Listen: How an AI-powered robot is speeding up solar installation

View Full Transcript

Taylor Kuykendall:

Welcome to Energy Evolution, the S&P Global Commodity Insights podcast all about the energy transition. I'm your host, Taylor Kuykendall. A lot of our episodes this year understandably have been about how artificial intelligence and robotics are changing the pace and direction of the global hunt for greener sources of energy, but today I just wanted to bring you a quick conversation with Deise Yumi Asami, the founder of AES's Maximo. AES is a large global energy company and Maximo is the company making a robot that works alongside construction teams to install solar panels in about half the time and cost of more traditional deployment techniques.

The goal of Maximo is to make solar installation more safe and scalable. Basically, it does the heavy lifting for workers when it comes time to attach these solar modules. Now, they've already installed nearly 10 megawatts of solar and they're shooting for 100 megawatts by 2025. AES is planning to use Maximo to help build up to five gigawatts of its solar backlog and pipeline over the next three years. Maximo is also going to be used in the construction of the two gigawatt Bellefield project in Kern County, California, which is the biggest solar plus storage project in the United States, and it's already got a contract with Amazon. Now, I spoke with Deise about the technology and how AI and robotics are changing the energy transition landscape. It was a really interesting conversation, so let's just dive right in.

Deise Yumi Asami:

I'm Deise Yumi Asami, founder of Maximo and AES Next. Maximo is the only AI robotic solution that does the complete mechanical installation of solar modules and uses advanced computer vision technology and generative AI to see its surroundings, solar parts, and it learns with every single installation of a solar module. Its key features ranges from both hardware where the Maximo system today is sturdier, it's faster, and really easy to operate, it can run for extended shifts, and it goes all the way to the digital capabilities and its brain. We have really been leveraging the new generation of generative AI and advanced computer vision that helps us to do the precise installation of solar modules.

Taylor Kuykendall:

Yeah. Tell me a little about the inspiration for Maximo. Maybe give us kind of the history of how this robot came to be and I guess the idea behind how we can go out there and completely automate the process of installing solar.

Deise Yumi Asami:

Well, taking a step back, AES is one of the largest renewable companies in the world. We lead the development of new technologies and approaches that drive the energy transition forward faster and at scale. The inspiration behind Maximo really comes from the urgent need to accelerate the future of solar. We understood very early on that solar demand would only continue to grow, and it was very clear for us that we would have to do things quite differently and more efficiently. For you to have an idea, at a solar construction site 70% of all the activities are highly repetitive, and out of the total 50% is solely dedicated on the installation of solar modules, so we really identified that this was a great opportunity for us to leverage new technologies and advanced technologies out there, such as new automation, robotics, and AI, to help us accelerate those repetitive activities on site.

Taylor Kuykendall:

Yeah. I want to come back to something you mentioned there a little bit ago. Tell us more about some of these AI functions and what it's doing on driving this robot. You mentioned that it learns with each installation. Can you tell me a little more what you mean by that?

Deise Yumi Asami:

Exactly. We are leveraging both AI, artificial intelligence, as well as generative AI. On the AI part, we utilize at a very macro level from a perspective of identifying the parts that involve the installation of the modules. The AI helps us to identify this is a solar module, this is a clamp, this is a torque tube, and really helps us understand how the robot is positioned relative to those parts. Then we are also leveraging generative AI where when you are in a controlled environment in a regular factory, it's pretty predictable, you always have a specific coordinate in which you pick up parts and you place parts in the exact location, but when you are outdoors, which is our case in the solar construction site, you really don't have that. Each time the module is going to be in a different position, each installation is going to be in a different place, and because you are in a uncontrolled environment, you have, for example, the effects of different light conditions. One of the things that happens with us on a solar site is the glare that is caused by the sun. Generative AI helps us to fill in the gaps that it's missing from what the robot is seeing, and it helps us to really know precisely where we need to pick the module and where we need to place it.

Taylor Kuykendall:

Maximo is not just an idea or a prototype right now, it's already installed some solar capacity, right? Can you tell me a little bit about just how much it's done and what you're kind of hoping to do here in the near term with Maximo?

Deise Yumi Asami:

Oh, yes. It has been a multi-year development. We did had very early on focused on the hardest things that we had to prove, as I mentioned before, taking the robotic systems from indoors to be utilized outdoors, and then today we are already with our second generation of Maximo and we have nearly 10 megawatts already installed with both machines that we have running.

Taylor Kuykendall:

Excellent. I'm wondering as we think about this, it seems like this could be transformative for the renewable energy industry. I'm just wondering, tell me a little bit about what your vision for how AI and robotics fits into renewables broadly, but maybe specifically solar as well.

Deise Yumi Asami:

Sure. For AI in particular, we really need a lot of computing for training and inferencing. As AI models get more complex from not only text but different mediums, like more images, spoken words, there's a lot more data and more processing that's going to be needed. At the same time, to keep up with these growing models and data processors are getting more dense and there's a lot more of it, so we will need more computing and we'll need more energy, and of course we all want these processors to be all powered by clean energy. Many future energy scenarios really shows that solar will be central to powering the future of electrification, and we do expect solar to support this growing demand.

Vincent Valk:

If you're enjoying this Commodity Insights episode, we think you'll really enjoy our Chemical Week podcast as well. Twice a month, the Chemical Week podcast dives into important issues impacting the global chemical sector. I'm Vincent Valk, one of the hosts from Chemical Week and Commodity Insights, and we bring you incisive and thoughtful conversations with executives and industry experts about petrochemical markets, sustainability, innovation, and lots more. Subscribe on your favorite platform or visit chemweek.com/podcast.

Taylor Kuykendall:

We're talking here about how this robot uses AI and how it uses data centers to operate, but this is going on in a lot of different industries. Can you maybe tell me a little bit about the amount of demand that's creating for electricity and why that maybe makes it even more crucial to be able to install solar quickly and cost effectively?

Deise Yumi Asami:

Well, if you think about just from comparison, just last year we had roughly 25 gigawatts installed. This really in this case translates to 150,000 modules that it's installed every single day. With this growing demand that I just mentioned of AI models getting more complex and requiring more energy to power all the processing needs in the future not only from Maximo in itself, but more generically speaking, that's what's happening across the globe with the more and more use of AI and more complex mediums of processing data, images, and audios, this will really require a lot more energy. Annual additions will continue to grow. By 2035, annual additions are projected to triple, so you can only imagine that 150,000 modules are now going to be very close to half a million modules installed every single day. For solar, it's very important that we keep up with that demand, with that pace, and with the growth and the demand that the world is requiring.

Taylor Kuykendall:

Backing back up to Maximo specifically, I wondering can you share any maybe numbers or data around exactly what kind of costs and time savings we're seeing from using this versus just normal human installation?

Deise Yumi Asami:

Maximo is really a tool that helps us to install solar modules faster. Our target has always been, and we've been proving that and already achieving the performance, of installing solar modules at half of the cost, half of the time. When we mean half of the time, is really the amount of time that it takes to complete the mechanical installation of the modules. We are doing that faster and consistently every single time, so we cut the time in half of what it would normally take to install those modules with the same amount of people. It's really making the construction teams on site more efficient.

Taylor Kuykendall:

Well, so it's interesting because you did say that it's the same amount of people. I'm wondering what's been the feedback from maybe some of the construction teams and the professionals around using this. Are they saying this is making their job easier, but also as they look at some of this happening should they be worried about their jobs in the future if there's more robots out there doing some of this work?

Deise Yumi Asami:

Yes, there's a lot of concerns when we bring robotics. I think people in general think directly about jobs. As I mentioned before, it's the same amount of people. It's really just helping the crew on site to be more efficient. A comparison that we typically use is the same way that years ago you would have people with shovels doing trenches on solar sites, today you have excavators. Maximo is really a tool there that it's helping the crew on site to be more efficient and have the consistency of installation and, quite frankly, also improving the working environment where instead of them having to lift those 70 to eight pounds over and over again hundreds of times per day, they are now utilizing Maximo and just installing those modules at a click of a button, but it's still the same amount of people.

The reaction that we have received from the industry is, one, we've heard from a financial institution as a rough reference that every 10 incidents on site, five of them could be related to or most likely are related to modules. This is not internal data, but this is external. It really shows that there's plenty of opportunity for us to improve that, to make sure that the installation of modules are going to be smooth as possible. From a Maximo crew perspective, yes, from an inclusivity perspective, we have an opportunity to also help people with different backgrounds and at different stages of their lives to also be able to continue to install solar modules when they can't actually keep lifting those 70 to 80 pounds modules every time, but with Maximo they can just click the button and continue to install modules.

Taylor Kuykendall:

Excellent. Obviously a really cool piece of technology and I'm sure you guys are very proud of it, but I'm sure you're also thinking about how to improve it in the future so what's the future of Maximo? Are there any features or qualities that you kind of have in the pipeline that you're hoping to maybe either change or upgrade?

Deise Yumi Asami:

Oh, yes. We have a very long roadmap of a lot of things that we want to execute for Maximo. We have a ultimate vision of what this system can be. We are rapidly expanding Maximo to a fleet level and we are also improving the system to continue to empower our teams to go faster for more competitive installations, and we are also focusing a lot now on the digital capability for optimal operations and generating more and more insights to our construction teams.

Taylor Kuykendall:

Looking at the broader maybe global renewable energy landscape, I'm wondering what you see as Maximo's impact both in terms of how it's going to speed up these installations and how that's great for getting more renewable energy out there, but also do you see your competitors trying to make similar kind of technology or is this one of a kind?

Deise Yumi Asami:

From what we know from the industry, Maximo is really the only AI-enabled robot that installs solar panels and does the complete mechanical installation of it. There has been companies out there that has a different approach to automation but delivering the same type of value proposition, and many of direct companies that compares to Maximo has been focusing primarily on the heavy lifting, which is a very important point to be addressed, but it still requires the collaborative work in which the installation might not go as fast as it can go with Maximo.

Taylor Kuykendall:

Deise, excellent. Thank you so much for joining us on Energy Evolution today.

Deise Yumi Asami:

Thanks for having me here today.

Taylor Kuykendall:

All right, listeners, thank you so much for listening to the latest episode of Energy Evolution. I want to give a special shout-out to our incredible podcast team members. I am once again Taylor Kuykendall, but other people behind this podcast include Dan Testa, Camilla Naschert, Camellia Moors, Christopher Coats, Karen Willenbrecht, and our wonderful agency partner, The 199. To stay up to date with our upcoming episodes, be sure to subscribe to Energy Evolution on your favorite platform. If you have any ideas for future themes or guests, we'd love to hear from you. All you have to do is shoot us an email at energyevolution@spglobal.com. Look, if you liked what you heard today, do consider sharing the podcast with others or leaving a review on your preferred podcast platform. We can't wait to bring you more fascinating discussions in the future. Until next time, take care and stay tuned for more from Energy Evolution.