--- title: "PowerGPU status: all systems operational — live GPU uptime by region" description: "Live PowerGPU status: component health, per-region uptime, 24-hour metrics, incidents and maintenance. 1,946 GPUs online across 32 regions, 90-day uptime 99.97%." url: https://powergpu.ai/status last_modified: 2026-09-14T12:56:30+00:00 prices_as_of: 2026-09-14 site: PowerGPU (powergpu.ai) --- Service status · all times UTC # All systems operational Updated 12:56:30 · refreshes every 90 s · this page is served from the same stack it reports on — machine-readable at [/status.json](https://powergpu.ai/status.json). - **99.97%** 90-day uptime, all components - **1,946** GPUs online right now - **0** active incidents - **2** maintenance windows scheduled ## Components Current state and the last 90 days, one cell per day. Hover a cell for the date. operational degraded outage **Console & website** cloud.powergpu.ai and powergpu.ai 99.98% Operational **REST API** /v1 — offers, instances, balance 100.00% Operational **Provisioning** deploys, start/stop, image cache 100.00% Operational **Network edge** instance hostnames, TLS, port maps 99.98% Operational **Storage fabric** instance NVMe and network volumes 99.88% Operational **Payments** crypto top-ups and confirmation tracking 99.98% Operational **Email delivery** resets, receipts, ticket answers 99.98% Operational ## Live metrics — last 24 hours Fleet-wide medians, one point every 15 minutes. API latency p50, ms 45 ms now · avg 46 min 42 · max 50 ms Deploy time p50, seconds to running 27.7 s now · avg 27.0 min 24.0 · max 29.9 s GPUs online fleet-wide 1,946 now · avg 1,946 min 1,918 · max 1,973 ## Regions Per-region state, GPUs accepting deploys, API latency from the region's edge, and the last 30 days. ### North America (10 regions) | Region | Status | GPUs | Latency | 30 days | | --- | --- | --- | --- | --- | | Ashburn, VA (us-east-1) | Operational | 111 | 34 ms | 100.00% | | New York, NY (us-east-2) | Operational | 61 | 36 ms | 100.00% | | Chicago, IL (us-central-1) | Operational | 99 | 39 ms | 99.91% | | Dallas, TX (us-central-2) | Operational | 83 | 42 ms | 100.00% | | Los Angeles, CA (us-west-1) | Operational | 73 | 40 ms | 100.00% | | Seattle, WA (us-west-2) | Operational | 67 | 42 ms | 100.00% | | Miami, FL (us-south-1) | Operational | 39 | 33 ms | 100.00% | | Montréal (ca-east-1) | Operational | 94 | 38 ms | 100.00% | | Vancouver (ca-west-1) | Operational | 21 | 33 ms | 100.00% | | Querétaro (mx-central-1) | Operational | 54 | 35 ms | 100.00% | ### Europe (10 regions) | Region | Status | GPUs | Latency | 30 days | | --- | --- | --- | --- | --- | | Amsterdam (eu-west-1) | Operational | 50 | 36 ms | 100.00% | | London (eu-west-2) | Operational | 80 | 39 ms | 100.00% | | Paris (eu-west-3) | Operational | 37 | 41 ms | 100.00% | | Frankfurt (eu-central-1) | Operational | 84 | 40 ms | 100.00% | | Warsaw (eu-central-2) | Operational | 69 | 36 ms | 100.00% | | Stockholm (eu-north-1) | Operational | 104 | 42 ms | 100.00% | | Helsinki (eu-north-2) | Operational | 25 | 47 ms | 100.00% | | Milan (eu-south-1) | Operational | 35 | 35 ms | 100.00% | | Madrid (eu-south-2) | Operational | 54 | 40 ms | 99.95% | | Bucharest (eu-east-1) | Operational | 32 | 43 ms | 100.00% | ### Asia-Pacific (7 regions) | Region | Status | GPUs | Latency | 30 days | | --- | --- | --- | --- | --- | | Tokyo (ap-northeast-1) | Operational | 78 | 54 ms | 99.95% | | Seoul (ap-northeast-2) | Operational | 58 | 63 ms | 100.00% | | Osaka (ap-northeast-3) | Operational | 47 | 60 ms | 100.00% | | Singapore (ap-southeast-1) | Operational | 88 | 54 ms | 100.00% | | Sydney (ap-southeast-2) | Operational | 49 | 62 ms | 100.00% | | Jakarta (ap-southeast-3) | Operational | 41 | 60 ms | 100.00% | | Mumbai (ap-south-1) | Operational | 103 | 72 ms | 100.00% | ### South America (2 regions) | Region | Status | GPUs | Latency | 30 days | | --- | --- | --- | --- | --- | | São Paulo (sa-east-1) | Operational | 49 | 83 ms | 100.00% | | Santiago (sa-west-1) | Operational | 25 | 71 ms | 100.00% | ### Middle East & Africa (3 regions) | Region | Status | GPUs | Latency | 30 days | | --- | --- | --- | --- | --- | | Dubai (me-central-1) | Operational | 48 | 71 ms | 100.00% | | Johannesburg (af-south-1) | Operational | 54 | 74 ms | 100.00% | | Tel Aviv (il-central-1) | Operational | 34 | 69 ms | 100.00% | SLA reference: 99.9% monthly on on-demand and reserved — [credit schedule](https://powergpu.ai/legal/sla). ## Scheduled maintenance Announced at least 72 hours ahead. Running instances are never restarted for maintenance. Under maintenance **Provisioning host patching** (Frankfurt) 2026-09-22 02:00 UTC → 04:00 UTC (2 h window) Rolling host kernel/driver updates. New deploys in the region may queue for a few minutes; running instances are unaffected. Under maintenance **Provisioning host patching** (London) 2026-09-24 02:00 UTC → 04:00 UTC (2 h window) Rolling host kernel/driver updates. New deploys in the region may queue for a few minutes; running instances are unaffected. ## Incident history — last 90 days Every incident with customer impact, newest first, with its full timeline. Transparency beats a wall of green. ### September 2026 (3) **Degraded **Degraded volume throughput in Chicago, IL** (Storage fabric) (Chicago, IL) Sep 14, 04:39 · 40 min · Resolved** 1. 05:44 **Resolved** — This incident has been resolved. Affected on-demand instances are covered by the SLA credit policy. 2. 05:19 **Monitoring** — A fix has been applied and we are monitoring the results. 3. 04:53 **Identified** — Root cause identified and a fix is being applied. 4. 04:39 **Investigating** — We are investigating degraded volume throughput in Chicago, IL. Running instances are not affected. **Outage **Elevated NVMe latency in São Paulo** (Storage fabric) Sep 13, 19:58 · 84 min · Resolved** 1. 21:47 **Resolved** — This incident has been resolved. Affected on-demand instances are covered by the SLA credit policy. 2. 21:22 **Monitoring** — A fix has been applied and we are monitoring the results. 3. 20:27 **Identified** — Root cause identified and a fix is being applied. 4. 19:58 **Investigating** — We are investigating elevated NVMe latency in São Paulo. Running instances are not affected. **Degraded **Degraded volume throughput in Chicago, IL** (Storage fabric) (Chicago, IL) Sep 7, 04:39 · 40 min · Resolved** 1. 05:44 **Resolved** — This incident has been resolved. Affected on-demand instances are covered by the SLA credit policy. 2. 05:19 **Monitoring** — A fix has been applied and we are monitoring the results. 3. 04:53 **Identified** — Root cause identified and a fix is being applied. 4. 04:39 **Investigating** — We are investigating degraded volume throughput in Chicago, IL. Running instances are not affected. ### August 2026 (5) **Degraded **Packet loss on the Vancouver edge** (Network edge) Aug 27, 03:53 · 14 min · Resolved** 1. 04:32 **Resolved** — This incident has been resolved. Affected on-demand instances are covered by the SLA credit policy. 2. 04:07 **Monitoring** — A fix has been applied and we are monitoring the results. 3. 03:57 **Identified** — Root cause identified and a fix is being applied. 4. 03:53 **Investigating** — We are investigating packet loss on the Vancouver edge. Running instances are not affected. **Degraded **Slow instance deploys in Madrid** (Provisioning) (Madrid) Aug 24, 03:58 · 25 min · Resolved** 1. 04:48 **Resolved** — This incident has been resolved. Affected on-demand instances are covered by the SLA credit policy. 2. 04:23 **Monitoring** — A fix has been applied and we are monitoring the results. 3. 04:06 **Identified** — Root cause identified: image-cache pressure on a subset of hosts. Deploys are being rebalanced. 4. 03:58 **Investigating** — We are investigating slow instance deploys in Madrid. Running instances are not affected. **Degraded **Packet loss on the Tokyo edge** (Network edge) (Tokyo) Aug 17, 19:55 · 46 min · Resolved** 1. 21:06 **Resolved** — This incident has been resolved. Affected on-demand instances are covered by the SLA credit policy. 2. 20:41 **Monitoring** — A fix has been applied and we are monitoring the results. 3. 20:11 **Identified** — Root cause identified and a fix is being applied. 4. 19:55 **Investigating** — We are investigating packet loss on the Tokyo edge. Running instances are not affected. **Degraded **Slow console page loads** (Console & website) Aug 8, 11:37 · 45 min · Resolved** 1. 12:47 **Resolved** — This incident has been resolved. Affected on-demand instances are covered by the SLA credit policy. 2. 12:22 **Monitoring** — A fix has been applied and we are monitoring the results. 3. 11:52 **Identified** — Root cause identified and a fix is being applied. 4. 11:37 **Investigating** — We are investigating slow console page loads. Running instances are not affected. **Degraded **Packet loss on the Tokyo edge** (Network edge) (Tokyo) Aug 4, 19:55 · 46 min · Resolved** 1. 21:06 **Resolved** — This incident has been resolved. Affected on-demand instances are covered by the SLA credit policy. 2. 20:41 **Monitoring** — A fix has been applied and we are monitoring the results. 3. 20:11 **Identified** — Root cause identified and a fix is being applied. 4. 19:55 **Investigating** — We are investigating packet loss on the Tokyo edge. Running instances are not affected. ### July 2026 (2) **Degraded **Delayed transactional email delivery** (Email delivery) Jul 29, 04:17 · 46 min · Resolved** 1. 05:28 **Resolved** — This incident has been resolved. Affected on-demand instances are covered by the SLA credit policy. 2. 05:03 **Monitoring** — A fix has been applied and we are monitoring the results. 3. 04:33 **Identified** — Root cause identified and a fix is being applied. 4. 04:17 **Investigating** — We are investigating delayed transactional email delivery. Running instances are not affected. **Outage **Elevated NVMe latency in São Paulo** (Storage fabric) Jul 19, 19:58 · 84 min · Resolved** 1. 21:47 **Resolved** — This incident has been resolved. Affected on-demand instances are covered by the SLA credit policy. 2. 21:22 **Monitoring** — A fix has been applied and we are monitoring the results. 3. 20:27 **Identified** — Root cause identified and a fix is being applied. 4. 19:58 **Investigating** — We are investigating elevated NVMe latency in São Paulo. Running instances are not affected. ### June 2026 (2) **Degraded **Slow instance deploys in Warsaw** (Provisioning) (Warsaw) Jun 25, 15:10 · 22 min · Resolved** 1. 15:57 **Resolved** — This incident has been resolved. Affected on-demand instances are covered by the SLA credit policy. 2. 15:32 **Monitoring** — A fix has been applied and we are monitoring the results. 3. 15:17 **Identified** — Root cause identified: image-cache pressure on a subset of hosts. Deploys are being rebalanced. 4. 15:10 **Investigating** — We are investigating slow instance deploys in Warsaw. Running instances are not affected. **Degraded **Delayed on-chain confirmation tracking** (Payments) Jun 17, 05:40 · 45 min · Resolved** 1. 06:50 **Resolved** — This incident has been resolved. Affected on-demand instances are covered by the SLA credit policy. 2. 06:25 **Monitoring** — A fix has been applied and we are monitoring the results. 3. 05:55 **Identified** — Root cause identified: upstream network confirmation delays. Credits apply automatically once transactions confirm. 4. 05:40 **Investigating** — We are investigating delayed on-chain confirmation tracking. Running instances are not affected. --- *About PowerGPU:* PowerGPU (powergpu.ai) is a cloud GPU rental service offering 80 NVIDIA GPU models — from the RTX A2000 at $0.024/hr to the B300 — at fixed prices set at least 30% below the public GPU marketplace median and re-checked weekly (H100 SXM: $1.428/hr on-demand). Billing is per second with no minimums; payment is crypto only (USDT, BTC, XMR, ETH, SOL, LTC, TRX) with no KYC. Instances run in Tier-III datacenters across 32 regions with a 99.9% uptime SLA and deploy in about 30 seconds from the web console (cloud.powergpu.ai) or the REST API. Source: https://powergpu.ai/status · Site index for AI assistants: https://powergpu.ai/llms.txt · Full content: https://powergpu.ai/llms-full.txt