Training goodput: stipulated 90%-versus-96% scenario endpoints
90% vs 96% scenariomodeled
| Value kind | modeled — Interpret this value according to its displayed kind, scope, source, and as-of date. |
|---|---|
| Scope | Stipulated endpoints for sensitivity only. They are neither measured provider outcomes nor universal targets. Chapter 14.1 reconciles productive-time boundaries; a site measures its own baseline. |
| As of | 2026-09-08 |
| Source | Guide analysis — stipulated sensitivity scenario; no claim of an industry measurement. · Chapter 14.1: stipulated sensitivity endpoints, not sampled provider results |
| Derivation | Stipulated 90% and 96% sensitivity endpoints; the named site selects and measures its own operating baseline. |
| Review | checking…review by 2026-12-08 · standard cadence |
| Recorded changes | last 2026-09-16 · 31 revisions tracked |
| Claim id | goodput-effective-training-time-industry |
Where the guide uses it
- 0.2 How to Read This Guide: Decisions, Consequences & Reference Data
- 0.3 Vocabulary, Mental Models & the Metric Stack
- 0.5 Reliability, Redundancy & Availability: The Design-Basis Primer
- 1.2 Training Data Centers: Synchronous, Dense, Checkpointable
- 1.7 The Requirements-and-Consequences Matrix
- 2.7 Simulation-Driven Design & the Digital Twin as a Design-Validation Tool
- 7.9 Software Ecosystems & Lock-In
- 7.14 Server & System Integration
- 8.1 Network Fundamentals & AI Traffic Characterization
- 8.4 Scale-Out Fabric: Protocols, Standards & Transport
- 9.4 Checkpointing for Large-Scale Training
- 10.1 Orchestration Architecture & the Scheduling Plane
- 10.2 Topology-Aware & Rack-Scale Scheduling
- 10.6 Observability, Telemetry & GPU Health
- 10.7 Fleet Reliability, Fault Tolerance & Autonomous Recovery
- 10.8 MLOps & Training Frameworks
- 10.9 Customer Onboarding, Delivery & Productization
- 10.11 Inference Serving Engineering: SLOs, Batching, Disaggregation & Goodput-Optimal Scheduling
- 12.2 The AI-Cluster Reliability Rethink: Goodput vs Facility Availability
- 12.4 SLAs, Goodput Contracts & Availability Commitments
- 12.5 Quantitative Reliability & Availability Modeling (RBD / FTA / Monte-Carlo)
- 13.2 Documentation, Scripts & Acceptance Test Plans
- 13.9 Cluster-Scale Benchmarking, Reference Training & Storage/Scheduler Validation
- 13.10 Staged Power/Load Ramp, Go-Live & Handover to Operations
- 14.1 Operational KPIs, Goodput & the Reliability Economics of AI Factories
- 14.2 DCIM, Telemetry & Observability for GPU-Dense, Liquid-Cooled Facilities
- 14.4 Reliability Engineering for Training (Operational)
- 14.5 Predictive & Preventive Maintenance of Power and Cooling Plant
- 14.8 Firmware & Software Lifecycle Management at Fleet Scale
- 14.14 Continuous & Re-Commissioning on a Live Campus
- 16.3 Software, Orchestration & Efficiency at the Frontier
← Full numbers register — every date-stamped figure in the guide, with revision history.