Built by grid operators, for grid operators.
Ampgrove started from a simple observation: the tools distribution utilities use to forecast feeder load haven't meaningfully improved in over a decade, even as grid complexity has accelerated dramatically with EV adoption, distributed generation, and changing load patterns.
Eliminate preventable grid stress events through better feeder-level intelligence.
Distribution utilities over-provision peak capacity because they can't trust their forecasts at the feeder level. The result is billions of dollars in unnecessary capacity reservation costs across the US utility sector annually, while reliability events that a better forecast would have prevented still happen.
Ampgrove's mission is direct: give distribution operations teams the feeder-level forecasting precision and dispatch guidance they need to stop reacting and start managing ahead of the load curve. We focus narrowly on this problem and build nothing that doesn't serve it.
We're based in Denver because Colorado is one of the most active clean-energy markets in the country. The grid complexity problems Ampgrove addresses are showing up in Colorado utilities earlier and more visibly than most other markets. That's where we started.
Three people who've spent careers working on this problem.
IHIngrid Halvorsen
CEO & Co-FounderYears building load forecasting systems for Western grid operations. Ingrid's background is in distribution-level probabilistic forecasting, specifically the problem of predicting feeder load under weather uncertainty. She started Ampgrove after working directly with operations teams and seeing firsthand how much manual burden comes from inadequate feeder-level tooling.
MRMarcus Reinholt
CTO & Co-FounderMachine learning infrastructure specialist with a background in real-time inference pipelines for high-frequency sensor data streams. Marcus architected Ampgrove's data ingestion pipeline and the model serving layer that produces rolling forecasts at 30-minute intervals across multiple feeder networks simultaneously. He built the SCADA connector framework that enables 2-week integrations.
PSPriya Sundaram
Head of Grid IntelligencePower systems operations and dispatch workflow specialist. Priya spent years in distribution operations, translating operator needs into system constraints. She designed Ampgrove's dispatch recommendation engine and alert logic framework to fit how operations teams actually work, not how software engineers imagine they work. She leads all utility customer integrations.
Why we built this now.
The founding insight for Ampgrove came from a consistent pattern across multiple utility operations environments: operators were running complex, high-stakes grids on forecasting tools that were fundamentally designed for an earlier grid era. The 24-hour ahead, substation-level forecast that was adequate for a relatively static load mix is no longer adequate when a single residential feeder can see its peak load profile shift dramatically quarter over quarter as EV adoption accelerates.
Machine learning for time-series prediction has matured enough that building accurate, continuously retrained feeder-level forecasting models is now an engineering problem, not a research problem. What was missing was a product built specifically for distribution utility operations teams: with the right integrations (SCADA/EMS connectors, not generic API calls), the right interface (operator console with threshold-based dispatch recommendations, not a data science dashboard), and the right commercial model (utility procurement norms, not SaaS startup pricing).
Denver is an appropriate place to build this. Colorado utilities are dealing with some of the sharpest EV load profile changes in the country, a strong solar penetration creating variable generation that traditional forecasting models struggle with, and a regulatory environment that's pushing harder on reliability standards than most other states. The grid intelligence problems Ampgrove is designed to solve are present and growing in our home market right now.
We're a small bootstrapped team. We work directly with our utility partners at the operations level. Every feature we've built came out of real conversations with grid operators about what they need, and every number in our results documentation reflects actual pilot data rather than modeled projections. That's our operating principle: narrow focus, precise claims, direct contact.
We're a small team. Talk to us directly.
No sales process layers. If you're evaluating whether Ampgrove fits your grid operations context, email Ingrid directly at [email protected] or use the contact form.