The Exchanges

Q There was a speech that you gave to the UN, uh, in May of 20, 22, where you You basically talked about some of the major challenges that are facing the global food system, and, and one of the things you mentioned was a lack of fertilizer. Why is there a shortage of fertilizer in the world right now?

A Oh gosh, that is actually the biggest problem now. So you're touching on probably the, the most terrifying challenge ahead of us for the next two years is fertilizer. So there are three types of fertilizer. Um, there are nitrogen-based fertilizers, which are basically require natural gas. Uh, there are potash, which is mined, and then phosphate, which is mined. So let's start with the first, nitrogen-based fertilizers. It's dependent on natural gas. Well, Russia, natural gas into Europe, energy crisis, right? It's all linked to the energy crisis, um, and 70 to 90% of the cost of producing nitrogen-based fertilizer is the cost of natural gas. So when the cost of natural gas quadruples, guess what happens to the price of fertilizer? So you basically have A massive sort of spike in fertilizer prices, but what you've also had is a lot of capacity being shut down in Europe as a result. So a lot of the producers just can't afford it. So that sort of cascaded though into challenges in places like Sub-Saharan Africa, where people just already use little fertilizer to using none. Now going into sort of twenty-twenty-three, you have a major, major problem ahead of us, because now you have an affordability and potentially an availability problem, conflating the two. And we can model out what that means to global food production. And we just finished rerunning a new analysis essentially as…

AI assessment note: “nitrogen-based fertilizers... dependent on natural gas. Well, Russia, natural gas into Europe, energy crisis”

Q could result in a version of that in parts of the world, right? Lack of fertilizer, obviously climate disruption, the Ukraine war, which is disrupting grain supplies, and then just supply chain challenges that are, you know, that make it challenging to move things around the world. Are we, are we looking at the possibility of Of, you know, significant famine in parts of the world in the next year?

A Absolutely. So if you look at the price changes in local currency since the start of twenty-twenty for major foods, so if you just look at core grains and sort of your vegetable oils, like the basics, right? Like we're not looking at the fancy stuff. Um, like meat, as far as I'm concerned, is a luxury, you know? Um, so the basics. If you look at the basket of basic food Products around the world since the start of 20 20. The price of a basket of basic products in Sudan is up 1900%. In Syria, it's up 700%. In Ethiopia, it's up 175%. In Argentina, it's up 300%. Even in the U.S., it's up 67%. Europe, 80%. No part of the world is immune right now. And what's going to come down to is which Economies have the resilience. Which governments have the balance sheets to be able to deal with this kind of need, right? And how you deal with it is going to be different. So my sort of concern is that this is so widespread, and every country is busy fighting its own fire, right? That, that it actually becomes so overwhelming to think about collective action, because every country is sort of focused on its own. But we've got to Do something. It's just really hard. And, and we've been spending a lot of time bringing as much attention as possible to the problem for that reason.

AI assessment note: “Absolutely. So if you look at the price changes in local currency”

Q Um, how does your, how does your business model work? Obviously there's, I'm assuming there's a subscription side to it, so companies pay a fee every year to access this data?

A Correct. It's purely subscriptions. So, um, we have different types of subscriptions people can buy. Um, so one of the things that's always mattered to us is sort of being able to serve Really small companies as much as we serve really big companies. Um, we work with financial institutions, again, some of the smallest and some of the world's biggest, um, governments, governments who can't afford to pay as much and, and those that can. Um, and so we've developed a really flexible business model that has allowed us to essentially scale up as organizations are, are bigger, but also start small. But everything is subscription-based and, and And what we've developed is a, is what we call our application store. So think of it as no different from the app store on your iPhone, where you go in and the app store is built on our platform that has all these models that I mentioned. Like when you have two million models, it's too overwhelming for one person to make use of them. So each application has a very specific use case and a very specific set of targeted sort of customers. And so there's a library of applications that people choose from and say, I want to buy this application. And one application can be 10,000 dollars Per seat. Another one can be 15. Another one's 50. Um, depends on the use case, and then depends number of seats, um, or whether it's an enterprise license or not, and…

AI assessment note: “Correct. It's purely subscriptions. So, um, we have different types of subscriptions”

Q So if you can forecast this now knowing what's happening with, with fertilizer production, what effectively could your data lead to? I mean, could it, could it, is it designed in, in, in theory to prompt governments to take action to subsidize fertilizer to, I mean, literally, you know, to, to ship fertilizer to countries that can't afford it?

A Um, it's everything from, uh, central banks around the world looking at what that effect is going to mean to their national balance sheets, because some of these are countries that are actually typically exporters. So when you have exports go down, that means that you have a balance of payments issue that you need to manage. So you can plan in advance for like, how do I mitigate this? Like, do you pre-purchase in the markets? It's like, how do you plan for this so that it's not an absolute catastrophe while it happens? We're telling you this as, you know, you're going into, sort of, twenty-twenty-three. It's being used by the likes of the World Food Program for emergency response. Where is it going to be the worst?

AI assessment note: “It's being used by the likes of the World Food Program for emergency response.”

Q Sarah, let me ask you about, about the business side of what you do. Um, how did you, you know, when you were setting, setting up your, your organization, um, how did you begin to, to gather team? How did you, how did you find people to work with you and to help you build this?

A Yeah, so, you know, I always say I'm the least qualified ex until I find the best qualified person to do that job, right, which is what it is like being an entrepreneur and starting a company from scratch, which is you're completely not prepared, you have no clue, you just have an idea, and then you have to make it work. Um, and so the first person I brought on is, um, is our COO and, and, um, my co-founder, Saweet, and, um, You know, she was a person that came from private equity, um, and finance, and I had known her for a long time from New York, but she'd moved to Kenya, and she was just somebody I just trusted, you know, like, that was it, and she's really smart, and, you know, we'd have to raise capital, and she's been on sort of the investing side, not on the, the operational side, and so I asked her to leave her cushy private equity job and, and go on this wild ride with me, and I was lucky she agreed. So that was step one, and then with every other team member, it was sort of And this was the benefit, I think, of taking time to sort of kick things off was as you're learning, as I, you know, I spent those two years from 20, 20 12 to 14, like I did so many crop tours around the world. I mean, I've done crop tours in the US and South America and Malaysia and China. I mean, I, I traveled the world and I learned just the agricultural industry inside out. And in that process,…

AI assessment note: “the first person I brought on is, um, is our COO and, and, um, my co-founder”

Q How are you able to differentiate your data from what governments provide? Because there are government organizations that provide this information, and it's available for, for, for industry.

A Correct. Um, so first of all, if you think about governments that report, the sort of most extensive is the U.S. government, the USDA, namely. Everything else is just tiny, tiny, tiny order of magnitude of what the USDA does. So to give you some perspective, even in the U.S., our models are predictive and accurate four to six months in In advance of when the USDA comes out with its numbers, um, within like, 98 to 99% accuracy. In places like India, in places like Russia, in places like China, in places like Africa, our lead time for sort of our insights is one to two years at times. So yes, governments report, but they oftentimes report way too late, and they're also not reporting the level of Depth that we report. So a lot of countries will report at the national level. We actually go down to the district level. A lot of governments won't tell you where the crops are grown. We have our own algorithms that looks at imagery to actually identify which fields are growing what crop, because that's how you can then determine how many acres are growing, and then use that acreage to determine the yield and all the stuff. So we really have harnessed, you know, the depth of knowledge of actually public data sets, Because that is necessary. Again, you need to train your models on something. You need a baseline. You need to compare it to something. Uh, but what we offer is fundamentally d…

AI assessment note: “our models are predictive and accurate four to six months in In advance”

Q While you were there, you spent, I think, um, Eight years, um, on Wall Street, you start to think about an idea that would eventually become this business, Grow Intelligence. Um, and I guess the idea really kind of began around 2008 during the financial crisis. What was going on and what led you to start to think about agricultural forecasting and what would eventually become Grow Intelligence?

A It was the oh eight, oh eight, oh nine financial crisis, which those of us that were there at that time will never forget. Um, and I had a colleague, um, Who, um, just thought the world was coming to an end in sort of a genuine way, like he was worried, and he was like, oh my gosh, and, and all I could think of was like, listen, like, honestly, if Morgan Stanley's stock went to zero, which at that time felt like any of that was a possibility, like Lehman had gone under and everything else, like, it would suck, but the world's not come to an end. You know, like, I have seen what close to the end of the world looks like, which was sort of my upbringing. And going back to the comment I made earlier, which is it keeps you grounded and it gives you perspective, right? And so his, his sort of way of hedging for that was buying as much gold as possible and, and lots of guns. And, um, and I one day said to him, like, what are you going to do? Like trade a sack of potatoes for a bar of gold? Like if you think the world's coming to an end, like that just gold seems like the worst investment idea. And, um, and so, um, It was actually that that led me to look at agricultural investing. Coincidentally, was also a time, though, when land, um, in places like Ethiopia, actually, and Sub-Saharan Africa more broadly, there was this really big push by governments to, um, make arable land availabl…

AI assessment note: “It was actually that that led me to look at agricultural investing.”

Q Because we know, right, that, that, um, the amount of food produced in, in the world is more than enough to feed people. In the US alone, um, 40% of food is wasted, right? Um, so when we talk about this idea of food insecurity, what exactly does it mean if, if there is plenty of food available and, and there are still people who don't have access to it?

A Yeah, well, you, you nailed it, which is food insecurity is about access and affordability. Um, because it's, it's, you can have a lot of food, but if you can't afford it, that's also not helpful. And so what we have as a world in terms of a challenge is both an access and an affordability problem, and it looks very different depending on what part of the world you're looking at, right? So if you're looking at food insecurity in Sub-Saharan Africa or in South Asia, It is an access issue, um, in the sense that not even enough is produced locally. So there is a, a food insecurity that comes from the fact that domestic production is not sufficient to meet the local needs. And then on top of that, there is an affordability issue, which is when you import it, obviously food is, is more expensive. Um, and then if you look at food insecurity in a wealthy country like the U S it's very different, right?

AI assessment note: “food insecurity is about access and affordability.”

Q How did you kind of develop the technology platform? Um, I mean, you're obviously super smart and come with incredible experience around commodities, but was there a steep learning curve for you to figure out that the technology side of it?

A It's a steep learning curve on every single one of it. Cause I was not a qualified agronomist. I'm not a qualified fertilizer trader. I'm not, you know, I was only a qualified energy trader. Like that was my qualification. Um, not a qualified CEO, right? Um, building a team is so different than, than, than trading a book. You know, everything was a steep learning curve for me, but one thing that I think has always driven me is that I love this work, right? Like I really, really, really believe in the work that we're doing. And so that just has made me completely relentless in learning anything and everything required at that point in time to make sure that we're successful and that, that I'm sort of doing right by the team and that I'm doing right by our investors, by the business, every, you know, just, it matters a lot to me. So I think I already had a deeply sort of technical mind in the sense that, you know, I was always technical even growing up. And even at Morgan Stanley, I, you You know, built my own options trading models and, and, and, and built out sort of these things, but it was always by teaching myself, and so I'm not sort of afraid to not know. I'm never afraid to, like, call people and say, like, I need to learn, and I need to learn from you. So in the early days, I literally used to, like, for example, there was this, um, this scientist at USGS out in, um, in …

AI assessment note: “It's a steep learning curve on every single one of it.”

Q starting to uncover. I mean, essentially, I guess you realize that there, there just isn't enough data for investors and businesses and, and, and governments to understand what might happen to the price of different commodities, right? Especially, uh, in agriculture. And I guess that lack of information, uh, Uh, can make problems like food insecurity even worse. Is that, is that more or less what you started to realize?

A Well, yeah. I mean, you know, when I was trading, I managed our natural gas options business. And if you think about sort of natural gas and, and, and any commodity, like the thing that drives the price of that product is like, how much supply is there? How much demand is it? And what's the clearing price that clears the supply and the demand? You know, hurricane season constantly disrupted, uh, Production. So during hurricane season, you're using weather data and weather forecasts to sort of see the probability of a hurricane hitting some type of production. Or in the winter, how cold is it going to be? Because that drives how much heating demand there is. So, you know, all of those are pieces that sort of drive that the price of any commodity. And, um, and, and agriculture is one of them. So grow just became this idea to, to do it for agriculture. And that didn't exist at the time.

AI assessment note: “Well, yeah... grow just became this idea to, to do it for agriculture.”

Q Lab. I'm Guy Raz, and my guest is Sarah Manker, founder of Grow Intelligence. It's a company that's using artificial intelligence to create forecasts for global food supply, demand, and pricing. So, alright, so you launched Grow Intelligence in 2014, and now your clients include food suppliers, um, agricultural businesses, financial institutions, um, How do you provide forecasting models for them? What, what kind of data points do you provide?

A Yeah, so we, so taking a step back, what we did is, you know, the company's gone through sort of a journey, right, since 2014. The first set of challenges we dealt with was, can we get enough data in our system? Like, what data is out there? Like, do we even know what's out there? It's sort of, you know, one of the things that made agriculture so much more complicated than sort of the work that we did on the energy markets is that agriculture is not a single product. You know, oil is oil, natural gas is a natural gas. Agriculture is tens of thousands of different products. Every single one is sort of governed by a different set of biological rules that govern how it grows. Um, supply is super fragmented. It can be in a half acre farm or a 100,000 acre farm, right? So our first challenge was saying, what data is even out there? And can we ingest it and come up with a technology that can take that data and standardize it And bring it in, in any language, in any format. So we shouldn't care if it's in PDF. We shouldn't care if it's images. We shouldn't care if it's in Mandarin or Portuguese or English. We should be able to take that, automatically translate it, and also standardize the format. So what that then gives you is too much data, right? So we get data from 50,000 sources around the world. Um, they come in all these different formats, languages, but data on its own is not …

AI assessment note: “we get data from 50,000 sources around the world... take that and how do you develop insights”