It is open, lightweight, and built to serve real-time, on-device use cases.
Developers can use it to create productivity tools, automate tasks, or enhance enterprise software, all with privacy and speed in mind.
Also Read: How Much Does It Cost to Implement an AI Chatbot?
This wave of launches arrives as Intel celebrates 40 years of partnership in Taiwan, a fitting backdrop to a future-focused reveal.
CEO Lip Bu Tan emphasized more than just the product lineup. He reaffirmed Intel’s focus on building an open and collaborative ecosystem. He expressed confidence that working together would lead to products that truly resonate with customers and unlock new opportunities in the evolving tech landscape.
At booth J0306 in the Nangang Exhibition Center, Intel has made it clear. It is no longer just about CPUs. The company is playing big in graphics, accelerating data workflows, and helping developers build smarter solutions with fewer limitations.
Professionals now have GPU options tailored for real work, not just gaming.
Scalable AI solutions from desktop to data center are no longer a pipe dream.
Intel is proving it is ready to challenge both established GPU giants and next-gen AI hardware startups.
Whether you are running high-fidelity simulations, crunching neural nets, or managing multi-display workstations, these GPUs are designed to handle mission critical workloads with rock-solid stability. They might just be the answer to all of the regular processor concerns.