If you’ve wondered what role AI can play in the clean energy transition, this is an episode to listen to. Host Bill Nussey nerds it out a little with Titiaan Palazzi, Co-Founder and COO of Myst AI. Myst AI is the artificial intelligence platform that is shaking up electricity supply and demand forecasting with highly localized time series data that dramatically improves the reliability and costs for systems of all sizes, even microgrids. Titiaan shares insights into how Myst tackles the massive data sets, how their platform can support modeling on the fly, and how their radically different approach to a data ecosystem can spawn exciting new business opportunities for innovators and entrepreneurs.
Here are some of the highlights from their discussion…
“We’re trying to help our clients find the models and the data sources with the highest predictive performance. So whatever data source [is needed, it] is our job to get it in to our client’s hands. … Our vision for the future of Myst … is that we’re empowering clients to build their own predictive models in house…”
“…one thing that excites [me about] local energy is the fact that … decision making [about the energy transition] is distributed. So we can all vote with our dollars for what car we all drive, how our home is powered, whether we’re cooking on gas or on electricity. And that could mean that this really goes fast. “
“Amory [Lovins] used to say that what we need for the grid to work like an orchestra. We need the ensemble to make beautiful music. What is necessary for all of this local energy to work together is clear view of the future and then an ability to control it. That doesn’t need to be a master controller. It could be distributed, but you do need to have that foresight.”
You can also listen to this podcast and others in our series on these platforms:
Check out these links to learn more
Full Transcript
Bill Nussey:
Hello, and welcome to everybody in the Freeing Energy community. We are excited here in a beautiful spring in the US to have another fabulous guest on another really interesting topic. And we’re going to go just a bit nerd today, because we’re going to talk about a topic that is one of my favorites, artificial intelligence and what this means for the energy sector. And we have one of the founders of one of the top companies, who’s going to tell us what it means for him. So let me go ahead and introduce our guest today. It’s Titiaan Palazzi, who’s the co-founder and chief operating officer of Myst AI. It’s a San Francisco based startups using artificial intelligence to create a data analysis platform intended largely for electricity demand and supply forecasting.
Bill Nussey:
So, Titiaan grew up in the Netherlands, and he studied in Zurich. He has a master’s of sustainable energy technologies. And he’s been a visiting researcher in the energy systems division at MIT, probably heard of that school. Not long after that, he had this really cool role, which is how I met him, as the special aid to the founder of Rocky mountain Institute, Amory Lovins. In fact, Amory was my guest a few weeks ago. And when I went out and started the book project, end of 2016, was able to get an interview arranged with Amory. And so I go there to meet him, and who sort of greets me is this tall guy, Titiaan. And he introduces himself and tells me that he works with Amory and he’s going to help arrange it.
Bill Nussey:
And he did. And then I finished up with Amory and Titiaan says, “Hey, you want to just hang out?” So we went, I will never forget, for this walk behind the headquarters in Basalt, Colorado, with beautiful snow covering the ground, right on the edge of a river. And we just talked for … I think it must have been an hour or two, and really became friends, and have been keeping up ever since then. And it’s been so cool to watch Myst come from an idea into a company and to a successful company, and on its way to being a rocket ship. So Titiaan, welcome to the podcast today.
Titiaan Palazzi:
Bill, it’s so great to be here. Thank you for welcoming me.
Bill Nussey:
So we’ve known each other for a while. And like we do, let’s talk a little bit about your background, even well, before we met. You’ve been pretty active in youth organizations around the world, and particularly in the Netherlands. So tell us about that work, particularly the work you’ve done with the Kairos Society, and share how you got into this work and why you got started, and some background.
Titiaan Palazzi:
Yeah, definitely. Kairos Society is a network for college-aged entrepreneurs. It was started about 10 years ago, a bit more, in the United States, with support from folks like Bill Gates and Bill Clinton. And I had the chance to meet the founding team in 2011 in the Netherlands, in the Hague. And after that, I was really impressed. So I told them I’d like to start and build a Dutch chapter. And so I did. And it became an opportunity to meet entrepreneurs in their early twenties, across the world, in the United States, in South America, in Asia, in many parts of Europe. And it was really a formative experience in which I had a chance to hang out, organize conferences, start businesses with these young, bright starry-eyed other students. So it was a terrific opportunity that led me to make friendships for life, and actually meet my later co-founder in that role, too.
Bill Nussey:
Oh, wow. Okay. One of the things that I’m always impressed and a little envious is of folks who grow up outside the United States, because I think you have such … and you are an example, you have such a global perspective on the world. And I think it helps you become a better leader, particularly for issues like energy, which is a truly global opportunity, and climate change, which is truly a global challenge. So we’ve had a lot of guests on the show over the years, and many of them are somewhat athletic or very athletic. It seems to be a trend among a lot of the energy aficionados. There’s a lot of outdoorsiness and getting out there and having the energy of life in addition to the energy of photons, electrons and hydrocarbons. And so we found that, it’s a great story, about you cycled 6,000 kilometers from San Francisco to the White House to teach hundreds of children how to grow vegetables and build solar panels. I got to hear, what’s the story here? What motivated you to do that?
Titiaan Palazzi:
Yeah, and I think it makes sense why people who work on climate, who work on energy, have this passion for the outdoors, because that’s where a lot of our motivation lies, across people. I spent part of my grad studies at MIT. And while I was there, I was soaking up the environment. And I became friends with some undergraduate computer science students at MIT. I went back to Europe to finish my master’s. And one of the friends at MIT decided that he was going to bring together a group to cycle across the US, to invigorate in kids the love for learning, not in an academic setting, but by helping kids build projects that they wanted to see come to life.
Titiaan Palazzi:
So we raised some philanthropic money. We bought eight bikes. Seven MIT students, one Berkeley filmmaker flew to San Francisco and then cycled from west to east, stopping at about 15 locations in between where we would teach one-day engineering workshops, mostly to middle school students. One of my favorite memories from that trip was outside of Kansas City, where we met a group of probably 30 homeschoolers. And we did a sticky notes exercise where they told us what they wanted to build. And a small group of them decided they wanted to build a trebuchet, so a catapult.
Bill Nussey:
Oh, I love it.
Titiaan Palazzi:
And so we spent a day building a catapult with them, which we used at the end of the day to launch watermelons about a hundred feet far.
Bill Nussey:
That’s a serious trebuchet.
Titiaan Palazzi:
Serious one.
Bill Nussey:
Yeah, we’ve got a little toy one here at my house my kids built when they were younger. And you can throw a golf ball 50 feet, but the watermelon’s pretty impressive. So tell us about Myst AI. What motivated you? What problem did you and your co-founders see that you needed to go after and solve?
Titiaan Palazzi:
Yeah, we started Myst coming from two visions. I founded a company with Pieter Verhoeven, my co-founder, also from the Netherlands. He was an early engineer at Nest, the company that is famous for its smart thermostats and was sold to Google for $3 billion. Pieter worked at Nest on large scale machine learning, for times series data. What Pieter did at Nest was he helped to build things like the demand response algorithms that allowed Nest to pre-cool buildings before a grid peak occurred.
Titiaan Palazzi:
What Pieter had seen at Nest was two things. First, there is a real opportunity for the latest knowledge in machine learning and AI to make a big impact in how you predict the future. The second thing was that actually predicting the future is really hard, not just for the reasons that you would expect, but also when it comes to time series data, because you are drawing in continuous reams of information from millions to billions of different data locations, and training and updating models all the time. So from an infrastructural perspective, from a software infrastructure perspective, it’s a really tough challenge that essentially every company that needs to solve this has to undergo. And so he saw an opportunity to really do something here from a technological perspective.
Titiaan Palazzi:
My starting point for Myst was in my final two years at RMI, I worked with small utilities, mostly electric cooperatives, to help them understand how they could plug in more solar and battery storage into their grids. And I saw firsthand by working with many of these CEOs of 50- to 100-person electric organizations, that they were really struggling with the intelligence to run their grids in a reliable way, as all these new distributed energy resources were coming online. And I knew that, and I saw that if we could give them better insight, that would be foundational for making this transition happen, that we’re all trying to work on. So we joined forces. We said, “We’re going to build a technological solution to this foundational problem to realize a carbon-free electricity system.” And so it went.
Bill Nussey:
It’s an amazing company. And it’s fun watching you guys take off and add customers and build out capabilities. But one of the questions that I’m fascinated by is the origin of the name. For those folks listening to us, it’s Myst with a Y, which immediately to me conjures up that silly video game I played in the ’90s. So what’s the story behind the name?
Titiaan Palazzi:
It’s really funny. Neither of us played the game. The name came from mystic, somebody who can see into the future. And we wanted a short, snappy name, so we shortened it to Myst. The good thing is we’ve met lots of people who have mentioned the video game. And all of them have positive associations with it. So, I don’t think it has hurt us.
Bill Nussey:
Let’s understand a little bit more about what you guys are doing, in a very practical, rubber meets the road way.
Titiaan Palazzi:
Bill, absolutely. So what Myst really does is we help companies predict the near term future. Our clients are typically companies who have exposure to electricity markets. So they could be utilities, they could be renewable generators, solar and wind, or they could be battery storage owners. And they need a clear view of the near future, so accurate near term forecasts, for energy demand, renewable production or market prices, to reduce the risk that they’re exposed to high real-time prices in markets.
Titiaan Palazzi:
We serve a variety of different types of electricity companies. We help companies that are called load serving entities, so who serve electricity to end customers. But we also have a lot of clients who are large, renewable generators, solar and wind, or battery storage owners, increasingly, for the load serving entities or for the utilities. One example is how we’re helping California Community Choice agencies to reduce their exposure to real time-price spikes. We’re working, for example, with East Bay Community Energy. East Bay Community Energy is a Community Choice agency in California. They serve about a million homes and businesses in the East Bay. And when we started working together in 2020, we showed we could reduce their exposure to real-time price spikes by 70%, which was tremendous. And it helped them to avoid blackouts during high stress periods.
Bill Nussey:
How has accuracy improved with the various components you can tap into for forecasting, like weather data, for example, or customer consumption data? How does that improve with your system?
Titiaan Palazzi:
We really see the data providers as our closest friends and partners. Our take on business with Myst is a bit different than some other predictive companies. So really, what we try to do is we try to empower companies themselves to find the models and the data sources with the highest predictive performance. So whatever data source that is, our job is to get it in our client’s hands. And really, when we see our vision for the future of Myst or our product vision for Myst, is that we’re empowering clients to build their own predictive models in-house, ideally in a world where they just say, “Look, I just want to predict this thing, rooftop solar production, as well as I can. You guys tell me which model and data best lead to this prediction.”
Titiaan Palazzi:
And then on our side, it’s our job to build a product that, from a wide array of data partnerships that we’ve all plugged in into our product, we surface to our clients those that have the highest predictive performance for their specific use case. So we have data partnerships with companies that collect energy markets data, companies that provide weather predictions and historical weather data, companies that do irradiance forecasting, like Solcast, companies that predict the wind speed at the height of the wind turbine where the blades are spinning, or companies that predict the rainfall on hydro reservoirs to give a sense of how much is in the reservoir.
Bill Nussey:
You’re describing Myst, not just as sort of a computational company, but the center of an ecosystem. It’s growing. You’re creating a whole downstream set of company, or you’re enabling a whole downstream set of companies to do what Myst has done. This isn’t like writing enterprise software. It has different economics. It has different challenges in addition to just the AI dimension. So, super learning here. Thank you.
Titiaan Palazzi:
Yeah. And Bill, for your listeners who come from the software world, I would even say we see ourselves as a platform that is connected to a data marketplace. And so the cost of data really should be born by the users who find the benefit in that, only when the benefit is accruing to them.
Speaker 1:
How hot will it be today? Will it be cloudy? Did the EV charge last night? Who’s working from home today? I don’t know about you, but hearing Titiaan talk about the number of variables and the amount of data needed to provide an accurate forecast of supply and demand for electricity is a little intimidating. Take just the weather data alone. Did you know that IBM’s The Weather Company processes around 400 terabytes of data per day to enable its models to predict the weather, days in advance around the globe? That’s a lot of data. Now add in energy consumption data and data about electricity being generated from a variety of sources, and a lot of other factors like time of day and seasonal elements, and you end up with gobs of data.
Speaker 1:
As more and more intermittent renewable resources are added, the demand for big data and sophisticated AI models to more accurately predict supply and demand becomes paramount to maintaining our grid’s reliability and stability. What was done with spreadsheets not too many years ago is quite literally impossible to do today. Grid operators, consumers, and energy traders all benefit from these new AI platforms. And as we are learning from Titiaan today, the powerful predictive capabilities of an AI system like Myst saves money, lowers cost, improves reliability and asset management, and creates competitive value as we make the transition to more and more renewable energy resources. Now, let’s get back to Bill and Titian to learn even more about AI and Myst.
Bill Nussey:
I think AI is one of the most misused, widely abused, and generally misunderstood ideas. For I think a lot of people, you think of AI and you think about Terminator, some robot that’s going to take over humanity, or some sort of mystical, completely ridiculous manifestation. So let’s just start with the basics. You live in this world, you’re pioneering it for the energy tech space. What is AI? And what’s the state of it today? And where is it going?
Titiaan Palazzi:
Definitely. So in a way, you can make it very simple. What artificial intelligence or machine learning, these terms are sometimes used interchangeably, really means is learning from data, often large amounts of that, to make predictions about things that are not fully known. In our case, that means predicting time series or predicting the future. But it could also mean deciding whether images are a cat or dog.
Titiaan Palazzi:
What makes it different is that there are ways to do this same thing by setting up explicit rules. For example, if the image is orange, then it’s going to be a cat. And these kinds of systems learn implicitly, so there’s no explicit rule. One of our main investors is Gradient Ventures, which is a investment vehicle from Alphabet, from Google, that focuses exclusively in AI companies. And so we grew up as one of the first companies that they invested in, and they now have over a hundred companies that they’ve invested in that all focus on applying AI to business use cases. So it’s actually very interesting to see how this is driving value across industries, or how people are trying to create value by using this technology across industries.
Bill Nussey:
Yeah, I think at its heart, AI is pattern recognition. I think in the simplest way, I think about it is like a regression analysis, which says that there’s this association that’s buried within all the data. And what you guys are doing and where AI is most valuable is to say, “If this association has occurred in the past, if these conditions exist in the future, it’s likely that this will be the outcome.” And I love your rules metaphor because you might say a Myst that’s not interesting and cool version of you guys in some alternate universe would say, “Well, if it’s supposed to rain tomorrow, we’re going to need less electricity for air conditioning. It’s a rule. If no rain, less electricity predicted.” But what you guys are doing, and I think this is something to be really embraced, is that you’re looking at millions, I’m guessing hundreds of millions of data points, and applying some crazy complex algorithms to it.
Bill Nussey:
And so, yes, it’s supposed to rain tomorrow, but there’s numerous other factors that can affect how much electricity is used, like what’s the ambient temperature supposed to be? What was it today? What time of year is it? And you can multiply those inputs by hundreds, and you apply the data. And what the machine learning does is it finds patterns that no human could do. You couldn’t put it into a spreadsheet. You couldn’t eyeball … I like eyeballing things, right? I’m always proud of the fact that I can sort of eyeball something down to a quarter of an inch if I’m cutting wood. But you can’t eyeball this stuff. And that’s where the computers are so crucial. And that’s essentially what you … Did I get that right? Is that kind of the way you guys do it?
Titiaan Palazzi:
Yeah, Bill. It’s spot-on. And I think another reason where I see AI being powerful is to quickly adapt to changing conditions. So I’ll give you an example. When COVID-19 started in 2020 in the United States, what we saw was that suddenly people were starting to work from home and businesses were shut down. And so for companies who had been consuming energy for dozens of years and were using that as rules to decide how much they had to buy tomorrow, suddenly those rules no longer applied. And what we saw was that when AI-based systems are set up dynamically, within hours to days, these systems were adapting and could essentially see these new kinds of trends.
Bill Nussey:
So why is this hard? Can’t I just get a spreadsheet or get some library and plug, plug, plug, and the next thing you know, it’s cranking out answers? I mean, my Nest thermostat, thanks to your co-founder, it seems to be pretty accurate, but then sometimes it does dumb things, and I just want to hit it. But seriously though, why is this hard? Why isn’t everybody doing this everywhere? What advantages do you guys bring, and how do you maintain a competitive advantage as this market gets more competitive?
Titiaan Palazzi:
Yeah. Well, one reason why it’s hard is that as you pointed to, you’re trying to incorporate data from vast amounts of data sources. So if you’re doing this in an Excel sheet, there’s no way that you’re going to incorporate, say, weather data from 150 airport stations across the state of Texas, or that you could incorporate the energy demand data for every building in your city. So that’s what makes it really hard, which also speaks to a different thing, right, which is that the reason this is a problem where a tech solution can help is because it is something that is necessary at tremendous scale and reliability. It’s really not just about sort of the AI smarts. It’s also about building the underlying infrastructure.
Bill Nussey:
And can you explain that a little more? What do you mean, “the underlying infrastructure”?
Titiaan Palazzi:
What I mean is this, is when you are creating predictions, let’s say we’re trying to predict the production for each rooftop solar system in a certain utility area, you are trying to collect the historical production data from those individual solar systems, they’re often logged in the inverters, in order to train that model. And so now you need to collect data from possibly tens of thousands, hundreds of thousands, millions of individual locations into one system, so there’s all kinds of data querying challenges, and then feed that into another system. And there will be data that’s missing. And there will be times where the system fails. And to do all of that in a way reliably, maybe every five minutes or maybe even every minute, is pretty tough to build, from a technical perspective.
Bill Nussey:
That makes sense. That makes a lot of sense. And I’m imagining that part of the art of what you guys do is there’s certainly hundreds of thousands of potential data sources, and choosing the proper ones, testing it, seeing where the model improves and seeing where it doesn’t, and also the amount of computational intensity for this is pretty high. So unlike a lot of software companies, you guys are actually incurring real costs to run the computers. If you’re providing enterprise accounting, the amount of computing time you’re paying Google or Amazon for is relatively modest, but to run AI models, I’m guessing Google sends you a pretty big bill every month. Is that accurate?
Titiaan Palazzi:
Yeah. I mean, after our talented team, our biggest cost is our cloud computing. Yeah.
Bill Nussey:
So you have a big incentive to make your models more intelligent, not to include every single variable, even if it could work. So it’s an interesting era of the computer industry, where the computation is actually a real cost. Well, listen, this is so fun, so cool. So as we wrap up our conversation today, I would love to take you through our four lightning round questions. This is something we put upon all of our guests that we’re privileged to have come spend time with us. And the perspectives are really incredible. So let me just jump right in. Are you ready? Are you ready for this, Titiaan?
Titiaan Palazzi:
I am ready, Bill. Let’s go.
Bill Nussey:
So, what excites you most about being in the clean energy business?
Titiaan Palazzi:
What I love is the clear mission. Everybody’s focused on the same end goal. I was at a summit on Friday with probably 100 founders of climate tech, clean energy companies. And it was so collaborative, and everybody is there with the same end in mind. We also see it with talent. It’s an amazing time. So many people from tremendous backgrounds are trying to make a meaningful impact. So that’s what gets me really excited. When I get out of the bed in the morning, the thing that gets me most excited is the people I get to work with and the clear focus of what we’re trying to solve.
Bill Nussey:
I did an interview couple weeks ago and someone said, “What’s the biggest difference to you, Bill, between being in the tech industry and being in the climate tech industry?”
Bill Nussey:
And I said, “For 25 years of my career, I’d meet two, three, four times a year someone that really inspired me. And in this industry, I meet someone once … two, three times a month.” And you’re right, some of the most talented people in the world have many options. And they’d rather work for something that has a mission, like Myst, or some of the other energy and climate-related businesses and organizations. So this is an exciting time. It’s not just hopefully make some money, but it’s also a chance to make a real difference in a way that’s tangible and will last for generations. So, second question, if you could wave a magic wand, a single wish that could change just one thing, what would it be?
Titiaan Palazzi:
We should go faster. So I think where we’re now is really … it’s a matter of speed. It’s not if, it’s when. If I could change one thing, I think it would be to change the minds of all the legacy leaders in charge of coal or oil or gas companies, and completely shift, from zero to one, the way that they’re allocating investments, because what we need to do is we need to turn over all our capital. We need to change the way we generate electricity, the way our transportation works, the way we are consuming energy. And that’s the one wish I would have.
Bill Nussey:
All right.
Titiaan Palazzi:
I’d also say one thing that excites me, similar to local energy, is the fact that what’s cool about this transition is that decision-making is distributed. So we can all vote with our dollars for what car we all drive, how our home is powered, whether we’re cooking on gas or on electricity. And that could mean that this really goes fast.
Bill Nussey:
That’s a great perspective. I do talk about local energy from time to time, and that is a perspective I haven’t quite heard before. But every decision we make about our car or our house is a vote, and a very tangible vote. That’s a great point. We’re going to run with that later on. I like that. Thank you for sharing it. So, third question, what do you think will be the single most important change in how we generate, store, and distribute electricity in the next five years?
Titiaan Palazzi:
Yeah, IPCC put out a report earlier this spring, at the end of the winter. And they had a table that showed, for the different ways in which we can change our climate impact, the expected gigatons per year that were available between now and 2030. And the table was really a wake up for me, because it showed that solar and wind are really the things that we must scale between now and 2030. I think they accounted for more than half of the economic emission potential in the next eight years or so. It was about eight gigatons. And for your listeners, I think our goal is to reduce today’s 60 gigatons of emissions a year to zero by 2050. So the biggest thing we can do is to accelerate adoption of renewables, solar and wind, and battery storage to support them
Bill Nussey:
Well, given that you’re in the AI prediction business, I think your answer to that particular question should be the one we go with, because you’ve got the giant oracle machine sitting there telling you the future. So last question, when people come to you and say, “Titiaan, I’m really excited about renewable energy and the future. What can I do personally to get involved,” how do you answer the question?
Titiaan Palazzi:
I’d say, see if you can work on climate, see if you can influence policymaking at a meaningful level, so get involved in your state. You’d be surprised how few people show up to state councils, whether it’s a public utilities commission or something else. If neither of those are options, I’d say influence the organization or community you are working in. I’ll give you a cool example. My younger sister works with a big tea and coffee producer. And she has no background in engineering or clean energy or finance, but she was passionate. And she convinced her CEO to install rooftop solar on all their commercial facilities.
Bill Nussey:
I love it.
Titiaan Palazzi:
So that’s an example of one person, just because they’re passionate, influencing the course of an organization.
Bill Nussey:
Well, listen, we are at the end of our time. I love this conversation. I love what you guys are doing. It’s so exciting. Thank you for all that you’ve done for me and for the industry. You are one of those true spirits that you have a call to help other people. And now you’re doing it in this way that’s making a big difference in the economics of energy. And I can’t imagine how many people have been [inaudible 00:30:37] fortunate as I, have benefited from your advice. And it’s been great to have you here today and to share your story. And thank you for all that you’re doing.
Titiaan Palazzi:
Thank you, Bill.