ASUS Launches Arc Pro B70 32GB Graphics Card
The new workstation graphics card brings 32GB of ECC memory to professional markets in Europe.
Updated on Oct. 26, 2026 in Semiconductors
Live Poll
Do you believe premium pricing for professional hardware is justified by regional taxes and design features?
ASUS has released its Arc Pro B70 32GB graphics card in Germany, Austria, and Switzerland. The new hardware is specifically designed for local AI inference and professional workloads requiring high data integrity.
Why it matters
The inclusion of 32GB of ECC GDDR6 memory allows professional users to handle data-intensive tasks with increased stability. The card is positioned to support demanding AI and workstation applications that require reliable error correction.
The Arc Pro B70 features 32 Xe2 cores and 256 XMX engines, with a dual-slot blower cooling design. The board is rated for 310W, providing 32GB of ECC GDDR6 memory in a single-card configuration.
The players
ASUS
ASUS is a multinational hardware manufacturer known for producing computer components, workstations, and consumer electronics.
The details
Workstation configurations can scale up to 128GB of total graphics memory by combining four cards, at a total investment of €7,196. The blower design ensures heat is exhausted directly through the rear of the workstation chassis.
Timeline
ASUS confirmed the regional pricing and availability for the B70 in October 2026.
The Tech Race
This release advances the Intel Arc Pro series architecture by offering a high-capacity 32GB configuration for professional tasks. It represents a shift toward equipping specialized workstations with higher memory density to compete in the professional GPU market.
Professional users can purchase the card for €1,799 in Germany and Austria or CHF 1,799 in Switzerland. The high-capacity memory configuration allows for local AI model inference that may previously have required expensive cloud-based solutions.
The takeaway
The Arc Pro B70 provides a scalable solution for professionals needing high memory capacity for AI tasks. Users should consider cooling requirements and power overhead when planning a multi-card workstation build.
Further reading
Explore the latest developments in professional hardware within the Semiconductors section.
Live Poll
Do you believe premium pricing for professional hardware is justified by regional taxes and design features?
