Why waiting months for EV data is no longer an option for utilities

Almost any conversation with power utility leaders today will surface the exact same challenge: the grid edge is changing much faster than our ability to see it.

The rapid adoption of electric vehicles (EVs), heat pumps, and rooftop solar is a massive win for the energy transition. But at the low-voltage network level, it’s creating a massive visibility gap. Utilities know the demand is growing, but they often have no idea which households are drawing that power, when they are doing it, or how it's impacting local transformers in real time.

Consider the sheer scale of this challenge through the lens of Sacramento Municipal Utility District. They already have 79,000 EVs in their service territory today—a number projected to skyrocket to 288,000 by 2030. Add to that 65,000 solar PV installations delivering 480 MW (heading to 680 MW), plus residential batteries and all-electric homes. It is all proliferating faster than any utility's traditional tools can track. Advanced Metering Infrastructure (AMI) gives utilities high-resolution consumption data, but it doesn't tell you what is actually driving the load.

Historically, solving this meant relying on lagging indicators like DMV registration records, customer self-identification programs, or back-office software that requires three months of historical data just to guess behind-the-meter behavior. On top of this, pulling massive amounts of raw AMI data into the cloud for analysis gets incredibly expensive, incredibly fast.

But what if the answer isn't building a bigger cloud? What if the answer is using the hardware that's already deployed at the household level?

That is exactly the paradigm shift that Sacramento Municipal Utility District (SMUD) just proved in a groundbreaking field test.

How SMUD is redefining grid edge analytics in California

As the nation’s sixth-largest community-owned electric utility, Sacramento Municipal Utility District decided to skip the centralized cloud and run grid edge analytics straight at the source: inside the smart meter.

Partnering with NET2GRID and Itron, SMUD deployed the NET2GRID EnergyAI® Edge application natively onto existing Itron Riva distributed intelligence meters across 840 households. The mission was simple: see if an embedded edge agent could deliver accurate, real-time EV detection without needing historical data training, expensive cloud computing infrastructure, or lagging DMV records.

Spoiler alert: It absolutely can. The results completely rewrite the playbook for what smart meters can do.

87% detection in 7 Days: The Power of Edge Intelligence

Traditional load disaggregation solutions are notoriously slow. They look backward, meaning utilities are always reacting to last month's strain on the grid.

By analyzing high-frequency energy data directly within the meter, the joint NET2GRID and Itron solution delivered a massive leap forward, as demonstrated by the breakthrough California field validation results:

• Unrivaled Speed: The application identified 87% of these EV-charging homes within just seven days of deployment, requiring zero algorithmic training or historical data.

• Massive Accuracy: When cross-referenced against vehicle telematics (the actual ground-truth data directly from the cars), the application achieved a 97% overall EV detection rate within 15 minutes of a charging session being detected.

• Granular Event Details: It didn’t just guess monthly kWh usage. It captured individual charging sessions with 89% Precision and 97% Duration Accuracy, tracking exactly when sessions started and stopped within minutes.

From Raw Visibility to Real-Time Grid Management

During a recent industry webinar on “Unlocking Grid Edge Intelligence: Turning AMI Data into DER Insights”, Nick Tumilowicz, the Director of Distributed Energy Management at Itron asked Trevor Lamb, SMUD’s Manager of IT Grants & Vendor Management, what surprised him the most about the results of this distributed intelligence field trial. Trevor’s response was a striking revelation for the entire energy sector:

"How much we did not know."

Coming from an industry leader like SMUD, the nation's sixth-largest community-owned electric utility that is aggressively targeting a zero-carbon goal by 2030, this admission underscores a universal challenge. It highlights a shared reality across the entire energy sector regarding the deep visibility gaps currently lurking at the low-voltage network level.

As Trevor put it perfectly later in the discussion:

"If you don't have the visibility of the behind-the-meter asset behavior, you can't move to the management."

This fundamental truth is exactly why precise, household-level insight with fine time granularity is such a game-changer. When a local neighborhood transformer is under heavy strain, a traditional back-office report that only provides total consumption data over long time periods is completely useless for real-time operational decisions.

By unlocking instant, household-level awareness, utilities can shift from passive observers to active grid managers. The data generated right at the edge can immediately be used to power:

• Transformer Protection & Congestion Management: Spotting localized grid strain before it causes an outage.

• Targeted Demand Response: Engaging the exact customers who own EVs behind a certain transformer with personalized, real-time Time-of-Use (TOU) notifications.

• Virtual Power Plants (VPPs): Safely orchestrating behind-the-meter resources to balance the grid when demand spikes.

Keeping the analytics inside the meter removes typical backend headaches. Massive cloud computing expenses completely vanish, data storage demands plummet, and consumer privacy is inherently protected because sensitive raw data never has to leave the home.

SMUD’s landmark field validation proves that the next generation of grid edge intelligence won’t be built in a centralized cloud server. It’s already sitting on the side of the customer’s home.

Click on the button below to watch the full webinar and download the full SMUD case study.

Find out more here