GPU Servers | KAYTUS

Leading AI computing provider fueling innovations across AI scenarios

Full stack

Most comprehensive AI server portfolio, covering solutions from single node to cluster, full-stack managing software and application optimizing services.
Extreme performance

Leading performance in model training and inference, fueling R&D and application across AI scenarios.
Widely recognized by global customers

Trusted by CSP and leading high-tech companies in AI+Science, AI+Graphics, and AIGC worldwide.
System-level open design

Cooperating with leading AI chip vendors developing mature OAI solutions.

All Products

Main Application Scenarios	Model	Height	Processor	Accelerator Card	Memory	Storage	Cooling Mode	Details
AI inference, deep learning, metaverse, AIGC, AI+Science	KR4268V2	4U	2x 4th Gen Intel® Xeon® Scalable Processors, TDP 350W	Supports 8x Dual-slot FHFL PCIe interface GPU cards, and supports ≥ 4 PCIe 5.0 x16 slots	32x DDR5 DIMMs, up to 4800MT/s	24x 2.5 or 12x 3.5-inch SAS/SATA drive bays in the front, supports up to 16x NVME or E3.S Built-in 2x M.2 NVME/SATA SSD	Air cooling	More
AI inference, deep learning, metaverse, AIGC, AI+Science	KR4268V2	4U	2x AMD EPYCTM 9004 Series Processors, Max cTDP 400W	Up to 10x full-height full-length double-width PCIe interface GPU cards	24x DDR5 DIMMs, up to 4800MT/s	24x 2.5 or 12x 3.5-inch SAS/SATA drive bays in the front, supports up to 16x NVME or E3.S Built-in 2x M.2 NVME SSD	Air cooling	More
AI training, AIGC, metaverse	KR6288V2	6U	2x 4th Gen Intel® Xeon® Scalable Processors, TDP 350W	1x NVIDIA HGX-Hopper-8GPU module, TDP up to 700W per GPU	32x DDR5 DIMMs, up to 4800MT/s	24x 2.5’ SSD, up to 16x NVMe U.2	Air cooling	More
AI training, AIGC, metaverse	KR6288V2	6U	2x AMD EPYCTM 9004 Series Processors, Max cTDP 400W	1x NVIDIA HGX-Hopper-8GPU module, TDP up to 700W per GPU	24x DDR5 DIMMs, up to 4800MT/s	24x 2.5' SSD, up to 16x NVMe U.2	Air cooling	More
Pre-training, fine-tuning	KR6298V2	6U	2x 4th Gen Intel® Xeon® Scalable Processors, TDP 350W	8xIntel Gaudi2/PVC OAMs, TDP up to 600W per OAM	32x DDR5 DIMMs, up to 4800MT/s	24x 2.5' SSD, up to 16x NVMe U.2	Air cooling	More
MotusAI	MotusAI	Fine-grained scheduling of GPU, Data acceleration strategies, Efficient distributed training, Fault tolerance mechanism						More

News & Events

CloudFest 2024

KAYTUS in WAICF24

STAC Summit Fall London

2023 OCP Global Summit

Hot Events

News