ASUS Enters the Intel Arc Pro GPU Market With a 32GB Battlemage AI Graphics Card
ASUS is stepping deeper into the professional AI hardware space with the launch of its first Intel Arc Pro graphics card, the ASUS UGen Intel Arc Pro B70. Built around Intel’s high-end Battlemage BMG-G31 GPU, this new workstation card is designed for artificial intelligence, local large language models, content creation, and professional compute workloads.
The move is notable because Intel Arc graphics cards have not received the same broad support from major add-in-board manufacturers as NVIDIA GeForce, NVIDIA RTX Pro, AMD Radeon, and Radeon Pro products. Until now, Intel Arc cards have mostly appeared from a smaller group of partners, while bigger brands have been more cautious. ASUS previously experimented with Intel’s early Xe-based DG1 graphics hardware, but it has been absent from the consumer and professional Intel Arc scene for the last two generations.
That changes with the ASUS UGen Intel Arc Pro B70, a graphics card aimed directly at the fast-growing AI workstation market.
The card uses Intel’s Xe2 architecture and comes equipped with 32 Xe-Cores and 256 XMX AI engines. ASUS lists the card with up to 367 TOPS of INT8 AI performance, giving it strong appeal for developers, researchers, and businesses running AI inference workloads locally. With AI demand continuing to rise, ASUS appears to be positioning this model as an alternative to more expensive workstation GPUs from competing brands.
One of the biggest highlights of the ASUS UGen Intel Arc Pro B70 is its 32GB of video memory. That large VRAM capacity makes it better suited for AI models, professional visualization, and memory-heavy compute tasks than typical mainstream graphics cards. ASUS also emphasizes multi-GPU scalability, with support for up to four cards in a single system. In that configuration, users can access up to 128GB of total VRAM, potentially enabling local AI workloads with models reaching up to 120 billion parameters.
The graphics card uses a dual-slot blower-style cooler, a design commonly preferred for workstation and server environments where multiple GPUs need to be installed closely together. Instead of pushing hot air around the chassis, the blower cooler helps exhaust heat more efficiently, making it practical for dense multi-GPU setups.
For display connectivity, the ASUS UGen Intel Arc Pro B70 includes four outputs: one HDMI port and three DisplayPort connectors. The card has a maximum total board power of 310W, placing it firmly in the high-performance professional GPU category.
ASUS is clearly marketing this card toward AI developers and workstation users rather than gamers. The company highlights compatibility with several major AI frameworks and development tools, including vLLM for XPU, llama.cpp through SYCL and Vulkan, Ollama, PyTorch with Intel Extension for PyTorch, OpenVINO, oneAPI, and Vulkan-based workloads. This software support is important because AI hardware is only as useful as the ecosystem around it, and Intel has been working to strengthen its developer stack for Arc Pro GPUs.
The ASUS UGen Intel Arc Pro B70 is also listed as compatible with ASUS Zenni Claw, suggesting that ASUS is building a broader professional AI platform around this GPU rather than releasing it as a standalone workstation card only.
ASUS has shared performance results for the card across AI-focused benchmarks and workloads, including MLPerf Client, agentic AI testing, and scalable local AI configurations using multiple cards. These tests are intended to show how the Arc Pro B70 can perform in real-world AI environments, especially when users need to run models locally instead of relying entirely on cloud infrastructure.
Pricing has not been officially confirmed for the ASUS version, but Intel Arc Pro B70 32GB graphics cards generally appear in the range of roughly $1,299 to more than $2,000 depending on the model, region, and configuration. While that is still a significant investment, the large 32GB memory capacity and AI-focused performance could make it attractive compared with some competing professional GPUs that cost much more.
The arrival of an ASUS-branded Intel Arc Pro B70 is an important step for Intel’s discrete GPU ambitions. Having a major hardware partner enter the Arc Pro market gives Intel more visibility in the workstation and AI segment, where demand is rapidly expanding. It also gives system builders and professionals another option when configuring AI workstations, local inference machines, and multi-GPU compute systems.
If this product performs well, it could encourage ASUS to continue supporting future Intel discrete GPUs. With Intel expected to keep developing professional GPU hardware, ASUS may eventually expand its Arc-based lineup beyond this first Battlemage workstation card.
For now, the ASUS UGen Intel Arc Pro B70 stands out as a serious AI-focused graphics card with 32GB of VRAM, strong INT8 compute performance, multi-GPU support, and a compact dual-slot design. As more businesses and creators look for powerful local AI hardware, ASUS entering the Intel Arc Pro ecosystem could make the professional GPU market more competitive.
