Intel Targets AI Creators with New Arc Pro GPUs and Gaudi 3 Accelerators at Computex

Chipmaker Intel announced its new line of GPUs, the Intel Arc Pro, aimed squarely at AI PCs and professional workstations on Monday at Computex in Taiwan. In addition, Intel now makes its Gaudi 3 AI accelerators in PCIe and rack-scale versions for enterprise and cloud AI services. It also revealed that its AI Assistant Builder software, meant to help devs create local AI agents, is publically available on GitHub.

"For the past 40 years, the power of our partnership with the Taiwan ecosystem has fueled innovation that has changed our world for the better," said Lip-Bu Tan, CEO of Intel, in a statement. "This week, we are renewing our commitment to our partners as we work to build a new Intel for the future. Together, we will create great products that delight our customers and capitalize on the exciting opportunities ahead."

The new Intel Arc Pro B60 and B50 GPUs are based on the Xe2 architecture (the updated Xe Gen 12 chips) and feature Intel Xe Matric Extensions AI cores and advanced ray tracing units to power the next generation of gaming and AI. The Pro B60 offers 24 GB memory, while the Pro B50 has 16 GB, and though Intel didn't release pricing details, it's likely the latter will hit a lower price point.

These B-series GPUs are compatible with consumer and pro drivers on Windows, said Intel, and support a containerized software stack to simplify AI deployment on Linux, as well.

Intel also revealed a new high-end workstation code-named Project Battlematrix that supports up to eight Intel Arc Pro B60 GPUs to run medium-sized AI models.

"The Intel Arc Pro B-Series showcases Intel's commitment in GPU technology and ecosystem partnerships," Vivian Lien, vice president and general manager of Client Graphics at Intel, said in a statement. "With Xe2 architecture's advanced capabilities and a growing software ecosystem, the new Arc Pro GPUs deliver accessibility and scalability to small and medium-sized businesses that have been looking for targeted solutions."

The company also revealed PCIe and rack-scale Gaudi 3 AI accelerator chips, NVIDIA competitors that help train AI models faster while also speeding up AI inference in practice. These can be used to train large language models, run AI inference across cloud or edge environments, and power enterprise AI workloads for recommendation systems, computer vision, or advanced analytics.

Finally, Intel's AI Assistant builder is available to developers via GitHub. Originally debuted at CES 2025, this open software framework helps devs build and run custom AI agents locally on Intel-based AI PCs.