CoreWeave Compute
PlatformBare-metal and virtualized GPU and CPU infrastructure optimized for AI training, inference, visual computing, and HPC.
- Dec 1, 2019
- —
- —
- —
- —
- 5
Products
| Name | Type | Category | Launched | Latest release | Status |
|---|---|---|---|---|---|
| NVIDIA Vera CPU Instances | Product | — | Sep 2026 | CoreWeave announces NVIDIA Vera CPU offering | announced |
| NVIDIA Vera Rubin NVL72 | Product | — | Sep 2026 | CoreWeave delivers Vera Rubin NVL72 production performance with Cognition | active |
| NVIDIA GB300 NVL72 Instances | Product | — | Oct 2025 | — | active |
| NVIDIA HGX B200 Instances | Product | — | May 2025 | NVIDIA HGX B200 instances become generally available | active |
| NVIDIA GB200 NVL72 Instances | Product | — | Feb 2025 | NVIDIA GB200 NVL72 instances become generally available | active |
| NVIDIA H200 Instances | Product | — | Aug 2024 | CoreWeave becomes first cloud provider to deploy NVIDIA H200 GPUs | active |
| NVIDIA GH200 Instances | Product | — | Jan 2024 | CoreWeave announces NVIDIA GH200 and HGX H200 instances | active |
| CoreWeave Tensorizer | Product | — | Jun 2023 | CoreWeave introduces Tensorizer for faster model loading | active |
| NVIDIA H100 Instances | Product | — | Mar 2023 | CoreWeave deploys NVIDIA H100 GPU clusters | active |
| CoreWeave Virtual Servers | Product | — | Jan 2020 | — | active |
| GPU Instances | Product | — | Dec 2019 | — | active |
Releases
Jun 22, 2026CoreWeave sets MLPerf Training v6.0 recordsBenchmark result
CoreWeave reported record MLPerf Training v6.0 results, including training DeepSeek-V3 in approximately two minutes on a large NVIDIA GB300 NVL72 cluster.
Mar 10, 2026CoreWeave introduces flexible capacity plansPricing change
CoreWeave introduced more flexible cloud-capacity plans designed to match customer workload duration and deployment needs.
Jun 4, 2025CoreWeave, NVIDIA, and IBM set Blackwell MLPerf training recordsBenchmark result
The companies submitted the largest NVIDIA Blackwell MLPerf Training v5.0 cluster and reported record Llama 3.1 405B training performance.
Apr 2, 2025CoreWeave posts first cloud NVIDIA GB200 MLPerf inference resultsBenchmark result
CoreWeave submitted the first cloud-provider MLPerf Inference v5.0 results for NVIDIA GB200, reporting more than 800 tokens per second on Llama 3.1 405B.
Jul 18, 2024CoreWeave announces large-scale NVIDIA Blackwell clustersAnnouncedProduct launch
CoreWeave announced NVIDIA HGX B200 and GB200 NVL72 Blackwell cluster configurations connected with NVIDIA Quantum-2 InfiniBand.