
Craig Rutfield has a startup mentality. Move fast, break things, improve. It’s a mindset often at odds with historical electric utility perceptions, which bemoan the industry’s slow embrace of innovation.
Rutfield spent two decades guiding software startups through the mazes of product design, iteration, and execution. Now, as the head of engineering for National Grid’s smart meter program, his scrappiness has helped the utility leapfrog peers in the grid edge intelligence race.
National Grid has so far deployed 400,000 next-generation smart meters — coined AMI 2.0 for the latest version of advanced metering infrastructure — in New York State. Over the next couple of years, the utility will install a total of 1.7 million electric meters and 640,000 gas modules in Upstate New York and 1.1 million electric meters in Massachusetts.
Like a startup, there’s no clear-cut playbook for National Grid. Most major utilities recognize the inevitable transition to AMI 2.0 but are still developing a deployment strategy.
“There are always challenges, but that’s part of the fun,” Rutfield said. “We’re also a first mover — National Grid isn’t always a first mover — which also makes it exciting to me.”
There are plenty of reasons to be excited.
AMI 2.0 makes AMI 1.0 look like a technological artifact. While AMI 1.0 meters gave utilities periodic data for energy management and billing, their primary function, they lacked the functionality to keep pace with the energy transition.
The latest AMI technology enables new capabilities including remote connect/disconnect, real-time data collection, load disaggregation, and customer opt-in WiFi connectivity— a significant advancement that allows for frictionless energy management for customers and distributed energy resource control. Each meter is capable of hosting a suite of applications, as well, much like your favorite cellphone company’s app store. This is where grid edge intelligence comes into play: AMI 2.0 meters will support computing at the point of consumption, opening up endless customer programs and rate structures, as well as improved power quality and resilience.
It took National Grid regulators more than a decade to buy into the AMI hype, though. Following an AMI 1.0 pilot with Itron involving around 20,000 smart meters from 2009 to 2012, the utility repeatedly struggled to secure approval for a full-scale deployment. Regulators contested that National Grid’s existing fleet of AMR meters had potentially decades of useful life left, and the utility couldn’t justify the investment.
That all changed in 2020, as regional decarbonization goals required higher-quality data and infrastructure to support economy-wide electrification. That year, New York approved National Grid’s AMI plan, and Massachusetts followed in late 2022. But by that time, AMI 1.0 was already giving way to AMI 2.0, so National Grid was able to skip straight to the latest and greatest metering tech.
Rutfield said the tone around smart meters quickly shifted “from ‘no, no, no’ to ‘how quickly can you get it done?”
How National Grid got it done is why utilities across the country are watching closely.
Deploy, deploy, deploy
National Grid’s early success with AMI 2.0 can be attributed to what Rutfield calls an agile deployment approach.
Instead of ironing out the details for every use case, National Grid got the smart meters — Landis+Gyr’s Revelo grid sensing platform — into the field as quickly as possible. They started with six customers, who were also National Grid employees, before scaling to 10, 15, and 50 installations per day. The utility now installs around 15,000 new meters in New York State each week. It’s deployment in Massachusetts is about a year behind due to its regulatory timeline but will take full advantage of the learnings in New York, installing around 50,000 meters per month.
Sean Cromie, Landis+Gyr’s senior vice president of the Americas, said that while grid edge intelligence has existed for more than two decades, next generation metering technology required a “connected ecosystem” to go beyond simple meter-to-cash and provide real-time, actionable insights. He added that National Grid “challenged manufacturers to provide more than a meter.”
National Grid’s new meters were first only used for meter-to-cash. The utility prioritized rapid deployment over full functionality, which can be layered on over time thanks to Landis+Gyr’s separation of meter-to-cash from its edge app ecosystem. “Starting small,” Rutfield said, allowed National Grid to provide value to customers sooner and learn lessons about the new technology along the way to minimize the impact of potential issues.
They also worked together with Landis+Gyr to add security robustness required by the utility to support enhanced customer engagement. The meters were then linked to a Snowflake cloud database and a Verizon 5G cellular network and placed into service.
“There’s one utility I’m talking to who’s just in the early stages and they’re doing what I’ll call the big bang or the waterfall,” Rutfield said. “They’re doing meter analytics and all these things and they haven’t even deployed a meter yet.”
One day, National Grid’s smart meters could allow the utility to consider time-of-use rates. But the suite of apps available on each meter present an even greater opportunity: creative tariff structures to better utilize and compensate demand-side resources, like customer-owned solar PV, battery storage, and electric vehicles.
Rutfield, and National Grid, didn’t let the grandeur of opportunity slow progress, though. That will come in time, and many of the possible use cases don’t even exist yet. Departments throughout the utility are taking notice, positioning AMI as a key foundation for advanced grid operations and system planning.
“The journey is all about the data,” Rutfield said. “You could never make these decisions before. Other groups are now saying, you have these meters out there, I want to start using the data, which is pretty cool.”
Timing the jump to AMI 2.0
National Grid was in a somewhat opportune position. Due to years of regulatory delay, the utility leapfrogged AMI 1.0 for 2.0. The timing was ultimately ideal to take full advantage of technological maturity, DER penetration, and the current political climate.
San Antonio-based CPS Energy, one of the largest municipal utilities in the U.S., is at a much different juncture. The utility completed a deployment of more than 1 million smart meters in 2018. The primary use case was being able to read meters over the air and reduce truck rolls— a perfect job for AMI 1.0.
In a given day, one of CPS Energy’s residential electric smart meters will make 96 interval readings per day. Multiplied across their full system, that’s a lot of data. How that data is stored, and ultimately used, is the most critical consideration for any utility AMI program. The jump to AMI 2.0 would exponentially expand that volume, bringing with it real-time data collection and edge computing.
Before a recent promotion, Michael Cervantes was tasked with managing CPS Energy’s AMI teams. But neither he nor his team members are trained data scientists— they’ve learned on the fly and taken advantage of advanced analytic tools, like Itron’s Operations Optimizer, to gain more value from AMI 1.0.
Timing the transition to AMI 2.0 could prove challenging for CPS Energy. Its existing fleet is only a few years old, and they’ve only scratched the surface on fully utilizing the data.
For now, CPS Energy is on a fact-finding mission. Engaging with utility peers, assessing technology, and building out future use cases.
“I would say we’re at that midpoint between AMI 1.0 and 2.0. There are plenty of use cases that we leverage off of our current solution that keep us busy,” Cervantes told POWERGRID. “We’re ready to go forward. We have a lot of technology changes happening over the next couple of years to help us leap into that direction.”
Part of the evolution relies on an internal culture shift, Cervantes explained. That can be a bigger lift for a utility that owns T&D infrastructure and power generation, like CPS Energy. They’re working to nail down how AMI data collection and analysis should influence broader utility decisions.
“We’re in the mode of spreading that knowledge,” Cervantes said. “Even if AMI 2.0 takes time to implement, there’s plenty of hard and exciting work that’s going to be done over time to really embrace all this information we’ve never had.
“I think every utility is going to face that.”
More than a meter
Tom Deitrich, CEO of grid edge intelligence provider Itron, doesn’t use the phrases “AMI 1.0” or “AMI 2.0.” Like many, he feels they fail to capture the value of the technology, which in turn slows innovation and progress.
The focus on “meters” is what requires a culture shift within some utilities. The latest iteration of smart meter technology is so much more than a meter— it’s a computer at the grid edge, capable of analyzing real-time data at the point of collection. The technology goes beyond customer energy management, too, by detecting and reporting power grid health with improved accuracy over first-generation smart meters.
Tampa Electric wanted to measure the performance advantages of that technological evolution. What would become a deployment of 800,000 distributed intelligence-enabled Itron smart meters began with third-party lab testing in 2020. During the study, distributed intelligence meter apps identified 100% of meter bypass incidents, residential neutral faults and high impedance occurrences. Compared to the accuracy of cloud-based analytics — 58%, 0%, and 0%, respectively — the case was clear.
“We get one shot to buy the right hardware to achieve the direction the utility wants to go in over the next 15 to 20 years,” Tampa Electric’s senior director of operations technology and strategy, David Lukcic, said. “Not all use cases are a fit for (distributed intelligence), but when they are, the value far exceeds back-office results.”
GO DEEPER: Inside Xcel Energy’s journey to deploy 3 million smart meters
It’s for those same reasons that Deitrich rejects any suggestion that the smart meter market is due for a slowdown, since smart meters already serve 80% of utility customers in North America. That’s before considering the impact of artificial intelligence, which is already helping utilities better understand the data they’re collecting.
“We have 11 million endpoints out in the field that are able to download distributed intelligence applications. With another 10 million in backlog,” Dietrich told POWERGRID. “I don’t have a single conversation with customers that doesn’t have something to do with this concept of agility and flexibility in infrastructure.”
One day, these endpoints could allow utilities to control customer-owned DERs without advanced grid software tools like DERMS or ADMS.
Landis+Gyr is working with another vendor to deliver a grid edge application for a utility that will connect a smart meter directly to a solar PV inverter. For the first stage of the project, Landis+Gyr is monitoring the solar PV from the meter and reporting performance. The next phase involves sending commands to the inverter to turn it on and off. Finally, the meter will be able to control the inverter.
Chris Patton, Landis+Gyr’s principal product manager, said the bifurcation of meter-to-cash from applications allows utilities to nimbly iterate the latest advancements.
“It’s kind of that walk-run mentality where you can actually test these things without having to go with the full waterfall approach where we wait until the very end to push this out,” Patton said.
‘Enough waiting’
Smart meters are now uniformly recognized as a foundational element of any modern electric grid. Even some of the last holdouts are coming around.
Public Service Company of New Mexico, which serves around a half million customers in the state, is one of the last remaining investor-owned utilities in the U.S. without smart meters installed on its grid. But that wasn’t by choice, at least on the part of the utility—state utility regulators have debated investing in smart meter technology for nearly two decades.
The New Mexico Public Regulation Commission finally relented on Aug. 16, granting PNM’s $344 million grid modernization request, which includes $171 million for AMI. The approval, the commission said, supports technology that is “neither novel nor new” and allows PNM to “catch up” with peer utilities and the evolving needs of the energy transition.
“We have waited eighteen years. That is enough waiting,” the regulator wrote, noting the commission first considered smart meter investments in 2006. “Cost management is important, but the time for achieving cost savings by waiting is over.”
When it comes time for PNM to deploy its new meters, National Grid’s Rutfield will share the single tip he delivers to every inquiring utility.
The message harkens back to a career focused on speed and scale. Can a utility think like a startup without mortgaging its obligations to safety and reliability? Rutfield and National Grid think so, and the success of the energy transition may depend on that new way of thinking.
“The idea of starting small and then crawling, walking, running, that’s working out really well for us,” Rutfield said. “Customers would not get benefits until much later if we used the waterfall approach.”