Aug 6, 2025

Guest Editorial | Microsecond Data on Ami 2.0 Meters Allows Utilities and Customers to Work Together in a Complex Grid Landscape

by Mike Phillips, Sense

Utilities are under pressure.

The grid is becoming increasingly strained and complex. Demand is soaring; forecasts call for a 200% to 300% increase from current levels by 2050. Home and vehicle electrification and renewables integration are simultaneously creating new decarbonization wins and varied intricacy. The need to measure, manage and balance all of this – while remaining affordable and resilient – weighs heavily.

There’s no doubt about it: the grid must be made more intelligent and more connected to support this transition. We can no longer afford to operate under the status quo user-and-supplier framework that’s existed for so long. It’s vital and urgent that we rethink the relationship between utilities, their customers and the grid they share. It must become a more flexible, responsive and efficient system. And utilities and customers must become more connected and responsive themselves – through the grid. Otherwise, we’ll all fall short of capacity needs and Net Zero goals.

The good news? It can be done. We just need to give a makeover to one of the most critical technologies at the forefront of the energy transition: the humble home energy meter.

Meters are key

Utility meters have traditionally been thought of as simply data collection devices. The first smart meters served a fairly small range of functions when they were installed in the early 2000s. They eliminated the need for manual meter reading, but weren’t much more than just that; basic reading and billing devices that operated on 15-minute interval data. That’s not nearly enough resolution to support today’s energy system needs. And it most certainly won’t be enough to support the grid of the 5, 10 and 20 years.

AMI 1.0 meters are hitting the end of their 20-year lifecycles now, too. It’s estimated that 25% of home energy meters will need to be replaced by 2030. The need to replace meters and the rapidly changing energy demands of the future meet at an opportunistic and timely intersection.

Enter the next wave of next-gen smart meters. AMI 2.0 devices, equipped with embedded intelligence and the ability to process microsecond-level data, are multi-purpose solutions for electric utilities and their customers. They turn meters from static devices into the distributed sensing, compute and control platform for the modern grid. By living in the intersection of homes and buildings and the grid, these next-gen meters can provide a real-time view of the entire distribution system. This insight is deep and nuanced, carrying all the way down to the device level.

So, what will it take?

To make this quantum leap in visibility and control all AMI 2.0 meters need three things:

High-resolution data: Meters have recently been announced that can continuously sample voltage and current waveforms up to one million times per second. That’s 50 million times more data processing than first-generation smart meters. It’s useful, too. This level of resolution turns data from incomplete and siloed, into comprehensive and decision-informing. This drives proactivity across multiple planes; from the present moment to the long-term, with learnings that prioritize targeted upgrades and investments.

High-resolution data lets utilities and grid operators track device-level consumption behind the meter and detect minute anomalies across the grid. Subtle variations in voltage and current waveforms can be used to detect and localize issues before they cause outages or damage equipment. These issues can be as granular as seeing singular transformer arcs or vegetation brushes on lines–in real time.

To deliver these capabilities, AMI 2.0 meters should be equipped with:

  • Continuous and synchronous sampling of voltage and current waveforms at a minimum of 15,000 samples per second
  • Linear quantization with at least 16-bit resolution for current and at least 14-bit for voltage
  • Raw energy data accessible to the application processor – beyond just summarized information
     

Local computation, memory and storage: With 50 million times more data, transmitting everything to a centralized cloud location to then make these decisions is not possible. Instead, AMI 2.0 places powerful computation on the grid edge. Embedded processing creates insights and automated operational efficiency adjustments. Advancements in the mobile phone industry have made this possible. This level of computation is now energy-efficient and cost-efficient as well. And it has opened doors in the utility space. Hardware accelerators on next-gen smart meters can run AI models locally and process high-resolution data rates in real time on a per-needs basis.

The latest AMI 2.0 meters leverage these processors by using a distributed software model. Much of the processing happens locally on the meter–combined with real-time networking and cloud-side processing as needed.

To deliver these capabilities, AMI 2.0 meters should be equipped with:

  • 1000 DMIPS CPU processing power
  • 256MB RAM
  • 1GB Flash storage or other storage
     

Real-time networking: A distributed computing model reduces the strain on utility networks, but realizing the full potential of a responsive smart grid still relies on reliable connectivity.

While many applications can function with modest bandwidth, others require rapid, low-latency communications. This includes some of the most vital flexibility and resilience functions–like real-time DER coordination, voltage optimization and outage detection.

To deliver these capabilities, AMI 2.0 meters should support:

  • 30-1000MB/day of non-real-time data transfer, from meter to cloud
  • Low latency (less than 750ms round trip) for real-time applications requiring rapid control signals
  • Local home network access (WiFi or Ethernet) to enable seamless smart home devices
     

Benefits on both sides of the meter

More data enables new capabilities. Just as data made possible advancements in larger language models and AI, grid-edge processing of high-resolution data enables the next cohort of real-time intelligence needed to manage the largest and most complex single project that humanity has ever undertaken.

More data also breeds adaptation. By embedding edge-computing software, AMI 2.0 meters can adapt over time and accommodate the consumer-driven transition to smart homes, all without needing to replace hardware. This approach facilitates whole-home electrification. On the utility side, consistent software updates can be made without hardware upgrades and future-proofing equipment for an evolving energy landscape.

Real-time visibility is not new at the generation and transmission levels. However, the progression of visibility into last-mile feeder networks, distribution transformers and behind-the-meter is a strong evolution. A better understanding of what is happening on the grid is the result. And that allows utilities and grid operators to take targeted, prioritized action.

With the ability to granularly identify without delay comes the ability to influence. Automation and efficiency via meter-hosted machine learning allow utilities to exert demand flexibility not only for overall load and demand but also to avoid system peaks throughout the grid. This peak avoidance runs deep, to the service feed into homes and buildings. A view of real-time data across a vast number of DERs can also help utilities adapt to new load profiles.

1MHz sampling also opens a new chapter in dynamic rate management. Dynamic energy pricing models can accurately reflect real-time energy usage and grid operations.

Consumers benefit as well.

Again, this oncoming wave of AMI 2.0 meters is primed to transform the relationship between homes and the grid. These devices even give utility customers a similar POV as grid operators. They can see their assets (for example, a microwave) across their system (behind their meter) and use energy with their command control center (an app on their phone) in real time. It’s precise enough for consumers to see the wattage used on a per-lightbulb basis.

With real-time access to this data, consumers can make informed decisions about energy efficiency and even automate their energy management. This shift towards greater control empowers consumers to play an active, engaged role in optimizing their energy consumption and contributing to the shared grid’s overall efficiency.

A case study on resilience

Grid resilience is at the top of all utility stakeholders’ minds. Grids must withstand extreme weather events that are rapidly increasing in frequency and intensity. In just the past year, wildfires in Southern California left hundreds of thousands of residents powerless and Hurricane Helene caused outages that impacted more than five million customers. Outages threaten lives, particularly in service areas with elderly and disadvantaged populations.

Effective vegetation management is crucial for safety, reliability and expense reduction. Southern Company’s Georgia Power has collaborated with Sense, a leader in grid edge intelligence, on a pilot program to enhance vegetation management.

Between 80 to 90% of outages within the distribution grid are caused by object-on-wire faults. These faults are notoriously difficult to detect. They’re typically uncovered through line-by-line segment inspections carried out by field teams. Bucket trucks good, old-fashioned man-power and visual inspections are the preeminent tactics in this critical work.

Georgia Power uses predictive technologies, like LiDAR, to measure the proximity of limbs and lines. They’ve supplemented this with Sense’s high-resolution data to determine if better grid visibility can be leveraged to identify the occurrence of disruptions.

The hope underpinning Georgia Power and Sense’s pilot program is that response times can be improved, thus enhancing safety and reliability for customers. Georgia Power is installing Sense retrofit monitors into the homes of participating Atlanta-area customers. The installation of these AMI 2.0 devices, using ultra-high resolution data and embedded AI, will generate an improved view of the local distribution system and a better understanding of when and where grid anomalies like tree slaps occur.

Georgia Power intends to use this disaggregated data to detect vegetation issues earlier than before and manage them proactively. That tightens and prioritizes vegetation management efforts while also removing the aforementioned manual detection process. It also produces smoother, stronger regulatory compliance. Most importantly, it creates a safer, more resilient distribution system for all involved.

Georgia Power will also be able to unlock learnings that can help refine its response to future disruptive vegetation events. This is an in-progress, customer-impacting example of the quantum leap provided by AMI 2.0 meters with microsecond data rates. It’s also a great example of rethinking the relationship between utilities, customers and the grid they share. It takes innovation and proactivity, both directed towards shared values and goals.

Real-time visibility builds a cleaner, more flexible and resilient grid

The adage “you can’t manage what you can’t measure” holds true in the energy transition. Next-generation meters equipped to process 1MHz data and locally host AI computation are giving utilities the tools they need to proactively measure AND manage, in real time. By enabling deep and responsive insights and offering management automation, these meters foster a new, more collaborative relationship between utilities and their customers.

As we create a future powered by clean energy, now is the time to embrace high-resolution data at the grid edge. Over the next five years, we have a unique opportunity to accelerate grid-edge intelligence and maximize the capabilities and performance of existing infrastructure. By transforming home energy meters into smart, proactive, automated devices capable of sensing, processing and controlling energy usage, we can build a cleaner, more flexible and more resilient shared grid.

Co-founder and CEO of Sense since 2013, Mike Phillips brings decades of expertise to achieve Sense’s mission to transform the relationship between people, homes and the grid. Previously CTO and founder of a start-up with the first voice-enabled virtual assistants on mobile phones, Phillips is a pioneer in machine learning, bringing his capabilities to tackle the climate crisis.