EdgeRun AI Raises $3.2M Seed Round to Expand Edge Predictive Maintenance Platform
The Stamford-based startup closed its seed round to accelerate hardware development and expand pilot customer deployments in automotive and process manufacturing.
Technical articles, case studies, and updates from the EdgeRun team on predictive maintenance and industrial edge computing.
The Stamford-based startup closed its seed round to accelerate hardware development and expand pilot customer deployments in automotive and process manufacturing.
Threshold-based alarms are set high enough to avoid false positives — which means they also miss the outer race defect frequency until it is too late to plan a repair.
Academic RUL models are trained on run-to-failure data from test rigs. Real factory conditions — variable load, contamination, temperature cycling — break those assumptions.
The Maximo REST API is well-documented. The undocumented part is which asset attributes your maintenance planner actually reads before deciding to dispatch a technician.
Cloud inference costs $0.003 per 1,000 predictions. But your OT network has a 200ms round-trip to the cloud and the plant security team blocks outbound TCP on port 443.
Operators spend eight hours a shift staring at a Siemens WinCC HMI. The last thing they need is another monitoring interface. Here is how we add health tags to existing SCADA displays.
Three false alarms in a row and your maintenance supervisor starts ignoring the system. We have seen this happen twice in our pilots, and it is always the same root cause.
Zone 1 classification means every piece of hardware needs IECEx or ATEX approval. Most IIoT vendors stop at Zone 2. Here is what that means for chemical plant maintenance teams.
Every vendor claims predictive maintenance pays for itself in year one. Here is a calculation methodology that survives CFO scrutiny and does not rely on inflated downtime cost assumptions.
MCSA is often presented as a non-intrusive alternative to vibration sensing. For bearing faults before stage 3, the current signal is too clean to be useful.
Pump failures do not happen suddenly. The bearing outer race defect frequency shows up in the spectrum 3 to 6 weeks before seizure. Here is what to look for at each degradation stage.
A predictive maintenance system is only as good as its normal operating signature. Rush the baseline period and you will spend three months tuning false positives instead of preventing failures.