After months of intense product and development work, we are very happy to announce our very first dedicated multi-GPU server, the GEX430! The GEX430 combines four NVIDIA RTX PRO™ 6000 Blackwell Max-Q GPUs with 384 GB of GDDR7 ECC VRAM in total and is based on an Intel® Xeon® Gold 5412U with 24 cores and 48 threads. It’s built for LLM inference, fine-tuning, training and running several models in parallel. You can find the details here: https://htznr.li/GEX430_ms