BOXX revamps AI production line with compact APEXX W3 deep learning workstation
BOXX Technologies introduced its new APEXX W3 deep learning workstation at NVIDIA's GTC conference, featuring Intel Xeon W processors and NVIDIA GPUs. The company also launched the GX8-P deep learning server, APEXX P5 workstation, and GoBOXX SLM VR mobile workstation, all designed to accelerate high-performance computing applications, particularly deep learning and GPU-accelerated rendering.
Key Takeaways
- APEXX W3 supports four dual-slot NVIDIA GPUs and Intel Xeon W CPUs with up to 18 cores in a compact chassis.
- The new GX8-P deep learning server features a single AMD EPYC processor with up to 32 cores and support for eight full-size GPUs.
- BOXX introduced the GoBOXX SLM VR, marketed as the world’s thinnest and lightest VR-ready mobile workstation.
- APEXX P5 workstation combines an AMD EPYC processor (up to 32 cores) and support for four dual-slot NVIDIA graphics cards.
- Systems are optimized for V-Ray rendering using NVIDIA OptiX AI-accelerated denoiser technology.
Why It Matters
The shift toward local, deskside AI development hardware indicates a growing need for engineers to bypass cloud latency and costs during the iterative training of video-centric neural networks. For the streaming ecosystem, this provides a more accessible entry point for small-to-medium studios to develop bespoke AI-driven compression, restoration, and content recommendation models. The combination of high core counts and multi-GPU support suggests that the line between production workstations and enterprise servers is continuing to blur. Watch for similar small-form-factor releases using Intel’s newly branded 'Xeon 600' series to determine if they further erode the traditional rack-mount server market for rendering.
Additional Context
The release of the APEXX W3 occurs as the high-performance workstation market undergoes a significant architectural shift. Per TechPowerUp (February 2026), Intel recently introduced the Xeon 600 Series processors, codenamed Granite Rapids-WS, which are set to replace the Sapphire Rapids-based Xeon W-2500 and W-3500 series found in earlier high-end configurations. These newer chips offer up to 86 cores and 128 PCIe 5.0 lanes, representing a 61% improvement in multi-thread performance over previous generations. This transition is critical for video professionals who require massive bandwidth for local AI agent execution and 12K video editing.
Simultaneously, NVIDIA is expanding its 'agentic AI' push. Per NVIDIA (June 2026), the company recently unveiled the RTX Spark superchip, which pairs a Blackwell-based GPU with a 20-core Grace CPU to deliver up to 1 petaflop of local AI processing power. This move toward 'on-box' AI processing is driving a refresh cycle across the hardware industry, with major players like Dell and Lenovo also shipping systems optimized for local 120-billion-parameter model execution.
Industry data from Expert Market Research (February 2026) projects the global high-performance computing market to reach $102.6 billion by 2035, fueled by the demand for local AI compute nodes that can handle high-frequency data and complex simulations with minimal latency. For production houses, the focus has shifted from raw clock speeds to 'tokens per watt,' as hardware manufacturers prioritize the energy efficiency of local inference engines used in virtual production and real-time rendering environments.
Read full article at boxx.com
Enjoy our coverage?
Add StreamingMeme as a preferred source on Google to see more of our streaming news at the top of your Search results.
Add as preferred source