MAXSUN has introduced its Intel Arc Pro B70 series, succeeding the B60 lineup and aimed at professional workloads, AI computing, and multi-GPU setups. The new series combines higher hardware specifications with a tightly integrated software ecosystem developed in collaboration with Intel.
The B70 series features 32 GB of VRAM and 32 Xe cores, designed to handle large-scale AI models, advanced rendering, and professional visualization tasks. Users can expect stable performance across intensive workloads without memory limitations. Intel Arc Pro’s environment variable configuration reduces memory usage during dynamic quantization. This allows AI inference and professional applications to run efficiently while reducing system memory requirements and avoiding extra hardware costs.
With PCIe 5.0 x16 support and a dual-slot design, the B70 series improves latency and data exchange in multi-GPU setups. This makes it easier to deploy high-density compute clusters in workstations or server environments. The cards provide up to 608 GB/s memory bandwidth and 367 TOPS (INT8) of peak AI compute power, supporting AI inference, video encoding, and complex rendering tasks with consistent performance.
The B70 series includes native PyTorch support and ISV certifications to ensure stability across professional applications. Multi-Arc optimization and Docker containerization simplify the deployment of large-scale AI models and other workloads. Each card supports 3× DisplayPort 2.1 and HDMI 2.1a, enabling multi-display setups with high-quality, uncompressed visuals for professionals working with large datasets or visual projects.
- Turbo Edition (32G Turbo): Combines blower fan, vapor chamber, and metal backplate for sustained high-load performance.
- Fanless Edition (32G Fanless): Uses passive cooling for server environments, optimized for chassis airflow.
The MAXSUN Intel Arc Pro B70 series targets AI developers, visualization professionals, and high-density workstation users. Its combination of large VRAM, strong compute capabilities, and flexible deployment options makes it a robust choice for demanding computing environments.
