The Role of Innovation in the Energy Transition

Denver Yu (00:00): 

The customers that you'll be selling to in the energy world are thinking very long term. For better or worse, that's just how many of their jobs are set up. And so I think you have to plan your product roadmap in that similar vein. And sometimes that means even if you want to work at your pace, your startup pace, your customers might not work at that same pace. 

Michael Torrance (00:24): 

Welcome to Sustainability Leaders. I'm Michael Torrance, chief Sustainability Officer at BMO. On this show, we will talk with leading sustainability practitioners from the corporate investor, academic and NGO communities to explore how this rapidly evolving field of sustainability is impacting global investment business practices and our world. 

Disclosure (00:45): 

The views expressed here are those of the participants and not those of Bank of Montreal, its affiliates or subsidiaries. 

Michael Torrance (00:57): 

Fatih Birol, Executive Director of the International Energy Agency said "the transition to clean energy is happening worldwide and it's unstoppable. It's not a question of if it's just a matter of how soon and the sooner the better for all of us." Today I'm joined by a panel of four experts who are at the forefront of technological innovation related to the energy transition. Today in our podcast, we'll look at the big picture challenges of shifting to clean energy sources and how a combination of technology, new thinking and hard work are addressing those challenges. So let's get started. I'd like to introduce Denver Yu, an analyst with Generational P artners, an early stage venture capital firm. Joining him today are the leaders of three of generational partners portfolio companies, Kevin Kong, who is founder and CEO of Everstar. Mariah Frances Carter, who is CEO of Lumora Technologies, Josh Riedy, who is founder and CEO of a company called Thread. So welcome everybody to the show. It's great to have you on our podcast. Denver, let me start with you. Just for the benefit of our listeners, can you introduce yourself and also generational partners talking a little bit about your investment thesis on the topic of energy transition? 

Denver Yu (02:12): 

Yeah, definitely Michael, thank you for the introduction. We are a venture firm based here in LA. We focus on pre-seed investments in what we call industrial markets. So you can think of aerospace, manufacturing and obviously energy, which is what brings us here today. So our background is actually primarily in aerospace and defense and part of our genesis of starting the firm was tied very closely to a lot of the talent that we were noticing coming out of aerospace and defense sectors into working on companies related to energy, whether it be new batteries, whether it be new software that could optimize, whether it be new designs for utility companies. And we wanted to start capturing some of that. So I think a good example is actually one of our founders here, Mariah, her and her co-founder Jesse left Palantir to start a company to help utility owners better optimize their revenue based on their electricity trading patterns. 

(03:11): 

And I think a lot of the culture that comes from aerospace and defense can be applied to the energy sector here. That's definitely unique to our investment thesis and something that I think a lot of the industry is starting to appreciate, namely because I think the way that these founders bring a level of urgency and responsibility to a lot of the product building, but also the way that they interact with customers, the feedback loops are much closer, whatever the customers are really hoping to get out of Something that another one of our companies here Thread and Josh has done is listen to these customers closely as a co-development partner. They're based in the Midwest and they've basically become a partner to Xcel Energy, one of the major utilities out there and have built a product that scales from wind turbines to distribution lines to everything that also affects wildfires here today. I'll let him give the proper introduction, but I think that startups have a tremendous role to build product faster than what some of the customers themselves would be able to do. 

Michael Torrance (04:17): 

Thanks Denver. That's a great introduction. Maybe just to recap then, if you could introduce Kevin, Mariah and Josh just with a brief overview of their companies and their products and then I'll come to you, Kevin, with a couple of questions about the work you're doing. 

Denver Yu (04:32): 

Yeah, definitely. I'll give a quick introduction to our lineup here. We've got Kevin here at Everstar. He's building a nuclear compliance platform for the industry. Everything from workflows to regulatory affairs, bringing really the modern AI best in class practices to folks in the NRC folks at utility companies, giving them the best tools that they need to go and file everything that needs to be done. Traditionally with consultants or lawyers, these tasks can cost millions of dollars for a simple mistake and he's bringing the best tools for these end users. We've also got Mariah here at Lumora, her and her co-founder Jesse, they are building a software that can optimize the revenue of electricity trading for battery storage owners. And in the long term, they can also apply this for other types of renewable sources. So think wind turbines, think solar operators and across the entire stack in the us. Last but not least, we have Josh Riedy over at Thread. They were actually our first investment in the portfolio. We've known him for a long time, stretching back from our Starburst days. He is building a product that can help utility companies inspect, collect, and store all of their data in the cloud to give utility owners a digital health record of all their assets. I'll let him get into the evolution of his product and his company, but it's a fascinating story. 

Michael Torrance (06:09): 

Alright, well Kevin, let's start with you then. Nuclear power's been in the news a lot lately with large tech companies turning to nuclear to power AI data centers and it's really a great solution for Baseload power in the energy transition, but of course there are very serious and real safety concerns that have to be adequately managed. Can you tell us what was your thinking and how did you come to found your company and using your startup to help solve some of these challenges? 

Kevin Kong (06:40): 

Yeah, absolutely. So appreciate the intro and the setup, Michael. Everstar is a company on a mission to engineer energy abundance. This is born out of my 10 years being obsessed over this concept. I've tried to tackle this problem in numerous other ways on a day-to-day basis. I was head of engineering at Bird, started a enterprise AI company as CTO and co-founder. Most recently in my advisory investing, working very closely with startups kind of role, I've validated deeply technologies like nuclear fusion to ground solar space, solar high density batteries, all leading to the very clear realization that nothing is going to move on the timelines that we need them to achieve. Energy abundance quite like nuclear fiss. So it is the key to this unlock, but it also comes with a lot of bottlenecks that make it not so ready to deploy tomorrow. So the realization I came to after a lot of deep exploration and talking to industry experts is that in this country the paperwork is more expensive and harder to do than the physics of nuclear fission. 

(07:53): 

And with that realization, we started this company to leverage the power of AI to accelerate the licensing process. It is the critical step in the broader energy transition. Now, nuclear energy is the cornerstone, as you mentioned, to achieving carbon neutrality. It's got immense power density specific energy wise, it is 1.5 million times more energy dense than natural gas. It has a much higher capacity factor than even natural gas. And so 90% compared to 50. So in terms of what can provide reliable clean energy at scale, nothing comes close to nuclear fission energy. And now there's a lot of industry hurdles in the regulatory front that can delay or even derail these projects. It has historically taken about $400 million to about $500 million for the last few reactor approvals over 10 years. So if you want reactors now, it's going to take 10 years to evaluate the safety of these reactors for site evaluation. 

(08:58): 

It also takes about a year each to get each one approved. And so with the power hungry needs of AI data centers and very power hungry industrial needs, there's no time for business as usual. We need to radically rethink the cost structure and the timelines with which these reactors and sites have been approved for deployment. So with this mandate, we've built this platform Gordian so we can streamline the compliance and licensing process, reduce the timeframes, lowering costs, and ensure that the right to requirements can be met more efficiently. Our goal, our mission here is to disrupt nuclear power economics to be much more predictable and investible. That's the key, so that we can contribute directly to the shift towards a more sustainable energy landscape and we're working as quickly as we can to achieve this mission. It has been no small feat to try to replicate the human expertise needed to file these 400 page, 4,000 page documents. But that's the essence of our technical challenge at. 

Michael Torrance (10:02): 

And some of the challenges are beyond the regulatory side, but also about public perceptions. Does Gordian have any role to play in helping to overcome that kind of challenge to the future of nuclear energy? 

Kevin Kong (10:14): 

Absolutely. So Gordian addresses these head on by automating and optimizing the regulatory processes. We're working very closely with industry experts both on the regulator side and private practitioners who've been around this space for a very long time so that we can map the complexity and introduce a lot of clarity to the system. So we're in the process of mapping out all the different pathways to approvals. There's no single pathway, there's multiple different interlaced regulatory regimes give you different optionalities for markets that you can target, different timelines. There are things you could do in parallel to streamline these processes and to allow direct and more proactive communication with the regulator in an open, transparent way. Those features are all baked into the platform. So we're trying to provide predictive insights, automating the documentation, which takes weeks, months to produce each iteration of, and we're ensuring that all the submissions meet the stringent standards set by NRC, nuclear Regulatory Commission, department of Energy or Department of Defense. 

(11:19): 

So improving the efficiency and accuracy of these processes like that is the key thing that nuclear operators really need from a platform like this to bring confidence and clarity and by demonstrating compliance and safety more transparent and effectively through the platform, this all adds back to enhancing public trust in the process. Our goal is not to cut regulations nilly. AI must be leveraged so that we marry economics with building on the safety standards that have been set by the NRC over the last three. So we believe that there's a happy marriage to be made somewhere in the middle through application of AI 

Michael Torrance (12:00): 

Now long permitting processes and slow approvals, it's affecting really the entire value chain, and it's not just nuclear. It's affecting everything from mining of critical minerals to different kinds of renewable power. Do you see your product having application in different contexts beyond nuclear? 

Kevin Kong (12:18): 

Definitely. So our initial focus has been on nuclear. I like to tell folks we're starting our startup journey in hard mode. We probably have the highest safety documentation, stringent standards around those requirements, but the underlying technology we are building to solve these challenges has much broader applications beyond just nuclear to other various industries, adjacent ones as well. So renewables grid infrastructure, as you mentioned, mining minerals, even petrochemicals and traditional energy industries face a lot of similar, not the same, but similar regulatory and compliance challenges, NEPA compliance for site safety, OSHA for worker protections, these things are, this is why American assets and infrastructure is regarded as the safest places to work in the best standards. So we believe that to take these industries to the 22nd century, you need to adapt this permitting processes, safety, compliance, procedures, operational efficiency, those all need to be adapted to the AI first world to make sure we don't miss a beat on the things that we hold dear, like worker safety, environmental production, but you also start to chart towards globally competitive economics. 

Michael Torrance (13:36): 

Thanks a lot, Kevin. That's really great insights you're sharing. Mariah, let's turn to you. You're the CEO of Lumora Technologies and you're leading a startup that's founded by engineers and it's focused on energy storage. The platform, as I understand it, it helps energy operators decide when to charge or sell energy and it uses advanced data analytics to do that. How is this process done today and what are some of the benefits that this kind of data integration software can bring to the emerging new energy sector? 

Mariah Frances Carter (14:11): 

Yeah, for sure. Thanks again for having me on here. It's really awesome to be joined by Kevin and Josh, two folks doing other really interesting things with technology in the energy sector. So real pleasure and honor to be here. To answer your question, Mike, I think the first thing it's very important to understand about the way that data is utilized by folks, not just in the battery storage space, but sort of energy traders more broadly, is that the data is all stored in really disparate sources. And so today, where most people find the operational challenges emerge is really in getting everything in one place. And then secondly, it's just stored in lots and lots of different places. So you'll have one dashboard for all of your realtime energy data like prices and market information. Then you might have another data source where you're getting your weather forecasts to help you understand things like generation profiles and expected output. 

(14:57): 

Then you might have another thing that's tracking outages for lines or generator facilities and maybe even another one that's tracking your own internal resources on how performant your assets are. And so imagine you're trying to make decisions in real time or even for the day ahead market and you're looking at half a dozen or a dozen different data sources across all those different systems. It's really challenging to figure out what the right course of action is from a market perspective, especially if something unexpected happens like an outage being the primary example or perhaps an extreme weather event, which we're seeing more and more in markets where things are being disrupted. Texas being the primary case for that, and it's really hard to organize all that stuff. So what we're trying to do basically is bring it all into one place so that folks who are trading assets like batteries, but not only batteries, have that information at their fingertips to be able to respond and adapt to current market conditions. 

Michael Torrance (15:46): 

Thanks, Mariah. When I was at Climate Week in New York last year, I was learning a lot about smart grid technology and the idea of Watts that basically if you can be super efficient with your use of energy, you can actually reduce demands for energy across the grid. So I appreciate how important this kind of technology can be. What would you, in your view be the ways that utilities can begin to incentivize the investment in smart grid approaches work that you're doing and others are doing that can actually lower demand perhaps or make more efficient use of the energy that we have? 

Mariah Frances Carter (16:25): 

Yeah, definitely. It's a really good question and I think there's sort of two primary pathways that I have seen that I think are really promising. The first one is sort of figuring out how this kind of advanced software fits into the utility cost model. So the way that utilities make money today, sort of spending money to make money and being approved for the amount that they're going to spend on capital investments and the return on those investments by their respective public utility commissions historically hasn't always included software that tends to be reserved for things that are more infrastructure intensive, like transmission lines and things that when you look at the cost tend to start with a B for billion. And so when we think about the ways that this technology is going to save money and sort of as you said, rearrange demand or rearrange the actual costs that the utilities incur in order to realize the investment return that they're looking for, I think it'll be very interesting to see over the next couple of years if utilities can figure out how to make the case that technologies that make things more efficient are and should be thought of the same type of technologies that should be invested in from a capital expenditure perspective. 

(17:28): 

The other method is just very straightforward making money right here. Right now, batteries are being built all across the country. In some places the operators of those batteries are the developers themselves, and in a lot of other cases, those batteries are being contracted with the utilities in the jurisdiction or the territory that they're in. As off-takers, these are sometimes referred as off-take agreements or tolling agreements. So if the operator isn't the person who developed a battery, somebody is still making decisions about when to charge and discharge those assets. And so if that is the utility, then there's a huge opportunity to realize gains from arbitrage, perhaps delay transmission or distribution investments that need to be made if depending on where those batteries are located. But all of that hinges on the utility or the operator for that matter, being able to discern when and where it makes sense to have those batteries and to discharge them. So there's sort of two primary pathways. It's how can we get software into the capital expenditure sort of framework that utilities use to annualize all of their costs, and then how can we just simply help utilities make money off these assets because of the really dynamic nature of batteries in their ability to arbitrage power across time and across space depending on where they're located. 

Michael Torrance (18:40): 

Thanks Mariah. Data is really key to your business. How do you overcome the challenges of different quality or formats or sources of data across a national power grid? And I guess related to that, is there any policy changes or tools that policymakers should think of that would help to address some of those challenges? 

Mariah Frances Carter (19:03): 

Absolutely. Boy, lots to unpack there. Data is really messy for organizations who have been around for many decades utility during this category. The military is also in this category, and one of the things that my co-founder and I spent a lot of time doing when we were both at Palantir was helping different branches of the military unwind those really old antiquated data systems that have a million different formats, a million different standards, maybe some of them are stored in a dusty old server that's in a closet somewhere that only one guy who's 65 and is going to retire next year knows how to operate. One of my co-founder's favorite anecdotes is a system that she worked on for the Air Force was so old and antiquated that she literally had to make a phone call to the manager of that system every three days to get him to reset it so that it would start uploading data again. 

(19:48): 

And these are systems that are used for very critical parts of military workflows. This isn't just sort of optional data. And the reason I share that anecdote is just to say that there are a lot of challenges in both data access and data quality that need to be addressed separately. Bringing organizations into the 21st century, which people love to throw around in the utility sector, that is hard work. And how do you do that? You roll up your sleeves, you sit next to 'em at their desk and you show them how to use the systems step-by-step until they are comfortable with it. And so I think that the quality issues that I have seen and the lack of standardization issue is absolutely something that can be addressed by policymakers. It's really funny you mentioned that because I actually just wrote a chapter in a book about this with a professor over at Columbia called David Sandal who's basically building a climate change AI mitigation roadmap for the Japanese government that's sort of globally focused. 

(20:39): 

And one of the things that's interesting about that is that it's this issue is present all over the world, everywhere in the world where people are trying to do interesting things with smart grids, they are finding that the quality issues are stemming from a lack of standards. And so if you can't have data standards, you can't have quality data, you can't do any of the fancy AI predictive use cases that folks really want to do, it's very linear in that sense. And so yes, policy makers absolutely could enforce standards. They could absolutely encourage ISOs and RTOs and other regulatory organizations to come together and decide, okay, this is how we're going to do standards. 

Michael Torrance (21:11): 

Yeah, and I mean it's the role of policymaker to solve coordination problems. I mean, that's really what it would be about. It's not really about enforcement. It's actually trying to facilitate this kind of innovation to take place. That's really fascinating discussion. Thank you, Mariah. Josh, let's talk about thread your company. It's focused on a really essential part of the rollout of renewable energy, the distribution side, and we all know power lines are all over the crisscrossing the country. If one of those lines fails, it's going to be a huge problem in terms of the availability of electricity. There's also safety issues and wildfires and other things that can be associated with those networks. How are those lines typically being inspected today and how often and can you tell us about the role that you foresee thread playing in helping to better monitor those distribution networks? 

Josh Riedy (22:07): 

Happy to do so, and thanks for having you today. It's a pleasure listening to my colleagues. I've had the opportunity to meet both and there are tremendously talented. I would answer your question by starting with what does the world today look like? And in doing so, I'd give a huge shout out to frontline workers that maintain our nation's critical infrastructure, perhaps an overlooked part of the workforce, a many dedicated lifelong servants maintaining the physical assets. And I would expand your question to talk about all of the physical assets that constitute our nation's critical infrastructure because the conversation today has been about regulation. It has been about data and sensor outputs. What thread seeks to do is to digitize physical assets in a way that decisions could be made from the normalization of that information. And that starts by empowering frontline workers to give them new tooling that allows for the digitization of those assets. 

(23:10): 

That could be a cell phone, that could be manned aviation, that could be unmanned aviation, vehicle mounted sensors. But the reality is utilities must gain a better understanding of the physical conditions of their assets so that they can make decisions in critical instances such as wildfire mitigation. It is an overlooked part of the industry. I would say that utilities are very skilled at monitoring the flow of energy, yet the physical condition of those assets. In the case of power distribution, which can have a 70 to 80 year lifespan, it is largely tribal knowledge. And so the analogy I often use for thread is electronic medical records. There was a day that your health, if the physician that you were seeing or the file cabinet that your medical records were stored in were absent, meaning you move somewhere else as an example, that knowledge of your health was largely lost. 

(24:10): 

It wasn't until companies like Epic began to digitize those records that you had transferability portability of your records, but also a holistic understanding of all of the aspects that made up running a healthcare facility and organization. We view a utility similar to that, that once knowledge is gained regarding physical assets, and what I mean by that is not just the collection of imagery and the storage of that imagery, but the interpretation of that imagery so that it itself can become a data point that is normalized once that happens and that information, whether it be imagery or AI analysis or interactions between frontline workers and engineers, when that information is stored in a similar manner and it's retrievable, better decisions can be made and more timely decisions can be made, especially when the stakes are at their highest, which is in the unfortunate instances of natural disaster, be it hurricanes, be it blizzards or be it wildfires as the nation is painfully aware of or was painfully reminded of again recently. 

Michael Torrance (25:23): 

That's fascinating to see. The broader application of this one area that we've been working on around physical hazard analysis has been the use of geospatial data and satellites and that sort of thing. How do you see the types of data collection that you're doing and data manipulation interacting with those other sources of data that could be acquired about physical assets? 

Josh Riedy (25:46): 

Great question. I don't think there can be enough data to solve these problems. For me, it's a matter of using the right tool for the job, the right level of detail to understand what you must regarding the physical condition of those assets. And a piece that I would focus on that makes thread unique from the industry is self enablement. You must enable these frontline workers to provide the tooling themselves and to utilize that tooling. That goes back to Mariah's mention of capitalization. Capitalization is much easier obtained if indeed you can perform the work yourself to a degree, especially when it becomes critical to creating a digital representation of the physical world. That self enablement is a pathway to being able to not only create a data model that is useful, but to maintain that data regardless of the different layers that it may be and those layers, to be more specific to your question that we most generally see is satellite imagery. It is manda, unmanned aviation, and then it is foot patrol and what you would see cell phones used for or still photography or vehicle based sensors. 

Michael Torrance (27:04): 

Alright, so then what role do drones play in your data product in terms of streamlining the inspection process and what are you seeing in terms of uptake by industry and the use of drones for these kinds of inspection processes? 

Josh Riedy (27:19): 

I'll go back to the electronic medical record analogy. Drones would be similar to the MRI machine or an x-ray and the ability for frontline workers to use those drones on an as needed basis, the ability to mix in third party service providers that can use drones to inspect assets at scale. Those are both useful inputs to understanding the physical condition of assets. But I would say the piece that perhaps is most impactful is the ability to unify those data sets, the ability to perform a simple tasks such as inventorying at scale. There's a lot of missing pieces that need to be built between the intent to collect data or the purpose, the ability to collect data regardless of type of collection, again, be it satellite or be a drone based, and then the ability to bring those into a common system with a common vernacular to extract data from that. 

(28:19): 

Once that entire process can be normalized, you can begin to use data productively, timely, and at scale. I don't think our company thread could exist without a co-development relationship with a major utility. In our case, we are fortunate enough to have that with Xcel Energy, but what that enables us to do is to go deep inside the utility and understand their business from the ground up and all of the layers that it has. And so it's with that answer that I say, drones are a necessary component, but they're simply a piece of hardware that layers into a much larger solution. 

Michael Torrance (28:57): 

Thanks very much, Josh. Denver, I'll throw it back to you. To wrap up, you've introduced us to three really interesting innovative entrepreneurs, and they're entrepreneurial ventures really in startup mode, which is an interesting phase that would be a challenge for anybody to take these opportunities to market. What advice can you give to entrepreneurs that are focused on energy innovation when it comes to making ideas like this commercial ready? 

Denver Yu (29:25): 

I mean, I think this is a little bit of a cop out, but I think most important is realizing that talking to venture investors, just one part of this very, very small equation right now, we're sitting in a room of three founders, a bank and a vc, and I think we're just very one small portion of all the stakeholders that are involved to make a company succeed in this industry. I think a great example is the way that Kevin has approached countless customers. I can tell you that the number of engineers and end users in the NRC and all of these little nitty gritty details to nail down what does the product that the end customer want, it's so important because we all know that sometimes what the customer says to you and what they actually want can be very different. Another thing that comes to mind is sometimes you don't have to reinvent the wheel. 

(30:14): 

I think we've seen enough lessons learned from the past two or three Cleantech bubbles depending on who you ask in OA in the mid two thousands, but there's been so many ideas that on paper have been tremendous, but the execution involved is quite difficult, so I think about how to de-risk your business plan through the founders around you and past entrepreneurs who have walked that road. I think that sometimes a business plan can be as simple as demonstrated by leura is how do we help our customer make more money and how do we give them more time to do what they're good at and what they love? I'd say no need to reinvent the wheel. I'll round out with another idea that Josh inspires us is to think really long-term. He's one of the most gritty resilient founders that I've ever had the pleasure of working with can tell me how many punches this guy can take and keep rolling. 

(31:06): 

I think that's an inspiration too, to all of us entrepreneurs here, but people that might think of starting a company one day, the customers that you'll be selling to in the energy world are thinking very long-term, for better or worse, that's just how many of their jobs are set up. And so I think you have to plan your product roadmap in that similar vein, and sometimes that means even if you want to work at your pace, your startup pace, your customers might not work at that same pace, and so how do you keep the ball rolling and pursuing a North star that aren't just vanity metrics? I think that's really important. 

Michael Torrance (31:37): 

Thanks, Denver. That's a great wrap up, and thank you very much to you, Kevin, Mariah, and Josh for such an interesting conversation and for joining us today. Have a good one. Thanks for listening to Sustainability Leaders. This podcast is presented by BMO. You can find our show on Apple Podcasts, Spotify, or your favorite podcast player. Press the follow button. If you want to get notified when new episodes are published. We value your input, so please leave a rating review and any feedback that you might have or visit us at bmo.com/sustainability leaders. Our show and resources are produced with support from BMO's Marketing Team and Puddle Creative. Until next time, thanks for listening and have a great week. 

Disclosure (32:29): 

For BMO disclosures, please visit bm.com/podcast/disclaimer.