Superpower Daily
Anthropic Adds ant apply to Manage Claude API Resources as CodeGPT-6 Astra Hits 99.9% on ARC-AGI-3, but Scores 62.7% in a Shar...Crusoe Signs Reported $13 Billion, Five-Year AI Cloud Deal With...Microsoft Releases MAI-Transcribe-2 at 10 Cents an HourxAI Opens Grok Bot to Enterprises With Controls for Autonomous...Figure Signs $3.5B Nscale Deal for Up to 100,000 GPUs for Human...Anthropic Adds ant apply to Manage Claude API Resources as CodeGPT-6 Astra Hits 99.9% on ARC-AGI-3, but Scores 62.7% in a Shar...Crusoe Signs Reported $13 Billion, Five-Year AI Cloud Deal With...Microsoft Releases MAI-Transcribe-2 at 10 Cents an HourxAI Opens Grok Bot to Enterprises With Controls for Autonomous...Figure Signs $3.5B Nscale Deal for Up to 100,000 GPUs for Human...

Research observatory 18 / Service interruptions

AI Outage and Reliability Monitor

Published reports of AI outages, downtime, degraded service, availability incidents, and recovery events.

Archived snapshotv11Sep 4, 2026
View latest dataset
Reliability records
23

Published source-backed records

Source references
30

30 distinct URLs shown

Tracked dimensions
4

Service outage / Recovery / Infrastructure incident

Median importance
80

Editorial importance score

Interactive figureReliability
CSV JSON
Data status23 verified records across 2 periods

Each dot is a verified monthly observation of verified records. Lines connect observations in the same series; missing periods are not treated as zero.

Verified observationHover or focus any mark for exact valuesLast updated Sep 4, 2026

Version ledger

Frozen public editions

Each edition preserves the records, method, sources, and downloads available at publication time.

  1. v11 / latestSep 4, 202623 records / 30 sources

    +2 records; +4 sources; updated Reliability records, Source references, Median importance. The frozen snapshot contains 23 records.

  2. v10Sep 3, 202621 records / 26 sources

    +3 records; +6 sources; updated Reliability records, Source references, Median importance. The frozen snapshot contains 21 records.

  3. v9Sep 2, 202618 records / 20 sources

    +4 records; +5 sources; updated Reliability records, Source references, Median importance. The frozen snapshot contains 18 records.

  4. v8Sep 1, 202614 records / 15 sources

    Source-backed fields changed; the frozen snapshot contains 14 records.

  5. v7Aug 31, 202614 records / 15 sources

    +1 records; +2 sources; updated Reliability records, Source references. The frozen snapshot contains 14 records.

  6. v6Aug 28, 202613 records / 13 sources

    +1 records; +1 sources; updated Reliability records, Source references, Median importance. The frozen snapshot contains 13 records.

  7. v5Aug 28, 202612 records / 12 sources

    +1 records; +1 sources; updated Reliability records, Source references. The frozen snapshot contains 12 records.

  8. v4Aug 28, 202611 records / 11 sources

    +1 records; +1 sources; updated Reliability records, Source references, Tracked dimensions. The frozen snapshot contains 11 records.

  9. v3Aug 28, 202610 records / 10 sources

    +1 records; +1 sources; updated Reliability records, Source references. The frozen snapshot contains 10 records.

  10. v2Aug 27, 20269 records / 9 sources

    +1 records; +1 sources; updated Reliability records, Source references. The frozen snapshot contains 9 records.

  11. v1Aug 27, 20268 records / 8 sources

    Initial public snapshot with 8 records and 8 cited sources.

Coverage note

Only publicly reported interruptions in monitored coverage are included. Absence from the tracker is not evidence of uninterrupted service, and duration is not inferred when undisclosed.

Dataset ID
spd:reliability-monitor
Stable URL
/research/reliability-monitor
Version
v11
Coverage
2026-08/2026-09-03
Records
23
Fields
7
Updated

Read the data

The records behind the figure

CSV JSON
AI Outage and Reliability Monitor data records
PublishedRecordDimensionCategoryImportanceSourcesStory URL
2026-08-20Simple AI Opens 2,000 Hours of Robot Training Data Collected Without RobotsService outagemodels700/posts/simple-ai-opens-2-000-hours-of-robot-training-data-collected-without-robots
2026-08-22Stanford Won Databricks’ Agent Cup, but 18.8% of Questions Stumped Every TeamRecoverymodels802/posts/stanford-won-databricks-agent-cup-but-18-8-of-questions-stumped-every-team
2026-08-22AWS Brings GPT-5.6 Cross-Region Inference, With a Data-Location ChoiceInfrastructure incidentinfrastructure801/posts/aws-gives-gpt-5-6-a-global-capacity-route-on-bedrock
2026-08-24EverestLabs Puts AI Agents Near Recycling Controls, but Keeps Operators in ChargeService outageproducts700/posts/everestlabs-puts-ai-agents-near-recycling-controls-but-keeps-operators-in-charge
2026-08-25Guidelight Finds Few Public Rogue-AI Plans as California Requires Safety DisclosuresRecoverypolicy801/posts/guidelight-finds-few-public-rogue-ai-plans-as-california-requires-safety-disclosures
2026-08-25Perplexity’s Portable Computer Runs Agents Locally—but Needs a 24GB Nvidia GPUInfrastructure incidentproducts802/posts/perplexity-s-portable-computer-keeps-ai-agents-local-if-you-have-24gb-of-vram
2026-08-26Meta Adds Voice-Controlled AI and Navigator Home Screen to Quest, With Narrow Initial AccessService outageproducts800/posts/meta-adds-voice-controlled-ai-and-navigator-home-screen-to-quest-with-narrow-initial-access
2026-08-26OpenAI Details Agent Breach of Hugging Face, Halts Research Model and Tightens ControlsRecoveryinfrastructure902/posts/openai-details-agent-breach-of-hugging-face-halts-research-model-and-tightens-controls
2026-08-27OpenAI Tests Persistent Codex Mode That Keeps Working Until Put to SleepRecoverymodels801/posts/openai-tests-persistent-codex-mode-that-keeps-working-until-put-to-sleep
2026-08-28Netflix’s GenRec Pairs a 0.006% Gain With Catalog-Bound LLM RankingService outagemodels701/posts/netflix-s-genrec-edges-production-ranker-by-0-006-relative-online-with-fewer-labels
2026-08-28Meta Tests Data-Center Repair Robots, but Cable Bots Still Need Human SupervisionDegradationinfrastructure701/posts/meta-tests-data-center-repair-robots-but-cable-bots-still-need-human-supervision
2026-08-28Salesforce Uses SageMaker Placement Controls to Meet 2-AZ Agentforce Standard, Keep 8x SavingsInfrastructure incidentinfrastructure701/posts/salesforce-uses-sagemaker-placement-controls-to-meet-2-az-agentforce-standard-keep-8x-savings
2026-08-28SageMaker Feature Store Adds 25-Record Batch Writes and a Record-Discovery APIDegradationinfrastructure701/posts/sagemaker-feature-store-adds-25-record-batch-writes-and-a-record-discovery-api
2026-08-31Pixel Watch 5 Reaches Stores With Offline Gemini Actions and Europe-First Breathing AlertsDegradationproducts902/posts/pixel-watch-5-reaches-stores-with-offline-gemini-actions-and-europe-first-breathing-alerts
2026-09-01ChatGPT Work Tasks Stall as Microsoft Investigates Separate Exchange Online IssuesDegradationproducts501/posts/chatgpt-work-tasks-stall-as-microsoft-investigates-separate-exchange-online-issues
2026-09-01Qwen3.8-Max Leads Open-Weight Commerce Agent Bench by Two Passes, With Audit LimitsRecoverymodels800/posts/qwen3-8-max-leads-open-weight-commerce-agent-bench-by-two-passes-with-audit-limits
2026-09-01Orbis Lets Users Alter Prompts Mid-Stream as Visko Takes Its Video Model PublicService outagemodels903/posts/visko-opens-orbis-with-live-4k-video-backed-by-10m-pre-seed
2026-09-01AWS Publishes Rust Edge Pattern for 256 MB Gateways, Keeping Cloud Agents ConditionalInfrastructure incidentinfrastructure701/posts/aws-publishes-rust-edge-pattern-for-256-mb-gateways-keeping-cloud-agents-conditional
2026-09-02Aranya Raises $11M for Software It Says Can Ready GPU Clusters in 48 HoursInfrastructure incidentstartups802/posts/aranya-raises-11m-to-turn-bare-metal-gpu-servers-into-production-clusters
2026-09-02ASUS Adds RTX Spark to ProArt PCs for Local AI WorkInfrastructure incidentproducts801/posts/asus-adds-rtx-spark-pcs-with-up-to-128gb-of-unified-memory-for-local-ai
2026-09-02Empirik Emerges With $21M to Check Infrastructure Changes Before DeploymentDegradationstartups803/posts/empirik-emerges-with-21m-to-check-infrastructure-changes-before-deployment
2026-09-03OpenAI Applies Mitigation as ChatGPT and Codex Errors Hit Core ServicesRecoveryproducts612/posts/openai-applies-mitigation-after-elevated-errors-hit-chatgpt-and-codex
2026-09-03Four Major AI Services Suffer Rare, Overlapping OutagesRecoveryculture682/posts/four-major-ai-services-suffer-rare-overlapping-outages

Measurement technique

How to read this report

  1. 01Records require an explicit outage, degradation, availability incident, or recovery event.
  2. 02Each published development contributes one observation; repeated reports without a material update are excluded by the editorial pipeline.
  3. 03The monitor does not convert anecdotal user reports into measured downtime.

Sources

Evidence

20 publishers supporting 30 records. Expand a publisher to inspect its cited pages.

aws.amazon.comaws.amazon.com4 sources / 4 records
Introducing cross-Region inference for OpenAI GPT-5.6 models on Amazon Bedrock | Amazon Web Services1 supporting recordSpreading the load: How Salesforce met Multi-AZ HA with SageMaker Inference Components | Amazon Web Services1 supporting recordBatch write and discover records in Amazon SageMaker Feature Store | Amazon Web Services1 supporting recordBuild edge AI agents with the AWS IoT Greengrass Component SDK for Rust | Amazon Web Services1 supporting record
blog.googleblog.google2 sources / 2 records
You can officially buy the Pixel 11 phones and Pixel Watch 5.1 supporting recordPixel Watch 5: Proactive assistance and advanced health tracking on your wrist1 supporting record
CNBCcnbc.com2 sources / 2 records
OpenAI releases sweeping report on Hugging Face AI agent hack1 supporting recordMicrosoft Outlook and OpenAI's ChatGPT Work are experiencing user outages1 supporting record
databricks.comdatabricks.com2 sources / 2 records
Evaluating AI Agents Live at the Grounded Reasoning Cup1 supporting recordIntroducing OfficeQA Pro V2: A New Benchmark for Enterprise Grounded-Reasoning1 supporting record
finance.yahoo.comfinance.yahoo.com2 sources / 2 records
AI startup Visko closes $10 million pre-seed round and launches Orbis, its first Live Model1 supporting recordempirik.ai emerges from stealth with $21 Million to build the AI Agent for Infrastructure Change1 supporting record
status.openai.comstatus.openai.com2 sources / 2 records
ChatGPT Work Mode High Error Rates1 supporting recordElevated errors across ChatGPT and Codex - OpenAI Status1 supporting record
Show 14 more publishers
TechCrunchtechcrunch.com2 sources / 2 records
Frontier AI labs still won't say how they'd contain a rogue model | TechCrunch1 supporting recordSequoia-incubated Empirik launches with $21M to predict outages before they happen | TechCrunch1 supporting record
WIREDwired.com2 sources / 2 records
OpenAI Is Developing a ‘Persistent’ AI Agent1 supporting recordInside Meta’s Push to Put Robots to Work in Data Centers1 supporting record
Ars Technicaarstechnica.com1 source / 1 record
Four major AI models suffer rare overlapping downtime1 supporting record
arxiv.orgarxiv.org1 source / 1 record
GenRec: An LLM-Backed Recommendation Ranker at Netflix1 supporting record
bastillepost.combastillepost.com1 source / 1 record
AI startup Visko closes $10 million pre-seed round and launches Orbis, its first Live Model1 supporting record
firstpost.comfirstpost.com1 source / 1 record
Perplexity Launches AI Tool That Runs Locally And Seeks Permission For Cloud Help1 supporting record
futurism.comfuturism.com1 source / 1 record
World Plunged Into Chaos as ChatGPT, Claude, and Grok Suddenly Go Down Simultaneously: "Finally I Can See the Sun!"1 supporting record
openai.comopenai.com1 source / 1 record
The Hugging Face incident and the road ahead1 supporting record
press.asus.compress.asus.com1 source / 1 record
press.asus.com1 supporting record
prnewswire.comprnewswire.com1 source / 1 record
prnewswire.com1 supporting record
sequoiacap.comsequoiacap.com1 source / 1 record
sequoiacap.com1 supporting record
siliconangle.comsiliconangle.com1 source / 1 record
Aranya raises $11M to turn bare-metal servers into AI clusters in less than 48 hours - SiliconANGLE1 supporting record
venturebeat.comventurebeat.com1 source / 1 record
Perplexity partners with Nvidia to launch Portable Computer, a fully local AI agent with zero token costs1 supporting record
visko.aivisko.ai1 source / 1 record
visko.ai1 supporting record
Next report / 19AI Benchmark Saturation Tracker All research reports