Pod Insights a16z Podcast a16z Podcast | Data Down on the Farm About this episode Farmers are among the best hackers in the business. They can fix anything, and are endlessly tweaking their approach to a business that is up against the strongest force on the planet: nature.
As adopters of technology, the agriculture industry is both forward-thinking and, at the same time, hard to convince to make a change. For good reason -- you can’t A/B-test an almond orchard. You get one shot a year to grow a crop and make a profit. So whatever technology makes its way into the fields had better work.
U. C. Santa Barbara computer science professor Chandra Krintz, Granular CEO and co-founder Sid Gorham, and Fruition Sciences co-founder Thibaut Scholasch join this segment of the a16z Podcast to discuss technologies working down on the farm. Including how sensors monitoring soil elements, water flow in plants, and fertilizer schedules are yielding fruitful data. Given the high stakes -- a planet in the midst of climate change -- what we can all learn from how technology is accepted and implemented in the agriculture industry?
Transcript 0:00 Speaker 3 Welcome to the A16Z podcast. I'm Michael Copeland, and this is sort of our holiday special. We're going to talk about food and technology. And to help us do this, we have three people who are perfectly suited on the line or on Skype, I should say, is UC Santa Barbara Professor Chandra Krince. And with us also here at Granular in San Francisco is the CEO of granular, Sid Gorham, and Thibaut Skolash, who is the co-founder and CTO of fruition science. Gentlemen, Chandra, welcome.
0:32 Speaker 0 Thank you.
0:33 Speaker 3 Thanks for having us. So I we have a thesis at the firm that software is eating the world, and this is where puns could start coming fast and thick, which we'll try not to make. But in agriculture, I want to talk about how technology is making inroads. And each of you are working toward that in sort of overlapping but different ways. Chandra, we can start with you. You've got this concept of smart farm. But what where has agriculture been or where is it at present in terms of its adoption of technology, what kinds of things, and then what are you trying to get moving faster than it might otherwise?
1:12 Speaker 0 Sure, that's great, great question. So we believe that agriculture is tremendously underserved by technology. It's been left behind in both software and hardware, although hardware has done a little bit better than software. So we think it's a prime uh area that we can contribute to both on the research side and working with startups and industry on the industry side. Uh we see that there being uh uh increasingly open minds to trying technologies in order to make farms and the farm processes more uh efficient. And uh so there's been a lot of movement uh in the device uh realm in the hardware realm where sensors, very low-cost sensors have been developed, uh they are connected via radio and cellular networks on the farms to get to collect information to collect data.
2:03 Speaker 0 Turns out that these uh sensors, however, are um proprietary and they all do a little bit something different and have different interfaces, and so it's really hard to uh go between vendors if you wanted to try lots of different sensors on your farm, whether ones that run on your tractor or that go in the ground and tell you how much how much water your uh soil needs or how your plants are doing. And so part of our uh initiative at UCSB from a research perspective and an open source community buildings perspective is to try to design systems, develop systems that are freely available and open source that start to bring data together from different vendors, aggregate it, put it into a form that growers can control and access it, and that interoperates with all the great technologies that are coming out of the public cloud in a more of a hybrid cloud fashion.
3:00 Speaker 0 So you can do some work on the farm, some work in the cloud, and get the benefits of uh both local and more cloud-based technologies.
3:09 Speaker 3 Thibaut, you you guys have a lot of sensing equipment, both hardware, but then you know, more importantly, you you extract all this data from vineyards. How does your um technology and your hardware and software sort of fit into this picture that that Chandra is describing?
3:24 Speaker 1 So that there is an opportunity from running analytics on uh whatever information is on the cloud. The issue is what uh parameters is the most uh relevant to assess the specific plant needs and plant needs are different for different crops. So there is a lot of work that then needs to be done once you have uh uniformized uh access to all those um crop parameters or land parameters or environmental parameters to run the analysis, and that's where you need to uh enter with plant intelligence to then help the farmer who has so many other things to deal with, uh getting fast to the best uh best decision.
4:07 Speaker 3 Chandra, you were talking about how agriculture hasn't seen kind of the benefit of some technology like we've seen in other industries, let's say finance, for example. Granular is is kind of a workflow for farmers, and you know, we all think of farms and we think of people on tractors or you know, walking down roads, it's very kind of you know, prustian and pastoral. But can it be both that and also sort of up-to-date and and um utilizing the best of technology?
4:36 Speaker 2 Yeah, but I think what's interesting in this transformation is that a lot of people assume that um farmers are just kind of waking up to the modern world and saying, huh, maybe I can use technology. And I I actually think they're they're early adopters of technologies and in many ways have kind of a hacker mindset. They're very remote businesses, self-reliant businesses. They're juggling a lot of different you know complexities. And I think really what's happened is that cloud and mobile have made it possible to use technology on the farm, and the farmers are willing and ready to pick up whatever tool will help their business. Right. So it isn't so much a story of um uh sleepy businesses waking up to technology, as in my mind, technology or the right technology arriving at the farm.
5:22 Speaker 3 Right. Well, and so it's gotta work, and right and then the the kind of evidence is fairly clear, right?
5:29 Speaker 2 Yep. Yep. I mean, I think um many different things have worked, both on the hardware side and on the analytics and software side, and I think um granular's particular focus is on making farms better run businesses. Um we think it's critical that farms understand where their profit comes from, what resources generate a return for them so that they can be more careful about uh using those resources and make sure one of the phrases we like to use is we want them focused on their dollar yield, not just on their uh bushel or you know, pound or ton yield, right because really you want farms to be very, very uh careful about how they use resources and make sure they're only using resources that that generate profit for the business.
6:13 Speaker 3 Right. So all the inputs and the outputs at the end in terms of what whatever profit that's used to. So you guys have all, I think, sort of touched on resources. Thibaut and Chandra, how do uh sensors and this data help utilize resources better? Teba, you were talking about the right plant for the right job, sort of um, but how how does that that kind of precision and that insight help change what farms do and how they act and at the end of the day what what we eat or drink?
6:45 Speaker 1 So it's true that there is an explosion of new sensing technologies, and uh some are providing data that are just nice to have, but not necessarily leading to actions. Uh some are uh mission critical. So if I take the example of the water flow inside a plant, it's a little bit like the blood flow inside a human body. Um by diagnosing the the water flow, you can sense when the plant is actually in need of water, and uh if the need gets too uh severely limiting, the plant will uh underperform. But you can also see if the plant is not in need of water but still underperforms, you can then point your finger at other resources that are limiting, and that in the past were misdiagnosed for water deficit.
7:34 Speaker 1 So it uh exacerbates uh a little bit your ability to tease apart what the plant really needs and what the farmer really needs to adjust in order to m maximize the um the crop performances. So that's just one example by looking at the potential benefits from looking at a vital index such as the water flow. Shandra, what are you saying then along those lines?
7:59 Speaker 0 Yes, uh we see a lot of um opportunities and excitement around using the analytics like Tebow uh has been talking about to actually control a lot of the operations on the farm, like the irrigation scheduling can be done automatically based on historical and current inputs. Uh uh Tebow referred to them as parameters, it's exactly right. We see uh uh uh precision application of fertilizers and pesticides. You can even do this as at with robots, so completely eliminate the need for uh manual intervention there. Uh and that allows you to direct the pesticides or the fertilizers, which are harsh on the environment to exactly where the problem is instead of more globally uh across a uh a field, and that prevents you know runoff and all the other side effects that we have uh downstream from these techno these uh applications.
8:53 Speaker 0 It saves farmers money if you're not they're not don't have to apply tons of per pesticides and fertilizers. And I also see it a lot in harvest. If you can control and guide irrigation and fertilization very precisely, you can ensure and control and guide when uh the plants uh are are producing when they and and and make it more consistent. If they all produce at one time, say for almonds, you only need to do one shake of the trees to harvest them. That lowers uh farm costs that Sid was talking about, optimizing the business side. So the uh it's a really an end-to-end uh solution that can be applied in lots of different ways.
9:35 Speaker 3 No, and so just to be clear because I I've never been to an almond farm, but like the one shake is because like it's that precise moment where all the almonds are ready to go and ready to drop. And so you just at
9:47 Speaker 0 the same time. That's right. Give it a bump. A lot of times, if you uh part of the farm has less water than another part, um certain trees will be ready for harvest before the others, and that caught that that's costly because you'd have to. Perform two harvests. What you want to do is manipulate both the fertilization and the watering process so that all the trees are ready for harvest at the same time. So you put out one team, you get it done in a day, and uh you don't have to go back and do it again.
10:17 Speaker 3 Sid from your perspective, when when farms, whether it's small scale or large scale, sort of adopt these kinds of technologies to run more efficiently as businesses, um how do they change?
10:30 Speaker 2 Well, I think one of the interesting things is most farms in America are family-run businesses. Um even the very biggest farms are typically family-run businesses. Um and as such, and be and because nobody has often people spend their entire careers in farming, so they don't know how other businesses run. They they run with a very informal management style. And so I think one of the sort of side effects of using products like granular or other software packages on the production side or on the business side, it sort of pushes them to be more um uh more organized and more a little bit more um regimented in how they think about the business. I think that's true in every industry. What often what happens when you adopt software is it formalizes your process and makes you a little bit you know, better run business, you know, above and beyond the impact of the software, and we see that in farming for sure.
11:24 Speaker 3 Thiba, you guys started in the wine industry. Um is that because it's just such a high value, like why were they willing to do it? And you know, Sid, you said that, and I agree that farmers are these hackers who can fix a tractor and fix you know this and they run off the grid if they have to, etc. But why the wine industry and how because there's this weird kind of binary, maybe weird is not the right word, but like part of the wine industry is like plant the skull of the cow under a full moon, um and your wine will grapes will be great. Um and the other part maybe is just much more you know technology and data-driven.
12:00 Speaker 1 So first is because uh clearly the the wine is a high cash crop uh value. Um and then there is another thing is that historically uh vineyards were always um left to grow in areas where we couldn't grow anything else. So the the plant itself is uh programmed to perform in otherwise very challenging and tough conditions. So it prefigures a little bit what could become the future of agriculture as uh land aridity and global warming uh keeps making the land a bit more uh tough on the plant. So for those two reasons high cash crop and also because the the nature of the plant itself uh makes it very suitable to to push a little bit the boundaries of where we can still grow.
12:48 Speaker 1 Uh it's it's a very good uh beach head to to start developing this approach. The the plant is very uh good at resisting drought. The plant is very good, the divine plant is very good at um performing even when there is not so much uh fertilizer or nutrient availability. So it helps you understand a little bit to which extent we can still uh make it uh profit profitable to to farm uh a parcel of land uh while still enhancing uh the production of a very um profitable crop. And so the the the market
13:23 Speaker 3 is good to test new technologies for that. So you're saying that you know farmers are actually pretty eager to go on this, but um the wine industry seems fairly old school in some ways. Um I mean, did they uh embrace this immediately or do they need I guess my uh my question is do you need does everyone need evidence that this stuff works before they're willing to give it a try?
13:46 Speaker 0 It's highly varied. There are uh families and um businesses that are very eager to improve their business processes and um and make more money because of efficiencies that come with technology. There are some that are highly resistant because they like because exactly like you said, they need evidence. They it's a crop life cycle is very long. It's not like in software and SaaS where we produce something and uh uh and it's out there tomorrow. Uh you know, this is we we we do something and it takes a year, you know, and it's a huge risk, and their lives depend on it. So there's they're very resistant to risk in many cases. And so you have to have evidence, which means experimental farms and partnerships with universities and trying things and and showing and proving to farmers that this stuff is actually gonna make a difference in their lives.
14:38 Speaker 0 Um I have a really funny story. The I was trying to uh help a farmer, uh actually a grape grower uh identify leaks uh using a combination of electricity sensors and water sensors um on the irrigation system and that's leaks with an
14:52 Speaker 3 A, not leaks with two E's, right? Okay. A leak, a leak of water leak in there in
14:56 Speaker 0 their yes. And they said, and and the and the farmer said, well, I know where there's when there's a leak, I just look for the birds. And so they you know, they have all of these wonderful mechanisms, because if the birds are around, that means that they're they you know they're thirsty, they go find the the water underneath the ground and then they find the leak. So they have all these mechanisms to solve the problems already. They are geniuses at this, and so we have to come in and be um uh and and provide them with the evidence they need to take on something new.
15:26 Speaker 3 Tebo, how how open were the arms of the wine industry to sort of get this done? Uh
15:30 Speaker 1 but paradoxically, it's not so much the fact that they cope with uh farming practices that are a little bit more obscure or hard to understand, like uh you referred to by biodynamic. It's more of a societal and psychological predisposition. So as a matter of fact, for us, it's much easier to develop our ideas in California, which has a culture of a young society ready to embrace new technologies compared to um the old world, like France or Italy or Portugal. So there is this societal footprint. And then the fact that you have always been uh farming with strong respect for tradition is not at all in opposition with uh adoption of new technologies. In fact, the main driver here is how uh critical it is for me to not mess up the next crop.
16:20 Speaker 1 As Chandra was saying, you only have one chance per year to make a profit. And so if you have two or three years of bad um crop performances in a row, you know you the next year is the last one you have on the line. And so if you don't maximize your chances to make it happen well, you're screwed. So there is also these uh money incentives uh regardless of the psychological or societal predisposition.
16:48 Speaker 3 And and like, you know, in the software world you can A B test, but you can't A B test a vineyard or uh an
16:52 Speaker 2 almond tree, right? Exactly
16:54 Speaker 0 right.
16:55 Speaker 3 So Sid, what are people can how can they sort of turn the knobs, you know, from from your perspective on kind of managing this live or die situation.
17:05 Speaker 2 Yeah, I mean it is a tricky technology adoption uh challenge because I mean, just like in any other industry, all of us are um creating a category, right? You're not replacing some on-premise vineyard software with your new thing. You're you're literally bringing the first software and analytics package they've used, and we're doing the same thing on corn farms in Iowa. So I think you have all the challenges that come along with that, but then you have this dynamic we just talked about, which is you get one chance or 40 chances in your whole career to produce a crop, and uh you you're not gonna just take a flyer on some guy with a shiny software package who says uh this can help. Right. Um I can try in my business, I can try Salesforce, and if it doesn't work, I'll turn it off a month later and no harm, no foul.
17:53 Speaker 2 Um, but if I kill a crop or um, you know, allow you to make some really big business mistake in when you only have 40 cycles in your whole career to make money at, that's a really big deal. So it's I mean, not unlike medicine or education or some other things where there's there are really good reasons for people being um extra careful about adopting technology.
18:15 Speaker 3 What can we learn then from I mean you raise a really good point that there's a lot at stake here for farmers when they're when they're picking a technology and they're deciding to get behind it. Yeah. It's their livelihood. Um but like you say, healthcare, um, education, other places that that need to be careful in their adoption of technology, and then sometimes we complain that they're oh my god, they're so slow. Right. But yet they need to be careful. What what can we learn from the agricultural adoption of technology or what you guys have learned in terms of like building things that could help other areas and other parts of other industries and uh adopt technology?
18:52 Speaker 2 Well, I guess it's this doesn't necessarily work for our product because we're a kind of whole farm business management product, but I think what the classic way of adopting technology and farming is I'm not gonna put this on the whole farm. I'll try it on 10 of my thousand acres, right? And we'll we'll sort of A-B test that way. That's been the traditional way that farmers have adopted new seeds or new you know, fertilizers or new management techniques. So there is sort of a um a natural, you know, uh take a small part of the farm and and try it out there approach that uh that I think is is pretty established.
19:29 Speaker 0 I also think uh the working with the universities and the extension systems that they've put into place is also how adoption happens for in the ag in the ag sector. So there are experimental farms that have similar crops and similar soil types and um that that we can do these experiments on, show that they work and farmers can then take you know have take um confidence in that. Uh and so that'd be interesting to to see if there would be these sort of forums for experimentation. That we could bring from the ag sector since extensions been so successful for so many decades.
20:08 Speaker 3 Right. That's a
20:09 Speaker 2 great point. That's I mean, that's yeah, you can't overstate how important that's been to the development of US agriculture. So we are the like the runaway leaders in in agriculture in many ways, some uh uh because of the extension process or the extension infrastructure. Um and I think um, you know, we we we need to keep that going. And I think it's under a lot of pressure from a funding perspective, and then some of the latest technology is sort of outside of the um the experience base of some of the the faculty at the extensions, and so there's I think we're at a really important crossroads where the extension university system needs to evolve and keep pace and support farming for the next hundred years.
20:51 Speaker 0 We have to, as a society invest in that, and that will make a huge difference for ag and for the just the technology industry in the ag sector.
21:00 Speaker 1 I see two main uh points that uh are going to benefit from adopting Agtech. Number one in the context of of global warming, uh, the more data we learn uh and acquire and analyze today, the better prepare we will be when temperature will rise with it be only half half a degree Celsius on average. And the second point, if I again use the example of of vineyard, we see a lot of um mismatch between one specific varietal and the environment where this varietal is grown. In in some instances, there are places where uh again blindly uh points that it makes no sense to grow uh one specific varietal or sometimes just grow, period.
21:46 Speaker 1 Yeah, right. So but it's only by using uh analytics and and and raw number that we'll be able to uh select the places and situations where it makes sense from either an ecological, societal or business standpoint to grow. And that's for the future will be very important as we can no longer uh play um take any chance with uh land use given the growing uh population. That's right.
22:15 Speaker 2 And and I think you're exactly right that it's both a science problem and a business problem. If you look at um uh Australia, that's a very um hostile farming environment uh from a water perspective, and they have a uh a fully developed uh trading system for water so that if farmer A doesn't use all his water, he can sell it to farmer B through this market. Um and and I think to for that kind of system to happen in America, you need farmers to really understand their economics and be able to say a gallon of water is worth exactly this much to me, and I would rather leave my field fallow and sell it. Yeah, exactly. If it crosses that price, and right now you can't, well, number one, water is generally free. Um, but but I think the only way agriculture evolves with the the pressure of global warming is through those types of market-based solutions, and that that requires farmers know that know their business.
23:11 Speaker 3 Chandra, you mentioned the farming for the next hundred years, and I want you guys to sort of imagine what it's gonna start to look like. Are drones gonna be out picking you know peaches in the in the trees and and our is it gonna be self-driving tillers and you know, harvesters? Um how do you imagine things? And you know, in your smart farm, Chandra, it almost sounds like the Internet of Things might hit the hit the farm before it hits our homes. But when you spin this forward, what does it start to look like?
23:42 Speaker 0 I think the biggest change is that instead of thinking about some of the things we've been talking about here, water resources, amount of land, soil quality, the resources that we'll all be talking about that are going to be the most valuable to the growers themselves is their own data. We're going to have to develop the markets and the controls and the ways of sharing data that benefit farmers because in my opinion, data is their most valuable resource. Uh and I think the future is all about that, because there's so much intelligence and insights that can be drawn from that data locally and globally. Uh, and so that's going to be a huge game changer in the future. And secondly, I do believe that there'll be a lot of automation.
24:27 Speaker 0 Um it will be uh hand in hand with you know the domain experts, the farmers themselves, but they will be probably requiring much less manpower because drones and robots and ground robots and um can take pictures, can do all kinds of interesting things, can can pick and harvest and fertilize, and they can do it round the clock, in the rain, in the dark, you know, they are not uh limited to what humans are limited to doing. And so I think that's going to change um and then our and then the crop life cycle is probably gonna change with the changes in global warming. And so different plants are gonna be need need to be grown in in different areas, and and you know, sharing that knowledge and intelligence is going to be vital. So I think it's a real combination of human intellect and and drone and robotic technology and analytics and data management that are going to change everything.
25:23 Speaker 3 Shandra, can I just ask you, does I mean one of the things we hear about, and Sid, you talked about how how many of the farms or most of the farms in the United States are family farms by sort of unit, um, then there's massive industrial scale farms. But what technology does a really good job of is lowering the cost of things. Does what you talk about change the scale of farming, or does it enable you know, small-scale farms to even do a much much better job?
25:50 Speaker 0 I think it's this that that I think it's the latter. I think it's smaller farms are now uh hopefully uh economically can be economically more viable now because they can be more efficient. Technology brings that. Uh, but the super scale of you know the field corn and the soybeans, you know, it that that scale is still an important factor to have many, many acres to succeed at. So it's really going to be a balance between those two.
26:19 Speaker 3 Not to mention feed the world's population. That's right. That's right.
26:23 Speaker 2 Yeah, I mean, I think uh what I see going forward, and in addition to all the things we've talked about, is just the continued professionalization of the industry, right? And and the continued consolidation of it. I mean, I think you can't avoid I mean it used to be that 100% of the country farmed, and now 2% of the country farms, and it's gonna be 1% before we know it, right? It's uh it is an industry in which uh people are moving out of these family businesses to go do other things. So I think as a society, we need to accept that farms are gonna get bigger. Um, and that doesn't necessarily mean they can't be great stewards of the land and do all kinds of good things for society. Um in fact, I think the more the the thing we as a society should care about is professional management of farms and whether that's a small-scale farm or a big farm, it doesn't really matter.
27:11 Speaker 2 We just want these resources to be well looked after and to get the healthiest you know product into the system. So I think I think to some extent we we get a little distracted by we love big farms, or we love small farms but don't like big farms, uh, and really what we should be focused on is is the talent and management of of the business and the use of technology. Uh and I see that you know, it's just gonna continue to accelerate. Um, the witness by the fact that we're sitting in South and Market talking about farm technology. Sitting in San Francisco thing
27:42 Speaker 3 about farm technology, yeah.
27:44 Speaker 1 Why the data are gonna be uh more and more critical, I think it's easy to understand if you look at medicine, for instance. Everyone understands the value of monitoring biological indexes over as many people as possible to better diagnose trends and statistics and predict when you are at risk of heart disease, for instance. Well, in farming, I feel because we haven't really value um uh the meaning of such data, it's a little bit like medicine in in the 17th century where when you have a problem, you're gonna get a bleeding and and hopefully it's gonna cut it
28:16 Speaker 3 off. Yeah. Cut it off.
28:18 Speaker 1 And really, in in the future, uh it's only by pulling together uh data that we're gonna increase our ability to diagnose the needs of of plant as if they were individual. But your uh point uh seal about uh farming unit getting uh ever uh a greater size is also going to raise the problem of viability. Where do we get the data from and how do we extrapolate the data to diagnose uh the best way to farm according to this intrinsic variability that can be seen at the plant level itself or at the unit of production itself? And and it's it's it's posing a big problem on how we're gonna extract the value of this data. So it's not only a matter of the meaning of the data, the data in itself has a lot of meaning for future use, but how do we get to that data and how do we make sure that this data is meaningful, which depends upon how do we deal with variability?
29:16 Speaker 2 That's right.
29:16 Speaker 3 Again, I think there's a lot that we can learn from agriculture as these um questions start to get answered, and as technology seeps more and more into the fields and into the vineyards, and um you guys are are at the front of this, and um it uh it'll be fascinating to see how it plays out you know on our plates and in and in our economy. So I just want to thank you all, Chandra, Thibault, and Sid. Um, we will see you on the farm or in the vineyard. Thank you.
29:43 Speaker 2 Thank you very much.
29:44 Speaker 3 Thanks, guys.
Previous Company Building in Crypto: A Conversation between Brian Armstrong and Chris Dixon Next What's Next for Education Startups Page updated August 25, 2026 · transcript and summary are machine-generated