
Visual Inspection at Scale: Pooling GPU Resources Across Factories
A manufacturing case study on defect detection, throughput, and cost control with TensorFusion.
Customer Profile
A multi-site manufacturer running automated visual inspection across 9 factories. Workloads spiked during shift changes and major product launches.
The Business Problem
- Edge GPU resources underused outside of shift peaks.
- Throughput bottlenecks when multiple lines launched new SKUs.
- Training and inference competed for the same GPU resources.
Baseline:
| Metric | Baseline |
|---|---|
| Defect detection throughput | 220–260 items/min |
| GPU utilization | 25–33% |
| Model refresh cycle | 10 weeks |
| Quality escape rate | 0.9–1.1% |
TensorFusion Solution
- Edge-first inference with pooled GPU resources across factories.
- Burst training pool that activates only during model retraining windows.
- Policy-based GPU slicing to prioritize production lines.
Results
| Metric | Before | After |
|---|---|---|
| Defect detection throughput | 240 items/min | 420 items/min |
| GPU utilization | 30% | 72% |
| Model refresh cycle | 10 weeks | 6 weeks |
| Quality escape rate | 1.0% | 0.4% |
“We stopped buying GPUs for peak launches only. The pooled model paid for itself in two quarters.” — Manufacturing Systems Director
Why It Works in Manufacturing
Factories have predictable shift peaks. TensorFusion aligns compute to those peaks without overbuying GPU resources.
Author

Categories
More Posts

AI Infra Partners: Building a Federated Compute Network with SLA Control
A customer story on federating GPU supply across clusters while keeping SLAs, data locality, and operations sane.


How TenClass saved 80% on GPU costs with TensorFusion?
TenClass using TensorFusion to save 80% on GPU costs


Public Safety Video Analytics at City Scale with Elastic GPU Resources
A public safety case study using pooled GPU resources to reduce response latency and improve utilization across city-wide video systems.

Newsletter
Join the community
Subscribe to our newsletter for the latest news and updates