Supermicro's new AI platforms speed up enterprise AI deployments
Supermicro just rolled out seven AI data platforms built with NVIDIA tech, aiming to make enterprise AI setups much faster and easier.
These systems pack the latest RTX PRO 6000 and 4500 Blackwell GPUs, speedy Spectrum-X Ethernet networking, and tools like NIM microservices and NeMo for smoother training and inference.
They're made to handle huge datasets and work seamlessly with storage partners like IBM, Cloudian, VAST Data, Nutanix, DDN, Everpure, and WEKA.
One standout is the CNode-X solution (a joint Supermicro and VAST Data offering): think two NVIDIA RTX PRO 6000 GPUs; the platform's EBox component is updated to AMD EPYC 9005 CPUs.
The new systems were showcased at NVIDIA's GTC event
Unveiled at this year's GTC event (March 16-19), the new platforms showed off use case demos and a mobile AI Factory.
Supermicro CEO Charles Liang said these solutions mean faster deployment thanks to modular building blocks and advanced cooling.
NVIDIA's Jason Hardy called out how unified compute, networking, and storage help boost AI performance,
IBM's Sam Werner noted that quicker data access means fewer "hallucinations" in AI results,
while Cloudian's Amit Rawlani highlighted direct GPU data handling as a game-changer for speed.