Entrepreneurs for Impact (EFI) Podcast: Transcripts
#170:
Gilwoo Lee, CEO of Zordi — Khosla Ventures Investment. PhD in AI. Smart Agriculture. Robots for Sustainability. Delicious Berries. Inspiring Billionaires. Selling on Non-Climate Benefits.
Podcast Introduction
Chris Wedding:
My guest today is Dr. Gilwoo Lee, founder and CEO of Zordi. Zordi is a BC-backed agricultural platform blending artificial intelligence, robotics, and machine learning to reshape and enhance sustainable greenhouse farming. Gilwoo is a PhD in AI with experience at Facebook, DreamWorks, and Michigan Venture Capital.
In this episode, we talked about why Khosla Ventures invested in them, mainstreaming climate tech solutions by selling on non-climate consumer benefits, their expansion plans into new crops, business models, and geographies, the role of computer vision and machine learning—how her robots measure pests, plant health, and growth rates to optimize yields. Why she gets up as early as five o'clock in the morning, which billionaires inspire her to be more focused and ambitious, and a whole lot more. Hope you enjoy it. And please give Gilwoo and Zordi a shout-out on LinkedIn, Slack, or Twitter by sharing this podcast with your people. Thanks.
Podcast Introduction
Chris Wedding:
Gilwoo, founder and CEO of Zordi, welcome to the podcast.
Gilwoo Lee:
Hi, nice to meet you.
Chris Wedding:
So I know that's a standard thing to say—meet you again, see you again. So Gilwoo, okay, we can add to that part. Gilwoo, I want to start somewhere that's maybe an atypical place to start—except for the fact that it is the kind of lead for your website, which says, and I'm staring over here, "the most delicious produce anywhere available. Everywhere." Now, you do mention artificial intelligence, robotics, and machine learning, so awesome. But you go on to say, let's see, "Our berries are specialty varieties two to three times sweeter than average berries, picked at peak ripeness and delivered within 24 hours."
Meet our ultra-sweet, fragrant berries in your nearby stores. And I think for you, obviously, this is a very intentional—and we'll start here—but like for listeners, it's a great reminder, like we can, we maybe should sell climate tech not on climate tech. Over to you, Dr. Lee. What you got?
Gilwoo Lee:
Is the question whether we should be selling climate tech without mentioning climate tech?
Chris Wedding:
Well, I think it's more like it's a good thing to reach more mainstream, you know, buyers, supporters, maybe even investors if we don't just talk about the climate-related benefits of our, you know, of our gadgets, you know.
Gilwoo Lee:
Yeah, I think it suits well, at least in terms of our business strategy. And I can't speak for everyone because they have different kinds of consumers or clients. For us, we are building these sustainable greenhouses local to metropolitan areas, and what we really want to enable as the end result is for everyone to have access to high-quality, sustainable food. And they know that when they take our product, they are getting these products. So, in a sense, a lot of what we do and our mission is around climate and sustainability. But the first thing that consumers care about—and I want them to care about and enjoy—is that as a result, they're getting really nice, fresh fruits and vegetables.
Chris Wedding:
Well, I think we joked back in Q4 when we talked that, boy, I wish that Ural's solution was close to the Chapel Hill, Durham area, etc., because the berries that my kids adore on their homemade oatmeal pancakes often don't last very long. Right? Maybe even don't last 36 hours despite the exorbitant price we paid for those. So clearly you're solving a pain point. How did you all—well, actually, I want to ask, so many detailed questions—let's stop for a second. Give us the pitch on Zordi more broadly. Who are you? What do you do?
Gilwoo Lee:
Yeah, so our mission at Zordi is to feed the world with high-quality, sustainable food. To achieve that mission, we did a lot of research and study on how to make the production of fresh fruits and vegetables more sustainable. It’s been realized that it's not just about the production itself being sustainable, but also about how close it is to the market and how to make it economically viable. All of those combined is what we're building right now. We're building low-cost greenhouses that are combined with AI and robotics, which can be built closer to metropolitan areas and serve to bring fresh fruits and vegetables to local distribution centers to reach a broader population.
And the combination of these low-cost greenhouses with AI and robotics is what makes our solution not only energy-efficient but also cost-efficient and, therefore, much more scalable.
Chris Wedding:
And what makes your greenhouses low-cost?
Gilwoo Lee:
We rely a lot on the history of technologies developed for greenhouses. There’s a spectrum of greenhouse solutions—from high tunnels to glass ventilator greenhouses—and our robots and AI are designed to operate on all of those. However, you don't have to build super-expensive greenhouses everywhere. It's really more about finding the right technology and solution that works for the climate and serves that market. When we did a lot of climate studies across different parts of the US, trying to do this more broadly, we realized that, oh, you don't have to build super-expensive solutions. There are many regions where we can build these greenhouses in a more affordable, modular way, reducing CapEx overall and significantly cutting the time it takes to build greenhouses.
The interesting part is that a lot of modern technology with AI and robotics, currently being developed for greenhouses, is meant for large-scale greenhouses and not for these high-tunnel, modular greenhouses. So we started designing our robots and AI to work well in both types and across a spectrum of greenhouses, focusing on combining it with low-cost greenhouses wherever possible.
Chris Wedding:
And I think I have the answer, but maybe just clarify—how do you guys make money?
Gilwoo Lee:
Our early business model is to bring these fresh fruits and vegetables at low cost, ideally with affordable prices, to retail stores, so we can be recognized as a brand that offers better, more sustainably grown, high-quality products. That’s how we make money, at least in the near term. Our revenue comes from selling the produce, and our profit comes from scaling efficiently at lower costs. Now, we're entering a phase where our technology is improving, and we want to scale and partner globally to reach broader populations. We're exploring ways to diversify how we utilize our technology.
Chris Wedding:
And how have you been able to attract investors to Zordi? And who are some of those investors that have bought into the dream?
Gilwoo Lee:
We've been fortunate from the beginning to have support from prestigious investors like Khosla Ventures, which has invested heavily in climate, food, AI, and computer science. One notable recent example is OpenAI, which was one of the first institutional investors. I have a lot of respect for Impossible Foods, which was one of the early-stage investors—they understand various sectors and bought into my co-founder Casey's and my vision. To make a real change and disrupt this space, you need to show a working solution.The most effective way—though challenging—is to build a vertically integrated solution that tackles every part of the problem, making everything more efficient. That’s the vision Khosla and many other investors appreciated.
Chris Wedding:
And how did you go from a PhD in AI to, yes, let's combine robots, AI, and low-cost modular greenhouses to sell super sweet berries? Not the most obvious jump. I’m glad you did. But what led to this being the problem you wanted to solve and attracting investors like Khosla?
Gilwoo Lee:
By the time I was graduating, I wanted to apply my expertise in AI and robotics to an industry where I could make a bigger societal impact. Climate change and sustainability were passions of mine outside academia. I wanted to find a way to make those two areas work together, which naturally drew me into food and agriculture. As a roboticist, I found the controlled environment agriculture (CEA) space exciting because automation can make operations easier for robots and automation in general. The fact that greenhouses already incorporate many of these design features made it clear where robots could have a big impact.
It also became obvious when I looked into labor costs—many horticultural tasks require manual labor, and many greenhouses are looking to resolve these issues. Combining AI and robotics with that need felt like a perfect fit. The more I studied, the more excited I became about what I could bring to the table.
Chris Wedding:
And you mentioned earlier that your current revenue model is this vertically integrated approach selling berries in major metros. How might that change in, say, five years? What might that look like to reach more scale—or maybe it stays the same, I don’t know.
Gilwoo Lee:
I think it's crucial to continue with our initial business model—adding more crops, demonstrating the economics, and scaling. Over the next five years, we aim to bring our technology to other greenhouse growers, partners, and regional governments to establish more greenhouses where food security is a priority. We’ll pursue diversified partnerships to expand our reach.
Chris Wedding:
On crop diversification—how do you balance the benefits of diversity with the challenge that you can't do everything in food and agriculture?
Gilwoo Lee:
There's the benefit of diversification, but also the challenge of whether specific crops will be profitable in the near or long term. We chose strawberries for several reasons, including the generalizability of our technology—our harvesting and AI systems can translate well to other fruiting crops such as tomatoes, peppers, cucumbers, and eggplants. These crops account for a significant portion of greenhouse produce. We’ve already demonstrated at a demo level that our harvesting tech translates well without major modifications.
However, expanding to other crops comes with challenges. The economic viability of new crops in greenhouses needs to be validated, and we need to assess how automation can reduce labor costs or improve yields. I am confident it works for the crops we currently target, but generalizing quickly and targeting the right markets will be key.
For example, in new regions, if tomatoes are the main crop and strawberries are rare, we’d prioritize that crop over others.
Chris Wedding:
It seems that if you only diversify by metro and those four or five fruits and vegetables, which are easy to pick, that could be a very large business. Not that I’m a food and agriculture expert—I'm not a PhD in AI either—but help us understand one step further. When you mentioned AI and machine learning, I picture a simplified process: training robots to recognize specific features, like a thin thread attached to an oblong object, not quite spherical. Are you showing millions of pictures of strawberries on a vine to these robots?
So they learn what a strawberry is and how to pick it without bruising? Or is it something entirely different? Gil?
Gilwoo Lee:
Yeah, machine learning has come a long way. We certainly have a large dataset of images of strawberries—mainly because we continuously collect data during our operations. That’s part of the beauty of building and testing these robots—it’s similar to autonomous driving, where you add more data over time. We use computer vision techniques, various ML modules, and accumulated datasets covering different varieties, lighting conditions, and plant configurations. As a result, our models can identify stems, determine ripeness, and decide where and how to act.
This enables our robots to perform harvesting actions accurately. Additionally, our AI-powered scouting robots detect early signs of pests and diseases, analyze crop health and growth patterns—all heavily reliant on vision-based signals. We’re also exploring non-vision signals, different spectrums, and microclimate data to enhance detection.
Chris Wedding:
Right, so it’s more than just picking—pest detection, plant health, growth rate, et cetera. Okay, and when you expand geographically, do you build every new modular greenhouse yourself, or do you just provide the robots and maybe distribution insights? How much control do you need versus serving just a part of the overall value in other metros?
Gilwoo Lee:
Expanding our footprint is exciting, but we need to prove the unit economics of greenhouses in each region first. We want to demonstrate how to build and control greenhouses in different geographies. In the early days, we’ll do a lot ourselves. But for massive scaling in the next 10-20 years, partnering with investors or regional partners is crucial. We’re having conversations with governments and regional growers to understand their needs and priorities.
Ultimately, we plan to expand our own brand and ownership, but also seek partnerships to scale faster and maximize impact. The more aligned and collaborative we are, the quicker we can grow.
Chris Wedding:
That confirms an important point—sometimes you have to build it yourself to prove it, even if the footprint is smaller initially. Later, you can license or sell parts of the technology, but partnerships accelerate scale. Now, tell us about the toughest parts in the last four years—from idea to today.
Gilwoo Lee:
One of the hardest things has been working with the timeline of agriculture. We’ve run many experiments over plant cycles, trying to optimize signals and operations. But plants are plants; you can't fully simulate them. You have limits on trusting early signals. Scaling and waiting for results is challenging.
Another challenge is scale: we’ve demonstrated success in small settings, but moving to larger acres introduces new problems. We’re now managing about one acre across two locations, observing diverse production styles. Ensuring technology works reliably across locations and scales is ongoing.
The next challenge I’m excited about is deploying five to six robots simultaneously in production, demonstrating labor reduction at scale.
Chris Wedding:
When you say five to six robots, are they doing different tasks or the same?
Gilwoo Lee:
Some are scouting, some harvesting and packing, others spraying. Most are harvesting, grading, and packing—the most labor-intensive parts of greenhouse operations. Scouting robots, equipped with sensors and cameras, can cover large areas quickly.
Chris Wedding:
Is there a way that berries produced in this environment could look higher in "intelligence," perhaps reducing sprays or coming closer to organic?
Gilwoo Lee:
Definitely. Controlled environment agriculture already offers resource-efficient growth conditions, maximizing yield while reducing energy and chemicals. Our tech adds early problem detection, enabling preventative measures locally. For example, early pest detection in a small section allows targeted actions—removing leaves or releasing beneficial insects—before problems escalate, making the process more sustainable.
Chris Wedding:
Yeah, more surgical than broad-stroke approaches. Got it. Now, here's a brief message from our sponsors—just kidding, we still don’t take sponsors. But did you know over 100,000 CEOs belong to peer groups? If that makes you feel FOMO, and you’re a CEO or founder, you're in luck. I lead North America's top peer group community for growth-stage CEOs, founders, and investors focused on climate, tech, clean energy, and sustainability. Our members create billions in market value and reduce millions of tons of greenhouse gases through monthly meetings, retreats, and coaching.
If you're interested, visit entrepreneursforimpact.com and join the waiting list today. Gilwoo, let's switch from Zordi’s business to the person growing Zordi—give us two or three pieces of advice you share regularly with emerging professionals, whether coming from technical backgrounds or looking to enter entrepreneurship in the climate space.
Gilwoo Lee:
This is an exciting time, especially with AI and robotics. Many new companies are emerging—some may not find the right business model, but it’s a great time to broaden the scope of industries and problems you work on beyond big tech.
AI and robotics are finally reaching a point where they’re applicable across many sectors—medical, food, service, climate, and more. Broadening your focus helps you find interdisciplinary solutions. It’s not necessarily about everyone becoming entrepreneurs, but diversifying your problem-solving approach can be valuable.
Chris Wedding:
Sometimes crossing sectors without being an expert can bring new insights, unbounded by past biases.
Gilwoo Lee:
Exactly. Vinod Khosla talks about "gene pool engineering"—bringing talented outsiders together with industry experts to solve problems with fresh perspectives, often leading to brilliant solutions. I love working with my co-founder Casey, a sixth-generation farmer and former head grower at Plenty, who brings deep knowledge of the space. His understanding of what needs automation combined with my technical expertise creates a powerful interdisciplinary dynamic.
Chris Wedding:
We see this in our climate CEO peer group too—CEOs from diverse sectors can offer unique insights that help others grow. Now, what habits or routines keep you healthy, sane, and focused on this entrepreneurial journey?
Gilwoo Lee:
For me, waking up early—around 5 a.m.—really helps. I used to have insomnia, so I don’t force myself to wake up if I’m tired, but early mornings give me focused time to plan and prioritize. Good sleep is essential for clarity and energy. I also try to incorporate exercise gradually through small habits to improve day by day.
Chris Wedding:
Never thought about it, but maybe waking at 5 a.m. is a cure for insomnia?
Gilwoo Lee:
Actually, yes. Sleep scientists advise sticking to your wake-up time regardless of when you go to bed—it helps establish a routine.
Chris Wedding:
Any other sleep hacks?
Gilwoo Lee:
Interestingly, since starting my company, I have fewer sleep issues—not because I worry less, but because I recognize what’s feasible in a day. Prioritizing tasks helps me sleep better because I don’t carry unnecessary stress.
Chris Wedding:
So, you put your worries in perspective—things that can’t be done today aren’t worth losing sleep over. And you mentioned tiny habits—there’s a book about that, right?
Gilwoo Lee:
Yes.
Chris Wedding:
"Tiny Habits" by BJ Fogg, I believe?
Gilwoo Lee:
Yes, correct.
Chris Wedding:
For listeners struggling with building new habits, that book is a great resource. Now, can you share two or three books, podcasts, quotes, tools, or resources that listeners might find valuable?
Gilwoo Lee:
I often listen to podcasts or talks by entrepreneurs I admire, like Vinod Khosla. I respect how he thinks and builds. I also listen to or read biographies of founders—like Elon Musk or Masayoshi Shisong—which inspire me, even if I don’t agree with everything they do. One book I recommend is "How to Avoid a Climate Disaster" by Bill Gates—providing a broad perspective on climate issues and solutions across sectors.
As a founder, it’s easy to become myopic about your specific problem space. That book was a helpful reminder of the wide array of climate challenges and the industries involved—energy, transportation, agriculture, and more.
Chris Wedding:
That reminds me of a family reunion I had in small-town Kentucky. One of my uncles, who’s always lived there, told me, "Chris, you can’t change the whole world—only part of it." I initially thought that was limiting, but it’s also wise. We can only do our part. You mentioned earlier that you’re inspired by ambitious figures like Musk.
I was reading "The Laws of Human Nature" by Robert Greene on a flight—about our shadow side. Sometimes it’s good, sometimes it’s bad. It discussed how leaders like Musk can achieve incredible things but also have complex personalities. Wouldn’t it be great if they achieved all they did but were also more chill or benevolent? The question is, would they have achieved the same if they were? I don’t know. You partly answered that—you said you learn from them but don’t necessarily want to emulate their leadership style.
So, you don’t necessarily agree with everything about their leadership?
Gilwoo Lee:
Exactly. I was more referring to their mission, their drive. For example, I may or may not agree with the mission to Mars. Their leadership personas are only partially visible—people show different sides depending on what they need to achieve. I believe you don’t have to have their personality to accomplish great things.
Often, people focus on their strong personalities, but many successful leaders are more subdued or less visible—less interested in media or interviews. For me, the key is their vision and mission-driven approach to solving problems.
Chris Wedding:
Great clarification. A final word—who are the kinds of folks you want to hear from—partners, employees, others?
Gilwoo Lee:
I want to connect more with people focused on climate tech and sustainability. Our work with autonomous greenhouses is an enabler for a more sustainable food future. I’m curious about other sectors—energy, carbon capture, forestry—and how innovation is emerging across industries.
Chris Wedding:
Be careful what you ask for, Gilwoo. That’s a big ask! Everyone listening might reach out to you. We’re excited about your journey from AI PhD to a solution for sustainable food production. We’re rooting for your success.
Gilwoo Lee:
Thank you. Thanks for having me.
Chris Wedding:
Thanks for listening. If you want more insights on climate tech, habits, and deep work, subscribe to our newsletter at entrepreneursforimpact.com or drop me a note on LinkedIn. That’s all, y’all. Take care.