Go back

Unknown

23m 25s

Unknown

Syed Madini, co-founder and CEO of Verse, has spent his career bridging academic expertise in power systems and electricity markets with real-world energy operations. Starting with degrees in power engineering and market design, he advanced through roles at PG&E, Tesla, and a startup focused on energy asset optimization. His experience revealed a critical market gap: large energy buyers, especially data centers, face immense complexity in managing power procurement, contracts, and financial risk. As demand for compute grows and electricity markets become more volatile due to AI and electrification, data centers are increasingly adopting on-site flexible assets like large-scale batteries to avoid costly grid delays and transmission investments. Verse addresses this by offering AI-driven software that orchestrates these assets in real time—balancing customer needs, utility requirements, and market conditions. This enables cost reductions, avoids T&D and demand charges, and supports a more distributed, resilient grid. Syed believes this shift is not just technically feasible but commercially essential, with 100% of new data centers expected to include flexibility by 2027. Despite policy and regulatory hurdles, especially in utility execution, momentum is building rapidly. The future grid will be distributed, intelligent, and responsive—where energy buyers act as active participants, not just consumers—driving down costs for all while modernizing the system for a compute-heavy world.

Transcription

4034 Words, 22738 Characters

English
This is partner content from Latitude Studios. Syed Madini has spent a lot of his life trying to figure out how electricity markets work and how to make them work better. He started with an undergraduate degree in power systems engineering and a master's in electricity market design, then a PhD in applied operations research. Basically, he studied the grid from every angle. So I did all I could to train myself academically for designing and running power markets and power grids. Are you a power nerd, like at an early age? What got you into this? I'm not going to say when I was three, I was in love with electricity markets. But during undergrad years in college, I really loved financial systems and markets. And I loved engineering. Electricity markets combined both. But for Syed, the research was never the end goal. He wanted to put it to work in the real world. I didn't complete a PhD to be a professor, although my dad is a professor, is retired now. And I had that DNA. But for me, it's mostly applied. How do we apply these things? So he joined PG&E, helping manage the utility's power portfolio and California's wholesale electricity market. It allowed him to see how market theory held up when real money and real power assets were on the line. And from there, it evolved to, "How do we do this better? How do we build software using AI and all of those fancy things that we learned at school to make sure these large energy assets are run efficiently and they can be monetized properly in electricity markets?" That question took him into the software business. At Tesla, he helped develop an early version of AutoBidder, a trading platform for large batteries. Then he led a startup called AMS, which built software to optimize distributed power assets. When AMS was acquired by Fluence, he became the chief digital operator. Then he became the chief digital officer, focusing on scaling AI solutions for batteries and renewables. Each company had a common mission: help owners of energy assets make more money in the power markets. But along the way, Syed uncovered a blind spot. Those platforms and products, it would take large, big batteries and other renewables and try to monetize them in markets. So the end user essentially is the market bidding entity, the IPP, the developer, the trading shops, the supply side of the equation. And throughout those years, I realized that the buyer side of this equation is extremely under-served. So that was really the push for us to create Verse. Let's create a platform, a team, a product that takes all the complexity that energy buyers are facing with and make it go away. What did you actually see happening in the market that made you focus on large buyers? I would say in late 2020, In late 2022, the load growth story was already cemented, not because of just data centers, but because of the thesis of electrify everything. So we all knew that load is going to increase, prices are going to increase, the volatility in electricity markets are going to increase. Now we had an extreme catalyst, which was the birth of OpenAI and ChatGPT and therefore LLMs and AI and superintelligence, now we're in this big, big, kind of a massive problem that we're dealing with, but the trends were there. For big electricity buyers, higher prices and more volatile markets meant more contracts to manage, more risk to track, and more complicated decisions. Then, AI turned up the pressure, and that also changed what happens at the site itself. You actually need to deploy physical assets, because if you want to slash your costs by 20%, you can't just do it by reorienting your contracts or think that, you need to deploy physical assets, because if you want to slash your costs by 20%, you can't just do it by reorienting your contracts or think that, differently about your risk management, you need physical assets that can operate on your key sites, and the operation yields to avoided T&D charges and capacity charges and reduces the cost. Historically, on-site assets like batteries were used for backup or managing monthly bills. But for data centers, they're the difference between connecting to the grid today or waiting five to seven years. It helps with the speed interconnection by showing, I'm flexible, I can be dispatchable, and as a result, I can be on a more speedy path to get interconnection. And obviously, with all of these large assets, you can orchestrate them well to bring the bills down as well. Data centers have a stark choice. Wait years for the grid or go entirely off-grid. But there's a third pathway emerging, using flexible resources to earn faster. In this episode, Syed Medini, the co-founder and CEO of Verse, talks with Stephen Lacy about what that flexibility strategy should look like. They talk about the role of batteries, the economic benefits, and why software is so critical to making it work. So our vision is to become a one-stop shop for these large enterprises that spend hundreds of millions of dollars, if not billions on power, which is growing every minute. Syed, what is your vision for the future of the grid? Syed Medini: Historically, we've seen this as two ends of a spectrum. Either you sit and wait for the grid connection, seven or however many years, or you build it entirely off-grid, and you're saying that there's this third pathway, which is to become more flexible. So what are you seeing in terms of customer adoption in their willingness to be flexible, what kind of flexibility they're willing to tolerate? seven or however many years, or you build it entirely off-grid, and you're saying that there's this third pathway, which is to become more flexible. So what are you seeing in terms of customer adoption in their willingness to be flexible? What kind of flexibility they're willing to tolerate on the data center side? Yeah. So I would say the majority of these customers, they don't want to be in the business of power, they want to be in the business of compute. But they're so entangled, and you can't really separate them. So even data centers are measured in units of megawatts, how much power they're consuming. So with that mindset, their number one goal is to get connected to the grid sooner. get connected to the grid sooner. And essentially at a high level, without going into too much details of the nitty gritties of all the policy proceedings and the utility tariffs and what's happening, at a high level, the reason that the data center can get connected, if you look at the engineering problem associated with it, it could be there's not enough capacity and power at the grid level to serve that data center. Or it could be there may be enough power, but there might not be enough transmission and distribution infrastructure power lines to get power to your location. So in this case, what's happening is the data center cannot say, if I'm using a 100 megawatt example, cannot come forward and say, I'm just going to use 100 megawatts around the clock. They need to be curtailed. They might apply for 20 megawatts as firm power. And the 80 megawatts, they're going to say, cut the power from me. I'm fine. That doesn't add up with the SLAs they have. They can't just cut compute by 80 percent over a sustained period. So what happens is you need to equip them with on-site energy assets. And these energy assets, we're not talking about small batteries in a basement. We're talking about a couple of hundred megawatt systems, which five years ago, they would have been the largest in the world. Now they're behind a meter. So if you equip them with that system and you solve one of the most complicated problems that are out there, which is orchestrating, you're going to be able to do a lot of things. And you're going to be able to do a lot of things. And you're going to be able to these assets based on the data center needs, the utility needs, the grid needs, the ISO needs, and the market needs. And that's what we do through our software. If you do that, then you're not disrupting the energy usage of the data center. You're basically contributing and giving back to the grid. And in many cases, you can avoid investments in transmission and distribution infrastructure and really bring the bills down for the end customers. We just published a paper detailing the math behind this. And we're going to be able to do a lot of things. And we're going to So that's really the solution. And I would say there's a lot of willingness to adopt this in the market. But we've got to do a little bit more to make sure the rules are clear. And what is most complicated in orchestrating between market and customer and asset? Yeah. Having worked in the world of battery operations and energy asset operations, let's say you deploy a 500 megawatt battery system. That's huge. Once you energize it, it's not going to do anything. It's going to do nothing. It's just going to sit there and stare at you. By virtue of deploying a large battery, you haven't solved the problem. It really depends on how you operate it. And by operations, I mean every second, if not every millisecond, you need to tell it what to do. Essentially, you bought a car. It's on the parking lot. Well, you need a driver. And there's no human drivers here. It needs to be automated because the level of decision making is every second. So what do you need to do? You need to be able to do a lot of things. First, you need to serve the needs. In this case, let's say a hyperscaler, the needs of the hyperscaler in the data center. It could be AI load smoothing. It could be voltage stability. A lot of different things that they need. Backup power. So essentially, you need to operate the assets every second to serve the host customer. And the promises you made to the utility by avoiding T&D and by avoiding hours of consumption on hours when the grid is stressed, you need to fulfill it. So that means any custom tariffs that the utility imposes, any CP charges, coincident peak charges, any demand charges, which are essentially a symbol of when the grid is stressed, and managing the power at the connection lines you need these assets to operate transmission systems are congested fire up the battery so you can alleviate that transmission bottleneck and without cutting the power from the data center and call it a success so that's a whole host of requirements and then you also want to give back to power markets you want to participate in capacity markets energy markets demand response markets ancillary service markets so when you stack them all up this technology problem is extremely massive there's a ai and intelligence problem and then there's a control problem so we essentially stack all of that up and in a seamless over the air kind of deployment we can make these large physical systems cooperate and work seamlessly as possible but we have this commercial pull right now where data center owner-operator clearly see the value and flexibility for both speed and and operational cost what about the the push from the market what are you seeing out in the market in terms of policy regulation that is making flexibility necessary yeah i would say there's a lot of great work that's happening i mean for essentially all these proceedings around bringing your own new capacity or bring your own flexibility it's happening at the fork level they get different forms and flavors but fundamentally all they're saying is that we're not going to be able to do that we're not going to be able to do that so what we're saying is you got to get curtailed you can connect now but you got to get curtailed go and equip yourself with flexibility so you don't ruin your slas that's essentially the fundamental message so so far so good but then at the end of the day the entity that you're dealing with within a connection are the utilities and we all know the isos don't have jurisdiction over the utilities it's the commissions so that line of command really breaks down between the final level of implementation between high level policies and and the utility kind of execution so that's why the commissions are involved and each utility has its own taste and flavor and to the extent we can streamline all of this that will be phenomenal because you know i always say you know the vision of moving towards a distributed grid it's not news i mean people have been writing papers about it for decades now we're at this place that the world's largest financial institutions are willing to finance this distributed grid for the sake of getting power sooner it'll be a shame we fumble at the five yard line we just need to make sure the policies are there because at the end of the day this absolutely helps regular households to bring their costs down so why not do both win the ai race and bring their costs down and make sure we we double down on this you studied the value of find in terms of how it could potentially bring bills down in the state? Yeah, basically what we found is, and it's all published on our website, I had my team of PhDs in power markets essentially do this math, because what we wanted to answer was, what does good look like? We all know if we add 480 gigawatts in Texas in form of data centers, it doesn't make, it's not rocket science to understand that's going to drive up power prices, because new generation needs to be invested in, new transmission lines need to be invested in, the utilities take their rate base, and they essentially socialize it across everyone. That should not happen, because if that happens, we're talking about every household in Texas will increase their utility bill by 200 bucks a year. That should, that doesn't work. It's not good for everyone, and I don't think any of the hyperscalers have that intention, and they want to avoid that. So the next phase of this is, what if these data centers bring their own power? What that means is, they invest in generation capacity anywhere on the grid, and that helps. That really eliminates a lot of the kind of extra costs in households for energy bills. To the extent from 200 bucks for if we're adding 10 gigawatts of data centers, that brings it down to 10 bucks or 11 bucks. But the next level is, bring your own capacity and bring it on site. And create that flexibility, meaning that you're seamlessly operating to avoid any bottlenecks in the transmission system to get power to your location. Not only you're avoiding investments in generation capacity, but you're also avoiding investments in transmission systems. That actually helps by reducing the bill for the end customer, because you're avoiding investments in transmission systems, which everybody needs to pay for. So those are the. The benefits of a distributed grid, again, not rocket science. And we're at the moment that a lot of folks are willing to finance this, but we just need to make it happen. I'm curious about what you're hearing from customers about whether they think they'll be off-grid permanently, or whether they'll eventually connect to the grid. What do you think that bridge to the grid looks like? And will we see a lot more development that is separated from the grid? Yeah, I personally don't think anybody has a desire to be off-grid. They see more work. They see total independence, which they need to get into the 99.99999% SLAs reads. These are not micro grids. They're macro grids that we're talking about. They're large power systems that people are thinking about building. We support them as well. But I think inherently, people see that as the last resort. So if we can fundamentally clean up the rules, and make sure that these assets can be connected with the flexibility. I don't think people have this secret desire to go off-grid. They're doing it based on necessity. And how do we streamline it? It's going to be the key question that is out there. How quickly has this conversation moved? It felt like overnight, data center developers were taking batteries very seriously. Sometime in the last year, demand and interest really picked up. What did you see from your side? Was it happening well before that? And we've had this academic conversation for a while on the value of flexibility, but now the rubber is truly meeting the road. So I'm just curious about how you have seen this conversation and ultimately execution advance in the last year or two. I would say the concepts that we're talking about, if we were talking about this 10 years ago, it would have been the same, but we would not be using the word data center. We were talking about large load, the benefit of a distributed grid, DERs for bringing down essentially non-wire-as-all, and a lot of those alternatives, which was around 15 years ago. But obviously that momentum has changed because of the need for compute and compute as power. So I would say the conversations have picked up massively in the past 12 months. In particular, as we're seeing these policy proceedings are overall moving in the right direction by advocating for a data center that's independent, brings their own power, brings their own flexibility. Compared to 12 months ago, I would say a lot of major improvements. So we're at the final line of making sure the utilities adopt what the right framework is for the final layer of execution and interconnection. So that's all good. So compared to power system years, we have accelerated a lot. But compared to AI years, still a dinosaur. So I think that's the fundamental cultural shift that we need to make, which is, we need to make sure that we have the right framework, we need to make sure that we have the right resources, we need to make sure that we have the right resources, which is, in order to win this race, you can argue that this is a good race or a bad race. That's another episode probably with somebody else. But if we argue we want to win this race, then we need to act much, much faster. What's it like building a company right now that's right in the center of growth? You've gone from 25 to 65 people in the last year. You're expanding really quickly. What is that ride like right now? Yeah, I mean, that's kind of the fundamental. Building companies are art and science. The science, you can kind of think about your burn and what type of engineers you need, what's your go-to market strategy. The art is making sure you put your foot on the gas pedal when it's needed. I've seen a lot of companies fail by over-hiring at the wrong time, by under-hiring at the wrong time. So I think that's something that we need to think about. I think we need to think about how we're going to do it. Yeah, I think we need to think about how we're going to do it. All of this orchestration and flexibility doesn't fall from the sky. It takes a lot of smart people to put together these essentially engineering layers of software and AI to make it happen. The current mood about data centers is pretty dark in this country. Lots of claims about data centers raising power prices, sucking up resource use, et cetera. And so I think there are people in the power industry who see this as a really optimistic moment. Certainly not discounting some of the issues related to large-scale data center development, but seeing this as a chance to modernize the grid, to execute on the concepts that we've been talking about for years around data. I think distributed power and flexibility. And I'm curious about where you sit on that spectrum. How optimistic are you that we could use this moment to advance the grid in ways that we've been talking about for years? I'm very optimistic because it's not just a policy issue anymore. I think It's not bringing ISOs and utilities in a room and raising the fact that we think about this differently. There is a strong business need associated with it. Whenever there's a business opportunity, I think the opportunities just naturally increase. For all the folks out there listening to this, absolutely true. Just a plain vanilla data center would increase costs, and nobody wants that. Nobody can argue against that. But I don't think that's going to happen. A plain vanilla data center with no power, no flexibility, just connecting to the grid, it'll take years and years and it's not going to happen. So we're in this moment now that flexibility is key, and flexibility is what's going to help you bring your costs down. So as a person that's worked in this space, I can stand behind our numbers and basically say, if you deploy large DERs and if you operate in well, we're going to get a more modern power system, and we're going to bring costs down. Now, somebody is paying for it, for that essentially DER infrastructure. Those are the ones that have the business opportunity and they're willing to do it. Pick a date, let's say 2027. What percentage of data centers are going to have some flexibility component built into them? I would say 100 percent of data centers that go online, we're going to have a flexibility. Now, the question is, how many are going to go online? How many do you think? I would say 10 to 20, but we'll see. Yeah. So paint your vision for the future grid. You've been at this for a while, five years from now as we see these large loads become more interactive with the grid. What does that enable? What does the system itself look like to you? Yeah. I mean, about 10 years ago, I said the future of the grid are batteries and advanced energy assets all distributed with a bunch of algorithms. I think we're getting to there. I think the modern grid will be a grid that extremely flexible, resilient, and distributed. I think there's no question to ask. The way we want to play in this space is not only think about the orchestration that benefits our customers and the grid, but also serve the other bits and pieces of what the customers want. We talked a lot about the orchestration, but that just doesn't end there. Folks need power procurement, folks need risk management, folks need contract management. So we've essentially pumped out these products. It's a cliche statement, but that's how you become the operating system for energy as it relates to enterprise. I mean, today we're managing 30 gigawatts, which is 14 Hoover dams, and I think in five years, those numbers are going to increase tremendously. Syed, thank you so much. Syed Madini is the co-founder and CEO of Verse. The company develops software to help large electricity buyers manage power procurement, contracts, and financial risk and coordinate energy assets like batteries. Verse's dispatch intelligence platform is designed to operate those assets for the benefit of customers and the grid. Learn more at verse.inc or follow the link in the show notes.

Podcast Summary

Key Points:

  1. Syed Madini transitioned from academic research in power systems and electricity markets to apply his knowledge in real-world energy operations at PG&E and Tesla.
  2. He identified a critical gap
  3. The rising demand for data center power, driven by AI and electrification, has intensified market volatility and created a need for on-site flexible assets like large-scale batteries.
  4. Verse was founded to provide AI-powered software that orchestrates large energy assets in real time, enabling data centers to reduce costs, avoid grid bottlenecks, and meet strict service level agreements.
  5. Syed believes the future grid will be distributed, flexible, and resilient, with large energy buyers deploying on-site assets to reduce transmission investments and lower end-user bills.
  6. A key shift in policy and market dynamics—especially at utility and ISO levels—is accelerating adoption of flexibility strategies for data centers.
  7. Current market trends show strong commercial interest in flexible energy solutions, with data centers increasingly viewing flexibility as essential for speed and cost control.
  8. Syed envisions 100% of new data centers incorporating flexibility by 2027, and a future grid where software orchestrates distributed assets to create cost savings and grid stability.

Summary:

Syed Madini, co-founder and CEO of Verse, has spent his career bridging academic expertise in power systems and electricity markets with real-world energy operations. Starting with degrees in power engineering and market design, he advanced through roles at PG&E, Tesla, and a startup focused on energy asset optimization. His experience revealed a critical market gap: large energy buyers, especially data centers, face immense complexity in managing power procurement, contracts, and financial risk.

As demand for compute grows and electricity markets become more volatile due to AI and electrification, data centers are increasingly adopting on-site flexible assets like large-scale batteries to avoid costly grid delays and transmission investments. Verse addresses this by offering AI-driven software that orchestrates these assets in real time—balancing customer needs, utility requirements, and market conditions. This enables cost reductions, avoids T&D and demand charges, and supports a more distributed, resilient grid.

Syed believes this shift is not just technically feasible but commercially essential, with 100% of new data centers expected to include flexibility by 2027. Despite policy and regulatory hurdles, especially in utility execution, momentum is building rapidly. The future grid will be distributed, intelligent, and responsive—where energy buyers act as active participants, not just consumers—driving down costs for all while modernizing the system for a compute-heavy world.

FAQs

Syed was drawn to electricity markets during his undergraduate studies because they combined his interest in financial systems and engineering. He saw the potential to apply academic knowledge to real-world problems in power systems.

Verse is a software platform that helps large energy buyers manage power procurement, contracts, and financial risk. It solves the challenge of complexity faced by buyers in volatile electricity markets by enabling seamless operation of on-site energy assets.

Data centers face rising electricity prices and market volatility, requiring them to manage risk through flexibility. Deploying on-site assets like batteries allows them to avoid transmission bottlenecks, reduce costs, and connect to the grid faster.

Verse uses AI and real-time dispatch intelligence to operate large energy assets—like batteries—every second to meet data center needs, avoid grid congestion, fulfill utility requirements, and participate in energy markets simultaneously.

The future grid will be flexible, resilient, and distributed, with large energy assets operating in tandem with customers and the grid. This shift enables cost reductions, avoids transmission infrastructure investments, and modernizes power systems.

On-site assets reduce energy bills by avoiding costly transmission and distribution infrastructure investments and by managing peak demand, which in turn helps lower electricity costs for data centers and other large users.

Chat with AI

Loading...

Pro features

Go deeper with this episode

Unlock creator-grade tools that turn any transcript into show notes and subtitle files.