Lisbeth Kaufman, Head of Climate Tech at AWS, sits down with Ryan to talk about her work helping climate tech startups get off the ground and the role startups can play in addressing the climate change crisis. She highlights a few projects to get excited about, including ones focused on fusion energy and sustainable agriculture.
Lisbeth cofounded and coleads the Compute for Climate Fellowship, which funds climate tech startups using advanced cloud computing and AI. Applications are open until April 6, 2025.
Read about how some climate tech startups are leveraging GenAI.
Connect with Lisbeth on LinkedIn.
Stack Overflow user JohnsonYuan earned a Lifeboat badge by answering Why can't I enter the url on my phone's browser to view my live site?. Nice work!
[intro music plays]
Ryan Donovan Hello everyone, and welcome to the Stack Overflow Podcast, a place to talk to all things software and technology. I'm Ryan Donovan, your host today, and we are going to be talking about climate change and startups and how they're the only ones who can save us. My guest today is Lisbeth Kaufman. She's the Head of Climate Tech Startups at AWS. She's going to tell us how we can save the world. Lisbeth, welcome to the program.
Lisbeth Kaufman Thank you. Thank you so much for having me. I'm super excited to share a bit about the startups that we work with and about the Compute for Climate Fellowship that we just launched as well.
RD Yeah, absolutely. So we like to ask our guests how they got into software and technology, and you are at an intersection of technology. So how did you get into that intersection?
LK Yeah, I've been working at the intersection of tech and climate, as you said, for about 17-18 years now, and I maybe have a bit of an unusual background for getting into the space. I actually think I was inspired as a kid early on by my dad. My dad is a filmmaker. He makes independent films, cult horror films, and he made a movie about an environmental superhero.
RD Is your father Lloyd Kaufman?
LK Yes! Yeah, exactly.
RD Amazing.
LK Some folks may know, he's very indie.
RD I'm a huge fan, huge fan.
LK No way! I love it, that's awesome. So he inspired me. He made a movie called The Toxic Avenger that literally is an environmental superhero fighting toxic waste. So as a small child, that got the idea of the environment and environmental protection on my radar. Then when I graduated from college, I wanted to work in climate and so I started in policy in Washington, D.C. where I worked in a few different places and ended up getting recruited into the Senate to be a Senate staffer focused on energy, environment, and agriculture. So that kind of launched me into the climate space and was an awesome learning experience to be focused on the policy side of things. But eventually I wanted to get on the business side, so I got my MBA at NYU Stern and I ended up starting a startup there, a venture-backed tech startup. And then when that didn't go exactly as planned, we ended up getting that company to profitability and I left and got recruited into AWS where I am today. And I've been here for five years, and I had the chance to start the climate tech team here.
RD That's great. People may not think fighting climate change associated with AWS. Can you tell us a little bit about what you do at AWS?
LK Yes. So I'm the Head of Climate Tech Startups. My goal is to make AWS as helpful as possible to climate tech startups. Ultimately, we want to help climate tech startups grow and succeed as businesses and then also help them address the climate crisis. And so we have a few different ways that we help, a few different buckets of resources. We have technical support, we have funding, which is where the Compute for Climate Fellowship falls into, and then we also have some go-to-market support to help climate tech startups get customers and grow.
RD One of the interesting things in the initial email was saying that startups are uniquely poised to help with climate change. Why do you think that is?
LK I mean, these startups in climate are incredible. They are just so innovative, and the founders, these are not normal founders. They are often scientists and engineers at the top of their field who are then applying their deep expertise and scientific knowledge to address the climate crisis, and so I'm seeing climate tech startups in the forefront of addressing climate. Plus, with startups in general– your audience probably has a lot of startup founders so I’m sure they'll know what I'm talking about– but startups are just so innovative. If there's a problem, they get excited, and climate is one of the biggest problems that we've faced in human history. I think that they are the ones who are going to figure this out. They're just so good at tackling problems.
RD Do you think big companies are a little slow to move on sort of problems like that?
LK I mean, I think big companies are addressing this too, and actually at AWS, we're maybe an example. We want to support these climate tech startups and we're trying to do that as quickly as possible. And I think we're pretty well positioned to help climate tech startups and that's why we've started this Compute for Climate Fellowship to fund more climate tech startups and get their projects to life with funding and our technology. But to answer your question, I do think startups are particularly well positioned to just move quickly. Small, agile innovative, they can really tackle things quickly where maybe bigger companies have to take a slower pace sometimes.
RD And startups seem to originate as a single problem sort of a company. As a big company grows, it diversifies into all these other fields and they have to balance their business needs and a startup can say, “Well, I'm doing this one thing and I'm doing it well.”
LK That's a really good point.
RD So these companies are also using AI to address the climate crisis. Can you talk about how AI fits into all of this too?
LK We're seeing AI become a super useful tool for climate tech startups to address the climate crisis. In the climate crisis, speed is so critical. We need to move as quickly as possible. I have a three-year-old daughter and I think about what the world's going to be like in 2050 when she's 28, and we’ve got to move as quickly as possible to make sure that that world is a good place to live. And so simply put, AI and cloud computing in general are very good at speeding things up and enabling things that weren't possible before. Maybe one example is a company called Smartex, and they are helping to reduce waste and carbon emissions in the fashion space. So they have an AI platform that uses computer vision to identify flaws in fabric so that the fabric manufacturers don't continue to make flawed fabric. Because before this existed, the manufacturers didn't know that flaws were happening. They'd make a huge amount of fabric, they'd throw it away and it would end up in landfill. So with Smartex, these manufacturers can now stop the flaws and reduce all that waste. And so with our program using AI, they've been able to improve their models or algorithms to identify those flaws and now they're expanding into different types of fabrics and different types of processes so they can help even more types of manufacturers reduce waste in fashion.
RD That's really interesting. I've seen other use cases for computer vision, machine learning, where it's for tumors and such, but it's really good at understanding what a flaw is.
LK Yeah, exactly.
RD So I've read a couple articles in the last few years about the sort of the energy costs of both AI and tech in general. Are there climate startups that are looking to sort of reduce that energy and climate impact of AI and cloud computing?
LK Yeah, there certainly are. There's a whole suite of companies that are helping to, first off, create more renewable energy to power data centers and AI, and then also decarbonize the data centers. And actually, I don't know if a lot of people know this, but AWS has already done a huge amount of work to decarbonize our data centers. So I like to ask people, do you have a guess at what percentage of AWS's energy is powered or matched by renewable energy? Do you have a guess?
RD Do I have a guess?
LK I don't know. Maybe you've researched it.
RD I don't have a real guess. I would say 70%.
LK You're higher than most actually, sometimes I get lower, but the answer is 100%. So AWS met our goal to match all of the electricity for our operations across actually all of Amazon. We're matching it with 100% renewable energy, and we met that goal seven years early. So that's just one example of how we're working on decarbonizing our operations of all of Amazon, including our data centers and our AI infrastructure.
RD And the renewable resources energy makes a certain amount of business sense too. You set up the equipment and you just get the energy from the sun, you don't have to dig up more things to make the energy.
LK Yeah. And actually in the Compute for Climate Fellowship, on the topic of renewable energy, we have an amazing company called Realta Fusion. They're building fusion energy plants. So I don't know if your audience is familiar with fusion. Should I give a quick overview?
RD There may be a few, but give us a quick overview.
LK So fusion is basically clean, low cost, abundant energy with no waste, and it's essentially harnessing the power of the sun but here on earth. So they fuse atoms and then release an enormous amount of energy out of that. So there's just so much promise infusion, but understandably, it's very difficult to take a star or a sun and bring it down here on Earth. So Realta Fusion are doing this in really interesting ways with magnetic mirror technology to basically create a magnetic bottle to hold all that energy. And so they came to us because they were having trouble, or they need to do an extremely intensive modeling of the plasma modeling which is pretty critical for fusion, the whole fusion industry, but the only computers that could handle these models are giant supercomputers and there's only two of these supercomputers in the US and there's a year-long wait to get to these computers. So Realta Fusion wants to move as quickly as possible, so they came to us and joined the Compute for Climate Fellowship and we helped speed all of that up. We recreated the supercomputers in the cloud and we enabled Realta to run these models in the cloud for the very first time. So instead of waiting a year to get access to the modeling, they set up their models in a week. And so now they're able to do their R&D more quickly and they can get to actual commercial grade fusion more quickly, and this democratizes access to these models for the whole fusion industry. So we're hopeful that this can speed up the whole fusion industry to get to commercial grade fusion more quickly. So speaking of renewable energy, which you brought up, this is such a cool space of clean energy that can help them power AI or power everything we need.
RD I've definitely heard of some fusion projects. I know there's a massive one in France where they're trying to build a Torus reactor. I don't know if that's related to Realta.
LK In France they just had a new breakthrough I think in the past few weeks. This is such a cool space. People talk about the space race, but there's also a fusion race, and there are these different companies and research organizations that are moving as quickly as possible to get fusion to the place where it can be commercial grade and power industrial processes or power plants. So it's definitely related to what Realta Fusion is doing.
RD Super interesting thing I'm going to touch on. You said there was the two supercomputers that could do it and then you were able to duplicate or give that processing power with cloud computing. Do you know how that works and does that make these supercomputers kind of an obsolete thing?
LK So I don't know if they're obsolete. I'm sure there's still plenty of applications, but it was pretty exciting to be able to recreate them in the cloud. Realta Fusion used some AWS services to do that. So they used AWS's EC2 high performance computing instances, and then they used the elastic fabric adapter to basically stitch together a bunch of computers to run these simulations. It's definitely repeatable, it's doable. At AWS, we've done other enormous high performance computing runs similar to that, and we're hopeful we can help more fusion companies do this. And actually we have another fusion company now in the fellowship as well doing a different approach called Thea Energy.
RD Wow. That's cool. What is the different approach?
LK Realta Fusion has magnetic mirrors where they've basically made a magnetic bottle to hold the plasma, and the plasma is where the atoms get merged. Thea Energy, they're doing a stellarator, which is this donut shape where they have the plasma.
RD Oh, yes. That’s the torus.
LK Yes, exactly. It's really cool looking. It's kind of beautiful. So they have a donut. But what Thea is doing is that traditionally the stellarator has been extremely complicated to build and it takes years to manufacture because there's these individual pieces that are each different. So Thea Energy is building a modular system that uses these magnets that are basically repeatable across the donut and then are programmable to do individually different things. So that means that it'll be easier to manufacture these plants and then you can build more quickly and get to the Holy Grail, the commercial scale fusion more quickly.
RD So it's a plug and play a fusion reactor almost?
LK Yeah, something like that. That's the goal, to make it as easy as possible. And then they've joined the Compute for Climate Fellowship to use AI and high performance computing to design their reactors and their plants.
RD That's cool. So we touched on a couple really cool companies’ instances. Are there other companies that you think are doing– if you're going to pick one that you're like, “Oh, this one's going to save the world,” which would it be?
LK I mean, they're all amazing. And I mean, Realta Fusion is so cool. There's another company that I love and is a great example of using AI to address climate issues. This one is called Phytoform and it's focusing on creating climate resilient agriculture so that no matter what happens with the climate crisis, no matter how bad climate risks and impacts get, farmers will still be able to grow food. So Phytoform is out of the UK. They've built an AI platform that can invent and identify new genetic traits so that plants can be more climate resilient. And so in our fellowship, we've helped them improve on their models to have higher accuracy, and now their platform is bearing fruit, pun intended, because they've actually recently announced that they have invented a new tomato plant that uses less water, less land, and then produces 400% more fruit, more tomatoes. So if we have droughts, this tomato can still produce food for the world. So that could save lives. But actually, one thing I realized I haven't explained is what the Compute for Climate Fellowship actually is.
RD Absolutely. Please do.
LK So the Compute for Climate Fellowship is an R&D funding program for climate tech startups to build groundbreaking new projects to address the climate crisis. So startups apply with a proposal for a project or a proof of concept and then AWS funds that project with cloud credits, and then we also bring in experts who are PhDs in the space that the startup is working in and then we get them access to all of the most advanced technology and help them build over about two to three months, and then we try and bring them as much visibility as possible through speaking opportunities.
RD And I think the application period will be open as of this release, right?
LK Yes, that's right. So applications are currently open. Deadline is April 6th. And it's a global program, so we would love to have as many climate tech startups from all over the world apply. And this year, so this is our third year running this program, and this year we've more than doubled the size of the program. We recognize that this is such a big issue and as the impacts of climate change get even more complex, we want to really double down on this, so we're expanding the program this year to accept 20 companies, up from eight last year. So we really hope that your audience, your listeners can apply. We'd love to have them in the program.
RD If you've got an idea, you've got a proof of concept, get the application. We'll include it in the show notes.
LK Thank you.
RD But I want to get back to talking about some of the other companies that y'all have gotten into the fellowship that have done some interesting things with the cloud computing.
LK Okay. Let's talk about there's a company Cosma. They have underwater drones mapping the seafloor. Now, you might not be thinking about the bottom of the ocean when you wake up in the morning, but it's actually a really important area. Scientists know more about the moon than they know about the seafloor, the ocean floor, so collecting information about the biodiversity and the ecosystems down there is extremely important, especially as we're doing more ocean industries like offshore energy. So Cosma has developed an AI platform and they have these underwater drones crawling around on the seafloor and then they use computer vision to identify and track the ecosystem and identify rare species. They're in our program and they've used AI to develop their first benthic ecosystem monitoring system. So Benthic I think is bottom of the ocean, and they've shown me some of the images that their drones have collected and it's beautiful and incredible. There are these very reclusive octopi they've been able to see that are usually unidentifiable, and they can not only see them, now they can quantify them. So that means that then their customers who are maybe the offshore energy companies can now quantify their impact and make sure that they're protecting the ecosystem and biodiversity.
RD Right. And I'm sure an octopus like that is a kind of a canary for whatever environmental impacts are on the ocean. So for a lot of these startups, the drones in the bottom of the ocean seems like it's an application for NOAA or something. Are the customers for these startups largely governmental organizations or otherwise?
LK No, most of them are B2B. So Cosma is selling to energy developers or consultants who do the environmental impact reporting, and Realta Fusion will eventually sell to maybe utilities or energy consumers. Phytoform, I think they sell to farmers or the CPG companies that need the specific ingredients for their products. And I was thinking back to your question earlier about why startups are so well positioned to address the climate crisis, startups that we work with have solutions that are better and cheaper and faster than whatever the existing options are that address the climate crisis, so their customers don't even have to care about the climate crisis and they still want to use their technology just because they're better products. It's sort of the startup fundamentals. So Realta Fusion, when they get their energy out, it's going to be so affordable and consistent and clean and abundant that it is going to be a better option.
RD I mean, the dream of cold fusion has been around for a long time. If they are able to do it chopping whatever piece of matter in there and turning it into energy.
LK Yeah, exactly.
RD Is there anything you want to talk about before we go to the outro?
LK Yeah, thank you. I'd love to just reiterate that we have just launched the Compute for Climate Fellowship. It's a global program for startups. We want to fund groundbreaking projects built by startups addressing the climate crisis. And I should also mention it's a partnership with UNESCO. There's an AI organization within UNESCO called the International Research Center on AI, so they're bringing their expertise and their channels for visibility for these startups and we want to have as many companies apply as possible, or we want to help as many companies. We have a new focus area this year for the program. So historically we've had eight different focus areas that include clean energy, sustainable agriculture, low carbon transportation, circular economy, these various ways to address the climate crisis. This year we have a focus on indigenous solutions in addition to those other focus areas. So we're eager to support startups and indigenous technologists who are using technology and AWS services to address the climate crisis. We recognize that they have so much to offer in addressing the climate crisis with their traditional indigenous wisdom and incorporating that with technology.
RD That's right. Sometimes the old techs are still good.
LK Yeah, exactly.
[music plays]
RD Well, we are at that time of the show again, ladies and gentlemen, where we shout out a question– somebody who came onto Stack Overflow and dropped a little knowledge and got a badge because of it. Today, we're shouting out JohnsonYuan, who won a Lifeboat Badge for dropping an answer on: “Why can't I enter the url on my phone's browser to view my live site?” Why? So if you're curious about that, we'll leave the question in the show notes and you can check it out. I am Ryan Donovan. I edit the blog, host the podcast here at Stack Overflow. If you want to reach out to us, the email is podcast@stackoverflow.com. And if you liked what you heard today, leave a rating and review.
LK Great. And I'm Lisbeth Kaufman, Head of Climate Tech Startups at AWS, and I'm the Co-Chair of the Compute for Climate Fellowship. And you can apply. If you're a startup in climate, please apply for the fellowship and you can go to ircai.org for the application. Thank you so much.
RD All right. Thank you all for listening, and we'll talk to you next time.
[outro music plays]