 So our next speaker is Marisa Hamun. Marisa is the Chief Technology Officer at UTD Data Inc. And she's also a governing board member for LF Energy. For those who do not know what LF Energy is, LF Energy is one of the Linux Foundation project that bring the energy industry together. So that include like Greek companies and software vendors and even like academia. So she is one of the governing board members of that project. And today Marisa will be talking about the redefining electric grid operations from the Edge with Open Distributed AI. So please welcome Marisa Hamun. Good morning and thank you. I appreciate your attention and I hope to convince you by the end of this talk of three things. First is that the edge of the grid where the utility meets the customer is gonna change very quickly over the next seven to 10 years. Second, that our current tools are really not set up or have not prepared us well to handle that. And that third, that what we really need to do as an open source community is respond to that with open distributed AI. My name is Marisa Hamun. I'm the Chief Technology Officer for Utiladata, a real-time grid operations company that spent the last 10 years operating the grid both centrally and from the edge in. I'm also a governing board member for LF Energy. I started my career at the National Renewable Energy Laboratory studying how we were going to as a country and as a world quickly transition off of fossil fuel driven generation to solar and wind generation. We built digital twins before we had that word. That's now in vogue, but we simulated the grid at a very detailed level to understand where were the opportunities for changing how we would do things, for simplifying and making things more cost effective. And what became very apparent to me was that the edge of the grid where we use energy, where we decide what we wanna plug in had a lot more flexibility and inherent scalability than we'd ever given credit to. And from there I spent the last eight to 10 years of my life trying to bring technologies into the commercial space that really exploited that flexibility and that scalability in order to lower that cost of the energy transition. I joined LF Energy a couple of years ago and our company did at the invitation of the then-executed director, Shuli Goodman. She and I had been colleagues and friends for many years and she had a vision that if we could just bring together the community of technologists under this banner of open source technologies, open source standards specifications, that we would really lower the cost of that energy transition and speed up the pace with which we could do that. The other piece of that puzzle from my perspective is the commercial side of that. Where open source technologies meet commercialization is where we can really make that change happen. And so, Utiladata started about 10 years ago when the United States had a big push towards the smart grid. Smart grid technologies are now, it's almost like a bad word because we labeled maybe them a little bit too early. They don't turn out to be very smart 10 years ago. But the goal at the time was to digitize the grid, use that digitization to create a real-time picture of the grid and then use that picture then to optimize it. And so, over the last five years under my leadership for the research and development, we've been taking that technology to the edge of the grid. So, what do I mean by the edge of the grid? This is first and foremost where the utility, the electric company ends and where the customer begins. Both residential customers, industrial and commercial customers, it is where most of the change is going to happen. In this room, my guess is that probably only one in 20 of you own an electric vehicle. But by the end of this decade, I would bet one in two of you will own an electric vehicle. You'll probably want to plug that vehicle in close to your house, if not at your house. You'll also probably want to plug it in at work, maybe even when you're at the grocery store. For places where there's a dense population in urban areas, you can imagine that the street will now be lined with EV charging. Your grid is not prepared to handle that. The grid is an old, a very old object. It started probably roughly between 1920 and 1950 and was really designed with the idea that we would serve lighting, maybe eventually some heating, but was not prepared for things like electrification of vehicles. But yet, we are gonna see an explosion of those. So, not just electric vehicles, but we'll also see a transition to heat pumps. So, off of natural gas heating or even oil heating and to electrified heating. We'll also see an increase in the adoption of solar and storage technologies, in part due to economics and in part due to resiliency. Both of those things come into play. The economic piece happens really fast. You only have to look to Australia for an example of that. When the payback period for solar reaches less than three years, the adoption rate goes through the roof. So now about one in every three houses or businesses in Australia has solar on its roof. They had to go through a very massive change in how they thought about operating that grid in order for that grid to stay upright. In Japan, there are about 2,500 power plants. We think of these as the big power plants, right? Cold burning power plants, natural gas power plants. If every household, if every business, if every endpoint or edge of the grid suddenly has solar and storage and electric vehicles, now all of those are end points. Sorry, no, now all of those are power plants. That means there are going to be 80 million power plants in Japan. The utility right now can handle, you know, coordinating and dispatching about 2,200 power plants. 80 million is a much, much larger number. And in places where that change has happened quickly, what we have seen is, sorry, oh, is this? Okay, is this on? Great, all right, well, where that change has happened quickly, the grid is now considered at capacity. In the Netherlands, over the course of about 10 months, the grid went from being able to handle new connections for industrial plants or data centers to not being able to connect any new loads nor new solar. And it crept up on them because of some changes in the politics and economics of where they were getting their energy. I would contend that this is going to happen just about everywhere in the world, just at different times. So in some places where they are very dependent on natural gas today, this transition might happen very soon and other places it might take a little bit longer. But if we agree that the grid is going to reach capacity, then we have to figure out how we want to, how we want to respond to that. When the electric company decides that the grid is at capacity, they're using a very specific measurement for what capacity means. They take the worst case scenario. Everybody is running their air conditioner at the same time, plugging their electric vehicle in at the same time. For the single one hour peak load, that's what they measure the capacity to be. But if you took the average energy use over the whole year, most grids are at about 30% capacity. So we build an extra 70% to handle that one hour of coincident peak load and generally don't use that. And that means that there's a lot of opportunity for us to utilize more of the grid without actually changing our infrastructure, without necessarily building things. So I would contend we have three options for the future. One is that we do nothing. We slow down electrification. We slow down decarbonization. We give ourselves a chance to catch up under our normal build plans. The second is that we try to build our way out of it. I would say this is expensive, probably not really realistic. There are both material shortages, supply chain shortages, labor shortages that make that a very unrealistic path forward. Or the last is that the energy community needs to embrace technologies that other industries have already done. Things like manufacturing automation, just-in-time delivery of materials, the optimization of supply chains, even discoveries in medicine or fintech. Those, what those technologies all have in common is the use of digitization. So let's make measurements, turn them into data. And then making decisions in real time and coordinating those decisions. So if you were to apply that concept to the energy industry, what we start with on this far side here is visibility. You would be surprised how many substations in the world have almost no data being reported back to the electric company. Probably about 40% of them. So 40% of them, the utility company only knows if something went wrong, if it's literally on fire. That's a problem we could solve today. Once you have the visibility, you can move to better forecasting. From forecasting, you can move to then real-time management. And then by the time you get all the way to the future, we should be able to dynamically respond to the conditions that are in play today or at this moment and automatically create resiliency and efficiency. So if this progression is to happen, what I would contend is that we need a distributed AI platform that allows for this to happen. And it needs to be based on open-source technologies in order to increase the credibility of that tech and in order to increase the credibility, but also to support the innovation that we know needs to happen. There are kind of three things that I think I hopefully have convinced you of in the last 10 minutes. First is that the endpoints of the grid are going to become increasingly complex. There's gonna be an exponential change in how those endpoints are gonna operate. Second is that that change is going to produce a set of coupled systems that when solved really require a new type of mathematics in order to solve them. And then lastly, that by bringing the computation and the decision-making to the edge of the grid, we both avoid the latency issues of that round-trip communication, and also we avoid any privacy issues with the data. So, here's a small example. This is probably gonna seem incredibly trivial to you guys, but this is something that the utility has almost no visibility in today. What is happening at a particular endpoint in terms of the energy use does determine how the utility wants to respond to this. This is an example of how when we are measuring locally and computing locally, we can detect in real-time the difference between an oven starting and an EV starting to charge. No one should tell you when you can cook, but it turns out that charging your EV is actually a fairly flexible process. Generally, you're plugged in for longer than you actually need in order to charge the car. So, this demonstrates that you can have real-time detection of information, and you can use that information locally then to create the right kind of response. ElephEnergy is the home of a lot of such projects. They range from edge-of-grid operations all the way to central operations, market operations, you know, back through kind of power-flow optimization, and is really trying to underpin a change in how the utilities pursue these options. I think this is really important because this is how I got involved with them is a project called the super-advanced meter. We had run out of the word smart grid. We maybe overused that one, and so we decided this would be a super-advanced meter, and the super-advanced meter under ElephEnergy is a open-source standardization project, trying to develop a set of standards for what that hardware architecture would look like, as well as what the use cases are to justify that investment. Utiladata is helping to bring that technology to the market. We call it Carmen. Carmen is a platform for the software-defined grid. It is built on top of a, it's built on top of a NVIDIA IoT chip, the Jetson Nano, and it is designed to link the grid with the things that are, quote, behind the meter with EV chargers, with solar inverters and batteries. But it is also meant to be a key part of grid operations. So the measurements that you get right at the edge of the grid, right at the meter, can tell you what's happening in terms of power flow, but also tell you what happens behind the meter. So this foundation can be spread throughout the grid in places like data centers, transformer boxes, even commercial and industrial sites. With an open software-defined grid, we could set the stage for using distributed AI to create that efficient and equitable, and rapid transition to a decarbonized and electrified future. I'll leave you with this last simulation. So this is an example of how if you have the right computation and the right software-defined platform at the edge of the grid, we can dispatch locally the resources that right now we treat as fairly dumb. The inverters and the EV chargers, even your heating and cooling system, for the most part right now, don't respond to any sort of signal, but all of them have that capability, and it's up to us as a community as to how we wanna respond to that. So I think what I wanna conclude with, I guess as a call to action to this community, that the energy transition is going to require layers upon layers of technology, and the open-source community has the expertise and the skills and the experience to really drive that change. So please support LF Energy, and more importantly, keep doing the good work that you guys are doing. Thank you.