Artificial intelligence is entering a new stage of maturity. As highlighted at WAIC 2026, the industry is rapidly shifting from technology demonstrations to real-world deployment, where scalable computing, operational reliability, and long-term efficiency have become the foundation of AI success.

Built for this new era, Gooxi G6860T5 AI Server delivers the performance, scalability, and reliability required for AI training, large language model (LLM) inference, digital twins, and other demanding AI workloads.

Built for High-Performance AI Computing

G6860T5 supports up to two 4th or 5th Gen Intel® Xeon® Scalable processors, with up to 350W TDP per processor, providing exceptional computing power for enterprise AI applications.

With 32 DDR5 memory slots supporting speeds of up to 5600 MT/s and PCIe 5.0 connectivity, the platform significantly increases bandwidth while reducing data bottlenecks, enabling smooth performance for large-scale AI inference and concurrent workloads.

Flexible 8-GPU Architecture

Designed for heterogeneous AI environments, G6860T5 supports up to eight double-width, triple-width, 3.5-width, or 4.5-width GPUs, compatible with leading AI accelerator cards from both domestic and international vendors.

Its CPU-to-GPU direct communication architecture minimizes latency, while multiple CPU-GPU topology options—including Common and Performance modes—allow users to optimize system performance for different AI workloads. From LLM inference and model fine-tuning to multimodal AI and digital twin applications, G6860T5 delivers the flexibility modern AI deployments demand.

An intelligent cooling system featuring six 6056 fans and six 12038 fans dynamically adjusts airflow according to GPU utilization, ensuring stable 24/7 operation under heavy workloads.

Scalable Storage and Expansion

To support data-intensive AI applications, G6860T5 accommodates up to twelve 2.5-inch or 3.5-inch front-access drives, supporting SATA, SAS, and NVMe storage, including four NVMe SSD bays for high-speed data access.

Integrated M.2 and Slimline interfaces simplify system storage, while up to 13 PCIe expansion slots provide exceptional flexibility for networking, AI accelerators, RAID controllers, and other expansion devices, making the platform equally suitable for centralized AI clusters and edge deployments.

Enterprise-Class Reliability

Built for continuous operation, G6860T5 supports up to eight hot-swappable power supplies with 1600W, 2200W, or 2700W options and N+N/N+M redundancy to maximize system availability.

Security features including TPM 2.0, administrator password protection, chassis intrusion detection, and redundant BMC design help safeguard hardware and data. Support for Redfish, SNMP, and IPMI 2.0 enables intelligent remote management, monitoring, and automated maintenance, reducing operational complexity across modern data centers.

Enabling the Next Wave of AI Innovation

As AI moves from experimentation to production, infrastructure matters more than ever.

Combining powerful compute performance, flexible GPU expansion, scalable storage, and enterprise-grade reliability, Gooxi G6860T5 provides a robust platform for organizations building next-generation AI infrastructure—helping accelerate AI innovation from the data center to real-world applications.