BOXX launches APEXX P4i workstation with 192-core AMD EPYC processors
BOXX has launched the APEXX P4i workstation, featuring AMD EPYC 9005-Series processors and up to four NVIDIA RTX PRO GPUs. This system is designed for professional 3D content creation, rendering, and simulation workflows, including those important for effects and animation in streaming production.
Key Takeaways
- Equipped with AMD EPYC 9005-Series processors offering up to 192 cores and 4.8 GHz clock speeds.
- Supports up to four professional NVIDIA RTX PRO 4000 GPUs using Blackwell architecture.
- Features a liquid-cooled chassis designed to accommodate up to 4TB of DDR5-6400MHz ECC memory.
- Includes eight PCIe 5.0 x16 slots to manage high-bandwidth NVMe storage and multi-GPU configurations.
- Available in a versatile mid-tower form factor with an optional rackmount configuration for studio deployment.
Why It Matters
The APEXX P4i addresses the escalating hardware requirements of high-end streaming production, where 8K mastering and AI-assisted effects are becoming standard. By using server-grade EPYC processors in a workstation, BOXX provides the massive core density required for CPU-based rendering while maintaining the PCIe lane richness needed for modern GPU clusters. This hybrid approach allows studios to consolidate heavy production workloads into fewer, more powerful nodes rather than sprawling render farms. For the broader industry, this launch signals a shift toward local high-performance computing as a necessary hedge against fluctuating cloud costs. Watch for increased adoption in virtual production environments where real-time engine stability depends on high-frequency host processors and multi-GPU peer-to-peer communication.
Additional Context
The launch of the APEXX P4i aligns with broader industry pivots toward high-density on-premise hardware to handle generative AI and real-time production workflows. Per Tom’s Hardware in October 2024, the EPYC 9005 'Turin' series provides up to 2.7x the performance of competing server chips in video transcoding, a critical metric for streaming platforms managing large-scale VOD libraries. This generation introduces the Zen 5 architecture, which AMD claims offers a 17% increase in instructions per clock alongside a revamped 512-bit data path for AVX-512, significantly accelerating the floating-point math essential for 3D physics and AI inference. At the same time, the hardware ecosystem is shifting toward more efficient GPU orchestration. AMD highlights that high-frequency models like the 5.0 GHz EPYC 9575F can accelerate GPU-bound AI workloads by up to 28% by ensuring the cards remain saturated with data. This is particularly relevant as NVIDIA’s Blackwell-generation RTX PRO GPUs roll out into the professional market. According to Jon Peddie Research in October 2024, BOXX has targeted a 'first-to-market' strategy with high-bandwidth workstation components, previously being among the first to integrate Intel Core Ultra and Thunderbolt 5 support for motion media professionals. Industry analysts at Deloitte and Unified Streaming noted in early 2026 that high-end production teams are increasingly seeking hybrid infrastructure that combines local workstations with elastic cloud resources. As real-time rendering and virtual production techniques move from niche categories to mainstream broadcast tools, the demand for units capable of handling massive datasets locally — like the 4TB memory capacity of the P4i — has increased. This hardware trend supports the rise of content provenance standards like C2PA, which requires significant local compute power to embed machine-verifiable authenticity signals directly into the video stream during the packaging phase.
Read full article at boxx.com