replicate · 82d ago
Hitting GPU Capacity for H100s creating large queue times for some models
ResolvedYes — replicate had a problem on July 22, 2026, and replicate has since marked it resolved. replicate's latest word: “This issue is resolved”
Timeline
Started: 7/22/2026, 11:51:02 AM
Last update: 7/23/2026, 2:19:58 PM
Resolved: 7/23/2026, 2:19:58 PM
Updates
resolved
This issue is resolved
7/23/2026, 2:19:58 PM
monitoring
This issue is now resolved
7/23/2026, 3:05:49 AM
monitoring
GPU usage is back below capacity. We're continuing to monitor.
7/22/2026, 5:11:16 PM
identified
We are over provisioned on H100s currently which is causing long queue times for some models running on H100s
7/22/2026, 11:51:02 AM
More replicate incidents
- A100 hardware partial outage — resolved, 48d ago
- Delayed scaling due to node failure — resolved, 64d ago
- Significant degradation — resolved, 69d ago
- API degraded for A100s — resolved, 72d ago
- Degraded scale-out due to failed setups pulling from huggingface — resolved, 74d ago
- HuggingFace download issues — resolved, 89d ago