ASUS Enters Intel Arc Market With 32GB B70 GPU Targeting AI Inference and Professional Workloads

ASUS Returns to Intel GPU Market After Two Generations

ASUS has unveiled the UGen Intel Arc Pro B70, marking its first Intel Arc graphics card after remaining absent through the Alchemist and consumer Battlemage generations2

2

. While ASUS previously released an Intel DG1 card in 2021, this represents the company's first Arc-branded discrete model2

2

. The move signals growing confidence in Intel's GPU ecosystem as demand for local AI inference hardware accelerates across professional markets.

Built on Intel Battlemage GPU Architecture

The ASUS UGen Intel Arc Pro B70 utilizes Intel's Big Battlemage BMG-G31 GPU based on the Xe2 architecture3

3

. The workstation graphics card features 32 Xe cores and 256 XMX AI engines, with graphics clocks reaching 2,800MHz1

1

. FP32 throughput is listed at 22.94 TFLOPS, while AI inference performance delivers up to 367 INT8 TOPS1

1

. This performance metric applies specifically to the card's XMX matrix engines and differs from consumer GPU TOPS ratings due to variations in precision, sparsity, and software implementation1

1

.

32GB ECC Memory Configuration for Professional Workloads

The card equips 32GB of ECC-enabled GDDR6 memory across a 256-bit interface, providing 608GB/s of memory bandwidth1

1

. ASUS specifies 19Gbps memory speed for the configuration1

1

. The ECC capability ensures data integrity during AI and workstation use cases, distinguishing it from gaming-focused cards1

1

. Power consumption reaches 310W, exceeding Intel's reference B70 card rating of 230W, though ASUS has not published benchmarks demonstrating corresponding performance advantages2

2

. The board uses a 12V-2×6 connector and ASUS recommends a 750W system power supply1

1

.

Dual-Slot Blower Cooler Design for Multi-GPU Density

ASUS implemented a dual-slot blower cooler design optimized for densely packed workstation systems2

2

. The blower-style cooler exhausts hot air through the rear I/O bracket rather than recirculating it inside the chassis, making it suitable for multi-GPU configurations1

1

. This design choice prioritizes thermal management in compact environments over acoustic performance, as blower coolers typically generate more noise than open-air designs, particularly under continuous AI loads1

1

.

Multi-GPU Configurations Scale to 128GB VRAM

The card supports multi-GPU configurations with up to four cards, providing 128GB of combined ECC graphics memory1

1

. ASUS highlights this capability for running AI models up to 120-billion-parameter workloads2

2

. However, this configuration does not function as a single 128GB VRAM graphics card—performance depends on model quantization, framework support, and interconnect behavior2

2

. Display connectivity includes three DisplayPort 2.1 UHBR13.5 outputs and one HDMI 2.1b output, supporting up to four displays at resolutions reaching 7680×4320 at 120Hz1

1

.

Software Support Spans Major AI Frameworks

ASUS emphasizes compatibility with OpenVINO, oneAPI, PyTorch, llama.cpp, Ollama, vLLM, and Vulkan1

1

. The company shared benchmarks across AI workloads including MLPerf Client and Agentic AI, demonstrating scalable local AI inference results with up to four cards3

3

. The card also integrates with ASUS Zenni Claw3

3

. This broad framework support positions the card for content creation and professional workloads beyond gaming applications2

2

.

Pricing and Availability Remain Unannounced

ASUS has not confirmed pricing or retail availability for the UGen Arc Pro B701

1

. Intel introduced the B70 workstation GPU in March with a $949 starting price2

2

. Third-party Intel Arc Pro B70 32GB variants typically sell between $1,299 and over $2,0003

3

. Earlier gaming benchmarks showed the B70 performing close to NVIDIA's RTX 5060 Ti 16GB at 1440p despite its professional focus2

2

. The card's value proposition centers on its 32GB ECC memory, blower design, and professional software support rather than gaming performance1

1

. ASUS's entry into Intel's professional GPU ecosystem reflects rising demand for local AI hardware and could signal future Arc designs if Intel continues discrete GPU development3

3

.