[00:00:00] Episode Intro

[00:00:00] Audrow Nash: What do you think when a robotics company comes into an old industry? Do they shake it up? Tell everyone that their old technology is out of date? Make things so complex that only people with PhDs can understand what's going on? I think that's the fear of a lot of people, especially people outside of robotics.

But it's not always the case. In this episode, I speak with Ben Alfi, CEO and co founder of Bluewhite Robotics, a robotics startup in agriculture that just closed their Series C investment round. Blue White is doing something different. They're trying to blend in. They want to make it easy for farmers, so they can use their existing equipment, work with the same dealers, and do things their way.

All while adding robots to make things more efficient, and especially Get the work done in spite of increasing labor shortages. You'll like this episode if you're interested in robots and agriculture, the ethics of disruptive tech, and a clever business model that creates opportunities for more jobs. I'm Audrow Nash.

This is my podcast. I hope you enjoy my interview with Ben Alfi.

​

[00:01:09] Introducing Ben Alfi and Bluewhite Robotics

[00:01:09] Audrow Nash: Hi Ben. Would you introduce yourself?

[00:01:12] Ben Alfi: Hi, I'm very happy to be here today.

My name is Ben Alfi, and people call me Alfi, too many years in the Air Force, so we've been called by your surname, and I'm 50 years old, a very young entrepreneur though, only six years in the startup ecosystem.

[00:01:37] Audrow Nash: Wow.

[00:01:38] Ben Alfi: And, CEO and founder of Bluewhite.

[00:01:43] Audrow Nash: Yes. And tell me about Bluewhite.

[00:01:45] Ben Alfi: Bluewhite is a data driven, autonomous farm company. we are, healing and, assisting how to, create disruption around, autonomy in the agriculture market. And the way we do it, we are, bolting in existing tractors, transform them to autonomy. Adding a kit on it. And, that way one person can operate different type of detractors, in his farm, with an iPad or a laptop or a cell phone, whatever he needs.

And the ideas that, or the labor shortage today in the market and the idea of, food is getting more and more expensive. Bluewhite is creating a model that enables, The adoption quite fast by, dealing with the existing, people, existing people you have, and the existing, tractors you have in the field.

What's nice about what we're doing is that we're focusing in, permanent crops, meaning vineyards, citrus, apples, almonds, and such alike, what has been called as high value crops. So it's all year long. A lot of tractors per acre at the same area. So these are the places that the labor shortage is very high, and the need, and the demand is very high.

And on the technological side, which you are, very keen on, you and the listeners, these are areas with no connectivity, no GPS. So relying on standard GPS, or standard connectivity won't get you anywhere. And so we have, nice technology, which I'm sure we'll dig into along the discussion.

[00:03:51] Audrow Nash: We definitely will.

[00:03:53] Outfitting a tractor with Bluewhite's tech

[00:03:53] Audrow Nash: Okay. going back just a little bit, you brought up a lot of very interesting things there. First, you mentioned it's a kit, so you're bolting it to existing tractors. Tell me a bit about why take that approach and what that actually means. So you have a, box, you're just strapping onto tractors and then integrating into the system, or what does it mean?

[00:04:18] Ben Alfi: it means that, we are a commercial autonomous company. And so today the go to market is a dealership of John Deere or New Holland are selling and installing tractors, Bluewhite capabilities. And, how does it go? You were sending like an Ikea kit, to a dealership. Dealership has an, an order for, from a grower that he wants, his existing fleet to transform to autonomy.

They take his tractors. In the morning, the tractor is totally manual, nothing happened on it. And in the afternoon, it's autonomous. After two people worked on it for, A few hours and after checkups, after all the integration, after all the sensor integration and everything, from actuation to being connected to the cloud, so you can start operating it and that's about it.

This is just the first phase. The second phase is how to implement it in the form itself and this is also something that we're helping or the dealership can do it also.

[00:05:34] Audrow Nash: Okay. So tell me more about this kit. So when these two people are installing it in just a few hours, what are they taking as a component and then, they're, how are they connecting it to the tractor or whatever equipment going to automate.

[00:05:55] Ben Alfi: it's, it depends on how smart the tractor is. Most of the tractors today are just manual tractors with no computer on it. Some of them already are digital in some places. And some of them in the future will be drive by wire, so you can actually, can connect to some kind of a computer. Work It. And the way it's done today, the kit includes actuation, the gas brake, the digital electric gear, and another component will be the sensing, front sensing, and other sensors that will go and to understand what is happening, so it's the actuations are the muscles, The ability to move the muscles, the sensing is the ability to sense and understand what's going on around me.

And you cannot depend just on one sensor, it can be a GPS, LiDAR, Visual, and others. Because we're doing sensor fusion. A set of communication. And compute, because you need also onboard computing, because The tractor, if it doesn't have a connectivity, it needs to know how to drive safely in a good way, even without somebody looking after

[00:07:18] Audrow Nash: Very cool. It's really cool to me that you are outfitting tractors that may have no digital systems in it, like purely mechanical

[00:07:28] Ben Alfi: it. Most of them are today. We're guessing in five years it will be different, but most of them are just without anything. Maybe some kind of a CAN bus to identify faults like, oil status or fuel gauges or other things like that. I don't know. But most of them are really classic mechanical tractors.

[00:07:56] Audrow Nash: And then, so you'll have I don't know, it'll, it probably looks like a cam or like a little foot and you put it on a servo, and then that servo can push the accelerator and push the brake. Another one will push the brake. So part of these, the technicians installing this part of their work is to place these motors in a place that they're actually accessing the controls.

Is it correct?

[00:08:21] Ben Alfi: You're doing it on top. I look at it as if it's like a handicapped vehicle. And this is, But it's, these are, these are off the shelf capabilities, the big things are the algorithms and the being smart behind it and understanding how to operate with the same software and all those different type of tractors, different type of crops, day and night, and also different type of implements, whatever is in the back.

[00:08:52] What types of jobs are they automating?

[00:08:52] Ben Alfi: We're not just tractoring, we need to make sure the job is being done.

[00:08:58] Audrow Nash: Yeah. What, so not just tractoring, what other jobs or what other vehicles are you automating?

[00:09:04] Ben Alfi: So it's, when we say not just tractoring, it means that the tractor needs to mow, to spray, to herbicide, to pesticide. Whatever is at the back over there needs to be very accurate in it, if it's working correctly. We are saving so much money on chemicals, so much, to the earth on putting just extra chemicals on, the vineyard or almond trees, our sensors understand if there is or there isn't.

a tree, so you can stop the sprayer, start the sprayer. Is it a big tree, a small tree? is it the end of the road and now you just need to stop it? today it doesn't happen. Today people just spend more and more, chemicals and it's not good for the people that are driving next to it and dying from cancer, too early.

and it's not good for the environment.

[00:10:00] Audrow Nash: Definitely. Yeah. I do think that there's a lot of opportunity for robotics to improve environmental things by making it so that you don't have to do as much fertilizer or pesticides or something like this, because you can have more targeted. Use or through automation and also maybe more persistent use, because you can have it running more frequently versus having to spray more heavily because you're gonna run less frequently because people are scarce or something like this.

[00:10:29] Ben Alfi: And availability, you're doing it not at the correct timing, or the weekend costs you 150 percent because of the labor costs, or nighttime is too costly, and then you are not doing it at the correct time, so you just over spray, overuse of your chemicals, and also, Think of the amount of tractors that you need in the farm, and we are saving around 35 percent of the amount of tractors you need because they can do double

[00:11:01] Audrow Nash: That's so cool.

[00:11:02] Ben Alfi: And we're saving 85 percent on the chemicals because we're doing it accurately.

[00:11:08] Audrow Nash: Okay, so that sounds wonderful to me. Going,

[00:11:12] How do you do localization? What sensors do you use?

[00:11:12] Audrow Nash: I wanna, understand the system really well, before, and I want to get into a cloud infrastructure and everything that you guys provide. but so you have actuators that are moving the tractor and controlling things. what kinds of sensors do you typically use?

'cause you mentioned you're often in GPS or connectivity denied environments. what are you relying on? And you mentioned sensor, fusion, but how, are you approaching it from a sensing perspective?

[00:11:41] Ben Alfi: And I think this is a key question, also a key way of how we're approaching it.

We're talking about autonomous vehicles for quite a long time. There are a few, very few companies that can say that they are commercial in autonomous vehicles, like Bluewhite.

[00:12:00] Audrow Nash: Definitely.

[00:12:01] Ben Alfi: being commercial means that you need redundancies and you need to be safe, first and foremost.

It's, easy to demonstrate a tractor running in the farm or in the field, but it's totally different to have 50, 000 hours of those tractors running around day and night, a few tons tractors with, with the grower. So how do we do that? It's all about redundancies, and the way we are creating those redundancies is by creating parallel navigation solutions.

It can be a navigation solution done by GPS and RTK, where you have it. It can be a navigation solution by a LiDAR, which is an amazing broad sensor. A lot of people invested through the urban mobility companies, who invested in so many sensing and we're taking those capabilities off the shelf to the agriculture space.

LiDAR sensor, visual sensor. Odometry and other tricks that we have added inside. And think of it that you are driving and all the time the computer gives you four types of solutions. So you can have decision making who is now incorrect because they are checking each other. And if you see an abnormality, You can decide who should be the main navigation system.

For example, a vineyard during autumn, and it's open skies, a good reception. Give the GPS and RTK, let them be number one, and keep LiDAR and visual as obstacle detection only. Almond trees, August, 120 degrees Fahrenheit in Fresno, California, under the foliage at night. No reception, no nothing, so you'll give the odometry and the LIDAR, you'll let them be number one and two.

And then you'll put the visual number three. And then, only then, you just, not rely, but understand where you are in the other one. So this is, and all this is happening automatically. You as a operator are not deciding what is, because the operator is You know somebody who doesn't understand technology.

Everything needs to be very simple. Stop, play, Spray, mow. This is the speed I want. This is the block I want. Other than that, we are counting on, the machine to be safe. So the machine automatically knows, hey, I'm in an almond block, it's summer. This is what I see. I understand what should be the prioritization.

I will give certain prioritization, but if something is changing at real time, I will change prioritization accordingly. And unlike urban mobility, at the end of the day, it's only, up to five miles per hour. So I just stop.

[00:15:14] Audrow Nash: yeah, that's a big advantage that makes this problem a lot easier, even though it's super hard, is that you can just stop. something's weird with the connectivity just stop and it's not that big of a deal.

[00:15:27] Ben Alfi: It is a big overdeal on the quality of work and the idea that we're not stopping a lot is great but on the safety issue it's much a bigger of a deal if you have a safety event.

[00:15:40] Audrow Nash: Yes, for sure.

[00:15:41] Ben Alfi: this is how we're working. Okay,

[00:15:44] Audrow Nash: So you have several different tracks that are all figuring out localization independently, and you can figure out which of those you wanna listen to.

so I'm imagining like a weighted consensus for figuring out where you are. How, and you mentioned that it changes based on certain factors. if you have the almond trees and they're blocking out GPS or whatever it might be, then you switch to just using sensors on the robot. How are you picking between these?

'cause you mentioned that the farmer doesn't have to do it. Is it like you do a, an evaluation of the environment and then you use the environment and your sensor availability and maybe some other heuristics like time of day to select a weighting for these different localization types? Or how do you, pick from there?

[00:16:38] Ben Alfi: so I've tried to simplify it, but it's complicated. And so that, so the way it's happening is that you can have a basic assumptions upon time of day, where I am, what is the field, what is the mission, spray, mow. If I mow, there will be more dust. If I'm spraying, there will be more humid, things like that.

what do I depend on? After that, how old is the, the orchard? Is it a young orchard? An old orchard? A lot of things like that are happening also. And also in the algorithms that we're using, should we rely on the classic algorithms or the AI? The AI is much more flexible, yet less predictable. It needs more maturity.

So these are the balance. Do I have an AI machine, AI algorithm that is running? Well enough and mature enough and got a good scoring already that I can give it, let him be a part of the decision making, or it will be still piggyback, just riding along and collecting hours until the truth, true false and the false truth will be, in a good way.

So it helps, AI helps for scale. Classic algorithms helps for starting to run, and these are balances, and what is nice that we, it happens in a way that it's transparent, so you have around four types of sensors and around, let's say, anything between eight to sixteen types of solutions that are running, and that way you are able to run it, run.

Some of them is based on pre assumptions, but real life scenario, what actually is happening can change. The assumptions in real time. And then at the post-processing, the planning for next time will be okay. This block, the pre assumptions were wrong. These are the pre assumption that should happen.

[00:18:39] Audrow Nash: That seems very cool.

[00:18:41] Adding sensors to the tractor's tools + levels of support

[00:18:41] Audrow Nash: Okay. And so you have your tractors have these actuators to control them, and then they have these sensors that you are using to figure out things like localization. and then you have your implement, which is whatever you're towing. It's the mower, it's the sprayer, these kinds of things.

Are you putting sensors on the implement, on the sprayers? I'm calling it the right thing.

[00:19:08] Ben Alfi: Yes, you are. Yes.

[00:19:09] Audrow Nash: okay. Just making sure. so if we put it. Are you putting sensors on the implement or how are you evaluating what, how you're doing, with the implement? Or do you just have a model of what it is that you're towing and when you turn it on and off?

Or how sophisticated is your control of the implement?

[00:19:29] Ben Alfi: So just like the relationship we have with the tractor, is that we don't own the tractor. The tractor is the growers or the OEMs, and we are, blending in. Same is the idea with the implements. and the implements were we, because we're commercial, the benefit for the implement pro, providers or makers is that they want to work with us and we provide APIs and help them how to make their implement smart.

There is a sprayer and there is a smart sprayer. And so if I have enough information, and what, we're doing, we, have, four grades. That's four levels of how smart an implement should be. First one is on and off. Can I operate, can I just switch it on and off? Then can I, how can I assure the quality of the work?

Can I just The altitude of the mowing, the mower, with a sensor, yes or no. Sometimes the, the company that created that mower, they will talk to us and we'll help them how to adjust and what should be implemented and they will do it. And some places we'll just add a sensor in key areas. for scale, we see ourselves more of helping those companies creating this next generation and make them, what we call the implements.

Bluewhite Ready, so it's on off, sensing on the quality of work, what we call also about, we talk about preemptive maintenance, if I know that I can drive the certain PTO level, a certain engine, RPM, and it drives through miles per hour. And now I see that I need to have more force and nothing has really happened.

Okay. What is happening over there with the mower? What is happening over there with the implement? And then the last thing is the ability to report and to also control, control nozzles. Certain nozzles to shut off, certain nozzles to open. These are things that we're doing. And in that way. And working with the implement companies or enabling by ourself in certain areas.

This is how we're doing those capabilities.

[00:22:00] Audrow Nash: Okay. I like that a lot. And those are very exci, like the. The maintenance one in particular is very, exciting to me.

[00:22:06] Ben Alfi: It's huge. Real time for safety, what we call critical understanding. It's amazing how we are neglecting the understanding of a person that is running. Just understand something just went wrong. And how do we understand the same thing? Noise, vibration. and something is stuck. I need more power, all kinds of things like

[00:22:32] Audrow Nash: anomalous? Yeah, something's different. might've happened. Yeah, I think I, that's one of the huge gains, I think, from automating these systems. So it sounds to me like the core competence of your business of Bluewhite is to create this mobility kit for farm vehicles and then an added thing.

It's almost like a, side business within it. Is to expose, expose what you're doing and give yourself systems so you can control implements and you can help other companies to make their implements work well with your systems. and so it's outside of your work, but other companies can be lucrative, can have lucrative jobs of just retrofitting or designing their systems to work well with your system.

[00:23:31] Ben Alfi: first of all, we see ourself also responsible to, to make it happen with

[00:23:35] Audrow Nash: to help them,

[00:23:36] Ben Alfi: and I think

[00:23:36] Being a full solution for agriculture companies

[00:23:36] Ben Alfi: there is a huge, element that we disregard until now is the operating system.

[00:23:44] Audrow Nash: Ah, okay.

[00:23:44] Ben Alfi: So this is why we see ourselves as autonomous, data driven autonomous firm and not a aftermarket kit for autonomy. What does it mean?

It means that in a certain farm, you don't have just John Deere, you don't have just New Holland, you don't have just one type of tractor, you have different types and you don't want as an operator not to have six type of operating systems because you have six type of tractors, or implements, or crop. One operating system knows how to operate everything.

And the last layer is that those tractors are running. Sensors are there already. They collect a lot of data. There is no reason not to share that data with the grower. With the, farm manager, with the agronomist, with whatever agriculture company that the grower wants to work with that wants to do yield prediction or, weather or the his insurance company that he wants to show them that his farm is working correctly.

the ability to be also data enabler for the grower, for ourselves, to give him operational insight, to share it with third party that can give him also agriculture insight, all that. This is how we look at ourselves as a full package.

[00:25:06] Audrow Nash: Yeah, it's the whole vertical stack of farming using large machinery. Okay. Yeah. Okay. I love that.

[00:25:14] Sending data to the cloud

[00:25:14] Audrow Nash: Tell me more about the cloud component. what's actually how, I guess the first thing, 'cause I going up the stack. So we started from actuators, then we go to sensors, and then implements included in that.

Then after, it's how are we sending data to the cloud? Like how, does that actually work? Where, 'cause you're in these connectivity, design, these connectivity denied environments. Is there like a home base that the tractor returns to where it has upload ability or how does, how does this work?

[00:25:47] Ben Alfi: This is, one of the biggest things that we didn't know that we didn't know. and I think we have, achieved, first and foremost on the system engineering. And the second is how to actually make it happen. What is the relationship? of a cloud based operating system and, what we know from other markets, our IOT device.

So if we use a look at this robot as an IOT device, one thing, what is needed to be computed On the tractor. Okay. Now, how do I also relay, okay, you are the user you're operating. It was under the fully is running all that role for half a mile. Okay. What, how do I know what happened? When should I know?

Should I know it at real time? Should I do an assumption and then upload it? How do you also cache it correctly to the operating system so it won't be an overlapping of what really happened. So this is the coordination that we've done. There are some things that are, we've created a section of what is really, important.

Where are you? Or stop start. I want you to stop now. I want you to start. Okay. Then what, do I, okay. how is the, are you healthy? what is going on? Do you need my help? what is really happening? Oh, the data of the camera. to collect it. Collect it now on the, on your cell, on your onboard computer.

But when I need it. Upload it when I need it or when I ask for it, or at the end of the shift. Not all the information is important right now. And black box it, say it is a machine, something will happen. I need a black box to understand when a fault has happened. Do a record. I need you to record all the time and track what has happened so I can extract, logs and all those layers of information.

So some of it will be on the cloud. Some of it will be on the endpoint, on the computer itself, and it should be synced in a way that it's not just overwriting, and I know exactly from which tractor, and also privacy wise, also security wise, and all those layers.

[00:28:16] Audrow Nash: So it makes sense you are prioritizing what to send out first given your connectivity and things like this. when you were mentioning the black box, I was a bit confused. I, but I think what you're referring to, it's like in, airplanes, they have that box that's supposed to survive no matter what.

So you have, there's an analogous system that records logs and things like this on tractors,

that

[00:28:43] Ben Alfi: Although there is no definition that it should be. we are coming back from 20 years of autonomous vehicles, mostly air vehicles and others. And we believe that this is how it should be. So we created like other urban mobility, standards and, agriculture standards. So this is a great way to understand what has really happened, that I can also pull on demand information.

And to, troubleshoot, understand what is happening. And also, if something has happened, to know why it happened and where.

[00:29:24] Audrow Nash: Yes, definitely. And then, so if you're in the situation where you're operating in a fully internet and GPS, denied environment, so you just don't have access to those, maybe you then just tell the user, at first you are operating in those kinds of environments

where you have just

[00:29:44] Ben Alfi: And so we create, yeah, we created local 5G networks as an example, worked with Intel at the past.

[00:29:52] Audrow Nash: ah, do you put 'em throughout the orchard or something like that, or

[00:29:56] Ben Alfi: We try not to spend too much money on infrastructure right now, and also to the grower. You need it to be cost effective. Each farm somehow has an internet connection. It can be also at the farm manager's office. And from there you are like linking with an antenna if needed with local 5G networks. And we're working with third parties to enable that.

We will see more and more symbiotic, Private and public networks running together. And again, it should be transparent to the tractor and should be transparent to the tractor operator.

[00:30:41] Audrow Nash: That's really cool. Okay, so you'll, where there is GPS denied not GPS, maybe you don't need gPS but is

denied, you'll provide some infrastructure and maybe, the robot can go in and out of connectivity even still, but you still are checking in occasionally and prioritizing things like I can turn it off if I need, or, so it doesn't drop out for like huge of times.

It's just a, second here or a few seconds there.

[00:31:09] Ben Alfi: exactly, and you can also decide, how you want to, are the rules? Okay, if I have no connectivity, should I stop? Am I allowed to get until the end of the row? Before I turn at the end of the row, should I stop and wait until there is connectivity? Is it a place so the grower can decide what are the rules of, okay, how free is that place from any other, any others just to run and play?

[00:31:43] Configuring their robot's behavior + onboarding

[00:31:43] Audrow Nash: So what strikes me with this is that the growers often have a lot of configuration available to them. so if they, wanna set the rule, don't drive this far without connectivity or something. 'cause I wanna be able to shut you off at any time. Or, I don't know, maybe there's more specific, like farm procedural stuff.

One farmer wants to do it this way versus another wants to do it that way. How do you manage and expose them, the growers to that kind of configuration? I imagine it's like customer onboarding and it's a one-time

[00:32:20] Ben Alfi: Exactly. Exactly. Spot on. So it's the customer onboarding. every in orchard you spray around 20 times a year for 20 years.

[00:32:31] Audrow Nash: a

[00:32:32] Ben Alfi: the idea is just, these are the efforts at the beginning. and also we have, our, our home recommendations. So I want it to be with recommendation and automatic, okay, it's an almond orchard.

This is how it looks. These are the automatic, but I want also the grower to have the ability to be flexible if used is used to keep every two rows. So what, I will give him a planning, but it's automatically only every row. You just tell me how you want to do it. You want to skip every three rows, every four rows.

You like different speed. Anything that you want should be addressable. And again, he has either a Bluewhite team that is on the support online that can, he can get more, help or, also dealerships that are helping him on the onboarding to do it. So he's not alone. Okay. This is the main idea, not to be alone, but, tell me what to do, but also when I want to do something, let me have it.

Let me do it. And as long as it's, safe. And it doesn't, infringe safety, we will help.

[00:33:43] Audrow Nash: It's all good. Yeah. So when, you send out a person or a small team of people to onboard a new customer, is it, is it like engineers are going out

or is it

[00:33:58] Ben Alfi: not.

[00:33:58] Audrow Nash: are. Okay. So a lot of this is exposed at a high level configuration, and so you have that are

[00:34:06] Ben Alfi: It's, operators. it's even just, just operators. we used to operate in the past. Now we're moving more from a service company to a product company because it's mature enough. And, but the know how on how to do it and, to give, to help adoption, to help the adoption and to help, to help, the growers, to use it in a good way.

We're not talking about, a year of work, it's, days and you're good to go. Over a day, it depends on how we have growers who have, who have 100, 000 acres, okay, of orchards. So it depends on how big you are

[00:34:52] Audrow Nash: Yeah. Oh man. That's awesome.

[00:34:54] Working with farm equipment dealerships

[00:34:54] Audrow Nash: So it sounds like these dealerships are really playing a critical role for you guys where, so you were doing this yourself, but now you're letting them handle the setup and you're turning into more of a product company. tell me more about the dealerships

[00:35:11] Ben Alfi: and all the market is under transition. Think of it. Those leadership have sold metal for the last century and now they need to be a precision provider and there is a lot of discussion about it. Okay. How?

[00:35:29] Audrow Nash: the sold metal. That's

[00:35:30] Ben Alfi: Yeah, it was a big question of, are we, when a dealership is installing a kit on a tractor and we want to see that every section was done correctly. Will he be opening a Jira ticket? Yes or no. Can he do it? Yes or no. And we found that. And the dealerships, they understand that they must be able to do it. They must, and in a way we are working with them together hand in hand on how to do it correctly. So the technicians are also, it's not a software technician, it's not DevOps.

But it's, what we are starting to see in the world as a robotic technician. So it's, 95 percent of their job will be on the mechanical side. or concept of operation side, but when needed, yes, you open a laptop and connect.

[00:36:33] Audrow Nash: Wow. How, how large is your operation with everything? Like, how many dealerships are you working with? How many robots are you deployed

[00:36:44] Ben Alfi: we are just, the first years, we are now, seeing ourselves growing more and more. we have around, 100 tractors running around already, and there are 400, 000 tractors waiting for aftermarket, just in permanent crops. Our goal is to get to 10, 000 in the next few years with those dealerships together. is an amazing experience. Main idea is how do you grow while making sure that safety is a top priority. How do you grow where gross margin is important for yourself, for the dealerships. And also for the growers to have a positive ROI, we have all that fixed and now we're just scaling more and more capabilities and it's like a spiral development and more like the metrics in a way that, okay, I didn't know how to do, how to run in apples, trees, and now I know, I didn't know how to, do spraying, now I know, so another capability, along the life of the project.

[00:38:00] Audrow Nash: Yeah. Okay.

[00:38:01] Spiral development

[00:38:01] Audrow Nash: And you mentioned spiral development, so it means gradually. So can you describe the spiral to

[00:38:08] Ben Alfi: Yeah, sure. it's, it's,

[00:38:11] Audrow Nash: as you're going.

[00:38:12] Ben Alfi: yeah, it's an, it's an agile, process. yet you can see it in a long term roadmap. Okay. how, okay. How many type of tractors do you know how to, transform? So we know around 20 already. How many type of, crops do you know? So we started with almonds, then pistachio, then.

Trellis then citrus, then apples, then all the others. How many types of implements We started with spraying, then herbicide, then mowing, now we're doing harvesting with the harvesting company. So all those capabilities that you're adding And it's great because we are transparent also with the dealerships, transparent with the growers.

They know what to expect along their upcoming years. All the updates are over the cloud. and it's, again, in a way that you can enable progress without buying a new iPhone every year.

[00:39:22] Audrow Nash: Because it keeps shipping updates for this kind of thing, so they don't have to necessarily buy more tooling. I would even imagine that if, say you upgrade or something, you can just add more compute. You can leave the sensors, you can leave the actuators. And so the upgrades may be somewhat painless.

[00:39:42] Ben Alfi: Yes.

[00:39:43] Audrow Nash: like relatively in terms of cost

[00:39:45] Ben Alfi: The idea

[00:39:46] Audrow Nash: back into the same system.

[00:39:47] Ben Alfi: On the hardware, we're trying to maximize as much as possible. Those sensors are amazing, and we can maximize much more. I gave you an example. We started with using the LiDAR for navigation, and now we also know how to, with the blue spray, we know how to save on spraying, because while you navigate, you know that there is a missing tree.

Let's tell the sprayer to stop. these are examples of how you create more and more applications along the years with the same hardware. And our goal is not to change hardware as often as others might be. We just don't want to go through that effort.

[00:40:28] Audrow Nash: Oh, for sure. Yeah. Especially if it's already working. Makes sense. Just keep going.

[00:40:32] Scaling their operation

[00:40:32] Audrow Nash: Now tell me about scaling. your main, so you've gotten a hundred tractors operating how many hours?

[00:40:41] Ben Alfi: We've done more than 50, 000 hours already of autonomous running. Our growers have 300, 000 acres. Even a grower doing a roadmap with a grower. A grower would have anything between 10 tractors to a robot. We have growers with 400 tractors. So you'll go with four tractors and then 12 and then go up to 80 percent of the fleet.

You don't need all the fleet. And so this is the idea of how we're, so it's a land and expand mode in a lot of places. we've been in California. Then we, now we're also operating in Washington state. We're going to, we're going to be in Europe, I'm guessing by the end of the year, Checkups over there and 2025 with more goals over there.

And Australia is also with huge demand and asking us to come. I just find it hard to be awake on three continents at the same time. So we'll see how we're growing. So the need is huge enough. The market is totally huge. There's enough room for another 10 Bluewhites in the world. And this is how it is.

[00:41:57] Audrow Nash: to 400,000. Yeah. And it'll be exciting. How, just for more context, how quickly, how has growth been looking? I imagine it's on an exponential curve, like in the next year, I would imagine maybe you have a thousand tractors or something like

[00:42:14] Ben Alfi: It depends what will be the, method that we want to do. We are emphasizing ourself, this year. We are going out now with generation three. Now, generation one was to take, the person out of the cabin, where you still seeing other autonomous companies in the world and in other areas. You still have a safety driver in the tractor.

Generation 2 was how that you don't need to look at the tractor, it can just run and it will work and it will know when to stop or run and everything you have the redundancies. 3 that we are now going out with is all the tractors are operated by the customer and to be operated by the customer we are looking for all the feedbacks that will happen along the year and getting more and more understanding of what is needed.

Then boom, we can, we can send a kit to a dealership in Australia. We don't need to go there over the cloud. We know how to upload. He is doing the onboarding and we're done. Okay. So it's more of, stages and maturity and then you can keep on running.

[00:43:27] Audrow Nash: I see. So the, if I understand correctly then the scaling challenge at the moment is you are letting in only enough customers that you can handle the feedback and then as you get ready, you'll take on more and more until, basically it's, biting off what you can chew right at the moment.

[00:43:47] Ben Alfi: And in parallel also creating the, how do you educate a dealership to do all that work? How do you, deal with. two sides, the customers, the dealership, and also your, kids and, meantime between failures, meantime between losses and things like that, that you want to make sure that you are maturing it, and then you can, scale it up.

part of, this last round that we just closed, round C was, is around that area that we can, okay, be, go to scale to the, just like you talked, go to those, big numbers. And, getting it from zero to one is a one obstacle, then one to 10 is a huge obstacle, 10 to 100, and from 100 to 1, 000, this is the, area that we're dealing with

[00:44:42] Audrow Nash: Every change in magnitude. Yeah, for sure. Very exciting.

[00:44:46] How large is the tractor market?

[00:44:46] Ben Alfi: Just to get an, a guess of how, big of a market we're talking of, to operate a tractor today in California costed the grower 100, 000 yearly costs.

[00:44:59] Audrow Nash:

[00:44:59] Ben Alfi: Okay, 400, 000 tractors are running, and think about the yearly cost that they're charging. So in the US it's more expensive, in other places, in Europe it's also as expensive.

So think about how, much money we can reduce to the world on the operation cost, on the chemical cost, on the amount of tractors, and in that way also maybe to make sure that you and I can keep on eating food in a reasonable price. And because food is getting more expensive, almonds, good food is just getting more and more expensive. we just need to control that. And on the other side, the amount of money that the growers are willing to pay us because we save them money. So it's easy for them to say, Oh, you give me 70 percent a cost reduction. Yeah, sure. You can have 30, 000 a year on the tractor. No worries. Okay, so this is how big the market is.

[00:46:01] Audrow Nash: Do you think that kind of savings, like 70% is feasible or

[00:46:09] Ben Alfi: now. I think it can be more. And this is without dealing with the yield that the data can give you on even improving yield. We, but we just, our approach is. And our approach is first and foremost, let's cost on, let's take down the operation costs and because you can compare it and then talk about yield, because I can, okay, who will pay until yield is proven?

This was the big question. And,

[00:46:38] Audrow Nash: mm-Hmm.

[00:46:39] Ben Alfi: it's season and every time. So first of all, when you have certain costs and immediately downgrade the cost. Automatically, the grower sees it. We will

[00:46:52] Audrow Nash: That's really remarkable. What do you think? So as you guys get more economies of scale in what you're doing, it so that the dealerships are the ones interfacing mostly with the customers. So you're not, and the dealerships are not, fielding you many questions because most of the things the customer wants you already support As you get that level of scale.

what kind of reduction in tractor expenses do you expect? Do you think it could be like 90%? Like where's the upper limit of this,

would

[00:47:26] Ben Alfi: it's about trust. The more the trust and the adoption will be, it will be higher. We'll see, less and less. Our assumptions are around 85 percent on cost. And around 30 to 40 percent on the amount of actual tractors that are needing, you need them to run around. And this is a huge amount of capabilities.

It's the only way. Also, we're going to be 10 billion people in 2050. And the way we are making food today is just not enough to feed everybody. So this is, a critical part.

[00:48:14] Labor shortages and people not wanting to be farmers

[00:48:14] Audrow Nash: I agree. And then also, I getting into kind of the labor discussion on this, I think we're have the same opinion with this, where it's, we just don't have enough people for these roles. there's such significant labor shortages that were, a lot of farmers are probably trying to hire people and unable to.

a lot of people don't want to go into farming and don't wanna be tractor drivers and this kind of thing, but can you talk a little bit about the labor

[00:48:42] Ben Alfi: You know how to ride with the shift gear?

[00:48:46] Audrow Nash: Yeah.

[00:48:47] Ben Alfi: Okay, and most of the people at your age already don't know. you need to press a clutch to drive a tractor. We don't burn a clutch. Young people, they don't know. They just don't know how to drive a tractor and not only they don't know, they don't find any reason to drive for eight or ten hours.

with the chemicals on them and suffering or being at, the high sun and suffering from that part. So it's not, it's, we see even, in, Latin America and any other place, you don't have people who wants to do that job anymore. and the baby boomers are now starting to get the intention.

They are starting their pension. They are, they were the last.

[00:49:42] Audrow Nash: than half of them are retired now

[00:49:44] Ben Alfi: And, and they, yeah, exactly. And they were the last tractor operators and running. and the amount of tractors that is needed is huge. So I think, I think just in Washington State there is a lack of 30,000 tractor drivers.

[00:50:06] Audrow Nash: Wow.

[00:50:07] Ben Alfi: And the prices are, they're sold above 20 an hour a long time ago to drive a tractor in California.

So these, and the quality is not good enough. And when you want a person to drive 2 miles per hour, 8 hours, the same speed, okay, he's unable. And downtime, you need to stop, you need to, to have a rest, to eat and everything. in an eight hour shift, we found out that people are driving, the tractor is driving around four and a half hours.

When it's an autonomous vehicle doing the same thing, it's seven hours and ten minutes because you just stopped for refills on the

[00:50:56] Audrow Nash: Ha That's

[00:50:58] Ben Alfi: You don't stop for a chat. You can coordinate when, the refilling will be. So it'll be always one in refill and all the other three are still running. And somehow when it's manned operation, everybody's gathering around at the same time for a refill and, to have a chat about, Sunday days football.

And so this is, this is what's happened and. So it's not just on the quality, on being accurate, when you're asking a robot to drive two miles per hour drivers, exactly, it's if you want it to, when there is an abnormality, you react immediately, you don't postpone it, availability, all those issues are just huge, and what is also beautiful, is that we see people coming to Bluewhite, To Fresno, to Central Valley, to Washington State, to Yakima.

Coming from Seattle, coming from San Francisco, from the Bay Area, saying, Hey, I love this. I can work with the iPad. And diversity wide, it's not just men. We have men, women, handicapped, and everything. And with the ability to operate, and when we're talking to growers. They say, at last, there is a reason for my son or daughter to come back and work with me and not to send them to, to work in Amazon in Seattle or something like that.

So

we are transforming the new blue collar of 21st century from pilots and drivers to robot operators. And we see it live in front of us. This

[00:52:42] Audrow Nash: That's so cool. What do you, so I guess what role will the farmers have in this? So there, I, love the vision. where, is the person adding value to the operation in this, the farmer?

[00:53:01] Ben Alfi: is, the autonomous farm is not a farm without people. it's a farm with people that have robots that help them, and they have data system that help, that gives them recommendations. but bottom line, the decision making, the hunch, the what is needed, and it's a database, decision making.

And, this is what's happening. And we see it in other places. We see it in our classic day to day other job that we are doing and that we get recommendation. If it's how to drive from one place to another, yet you still have, you still drive, right? And other things like that, the idea is that we want those people. who have so many skills and hunches about how the farm should run and have the ability to just deal with this and not deal with, okay, the simple things of just drive that block of 200 acres, two miles per hour, we shouldn't put a human on that part. Put them on. Okay. Should it be today? Should it be tomorrow?

Let's look at the, Oh, it's going to rain right now. Okay. So let's go and do this and that. These are the areas where we see humans. Bring the maximum factor and it will be like that also for quite a while.

[00:54:29] Audrow Nash: Yeah, I like that. So they're the decision makers, and I imagine there's also a lot of tasks that are just hard to automate around the farm that plenty of people will be involved with too. But they can shift from also doing the tractor work to just doing these very hard to automate tasks. And it makes the.

Labor gap a little smaller with

[00:54:52] Ben Alfi: point. Great point. I will use that also when I talk the next time. Because yes, what we, for example, we insisted to keep the seat vacant. That you can in the switch of a button move from autonomy to manual. Because you cannot do everything. Or you, I don't know, you want to do it, take it there.

Cross it, through a, paved road, and to bring it to the other side of the block, to another place that is without, it's not, out there. all things like that. The ability to go from manual to autonomy, the ability to do those, hard tasks or those tasks that we talked about that are yet to be accomplished? definitely yes.

[00:55:43] Audrow Nash: Do you think that we're, so what strikes me is, so I, think this is an incredibly smart approach where it's you allow the humans to still use the machinery just as the, tractor, the just as they would before it was outfitted with autonomous capabilities, so that they can do certain un there's, certain unstructured things that would probably be a pain to get the robot to do.

And maybe it's just a one-off thing.

[00:56:10] Ben Alfi: lot of growers use it as meditation, by the way. They just say, hey, I need just one hour to drive with my tractor. Yes, don't take it away from me. I'm guessing it's like horses. Okay, you still like to just ride a horse. But at the time, but to do the job with the horse, no, I'm okay with that.

[00:56:30] Audrow Nash: Yeah, I like driving just a little bit,

[00:56:33] Ben Alfi: Yeah, exactly.

[00:56:34] Audrow Nash: wanna drive all the time.

[00:56:35] Ben Alfi: Exactly. Exactly.

[00:56:36] Audrow Nash: Eight hours a day is too much.

[00:56:38] Ben Alfi: Too much.

[00:56:39] Audrow Nash: That's so funny. They're like, ah, just go in there, see how it feel and see how it's running.

[00:56:43] Future of unmanned tractors

[00:56:43] Audrow Nash: I love that. But, so going just a little further, going further into the future, do you imagine that we're gonna start building agriculture?

Maybe, and I wonder the timeline for this, but, building tractors and things like this that are not manned, maybe we can optimize them so that they don't have, I'm sure that there's different considerations in making a tractor that would result in a slightly different form if we say, oh, it doesn't need a human on it.

do you think that we're gonna be moving towards there? And I think the approach of retrofitting existing tractors is much smarter than trying to initially. Create tractors that don't like a custom made, very expensive. They need to buy a whole new fleet again. but moving towards, in the long term, world where tractors may not have people on them ever.

tell me a bit about that,

[00:57:45] Ben Alfi: this is my personal view. Okay, and track it also from what we've seen in manned and unmanned aerial vehicles. In a way, what I think we'll see in the next, I'd say three to five years, we'll see more and more, tractors that are coming out of the line that are more digital, with, going from mechanical to, digital and from digital to drive by wire. Then the next generation would be with sensors already integrated inside.

And then we will see what is beautiful about tractor. That is a multi-mission. Okay, I can do it for so many things. And so the growers are more innovative than whoever invented that tractor and what to do with it. So you still want that flexibility, this multi-mission, vehicle. So I see the seat.

Staying there vacant for quite a while, we will see some classic robots where it's not a seat, like just for certain, just I don't know, landmowers or, that we see in our house, near our houses, or all kinds of, cleaning machines and things like that. So they're very simplified ideas. The thing will be the balance on the cost of material, and these machines are huge. And if now, because you don't have a chair over there, half of the season it just sits, and you're not using it, not the correct way to do it. and, so I think this, we will see until, I'm guessing 2030, 2035, at least, we will see still tractors going out with manual and autonomous.

2030, we'll start to see them, with the autonomous inside a capabilities, and still think about transition and how long does it take to adopt and things like that. So we're off to a hybrid after market area for the next, 15 years. 10 to 15 years. It's a, again, it's a huge investment to buy a tractor for the growers.

It's zero thought in a way. It's very easy once trust is made to take an aftermarket kit because you see ROI. Small example, just to have a cabin on a tractor with air condition and anti chemical capabilities because you are driving in chemical area environment. This alone costs more than the setup fee of an autonomous vehicle,

[01:00:50] Audrow Nash: Yeah.

[01:00:51] Ben Alfi: so Most of our tractors that you will see around are without a cabin at all.

[01:00:58] Audrow Nash: Oh yeah. 'cause so then the person's just sitting out there and for these really poisonous, tasks, you just

[01:01:05] Ben Alfi: you go autonomy. Yeah.

[01:01:07] Audrow Nash: Yeah. I like that. And it makes a lot of sense, what you were saying where it gradually bleeds into having more of these auto features that would enable autonomy.

I would think of it, it's almost vehicles like, cars would be similar where you add ABS brakes and then you add lane keeping and then you add, and it just gets more and more sophisticated. Now cars have ultrasonics on them and you're starting to see different sensors. Some have cameras, and then eventually it enables more and more autonomy

[01:01:37] Ben Alfi: you'll still need that Yeah, you'll still need that autonomous, that the operating system that you are used to so so you don't you still want the Bluewhite autonomous system that is able to operate a Operating system for any type of tractors whether it's Autonomy from the shop or aftermarket and you'll still want the connectivity on the implements to make sure that everything is running and going correctly.

You'll still want the ability to use the data to different various areas and so these are the things that, keeps it evergreen, the need for that and what we see also that we will see more and more Bluewhite. Algorithms and capabilities that have been matured with millions of hours implemented in the tractor companies inside.

[01:02:35] Audrow Nash: Yeah. That'll be so cool. 'cause I would imagine from your perspective as that shift happens in 15 years or. More to become like where you don't need the seat for the operator, I would imagine, Bluewhite will be incredibly well positioned to make the jump to autonomy without an operator seat. and that will be awesome.

Yep.

[01:02:59] Ben Alfi: want to be the leading company for data driven farms, off road, mining, and hopefully space. The first, the first farming in space will be, and also I'm guessing the last, will be by robots and by autonomous vehicles.

[01:03:16] Audrow Nash: Ah ha.

[01:03:17] Ben Alfi: we see, ourselves as farmers, working with the farmers, wherever and whenever humankind is, needs.

This is where we are.

[01:03:31] Audrow Nash: Yeah. And so the big advantage, I guess if you solve this problem for tractors. You can also solve this for big mining vehicles. It's a fairly similar problem. I would imagine. Maybe it's more waypoint routing and this kind of thing, rather than back and forth, but it's still following a

[01:03:49] Ben Alfi: Yeah,

[01:03:49] Audrow Nash: I

[01:03:49] Ben Alfi: it's just too costly to mature it over there. To do one million hours in mining and one million hours in agriculture, I can do it while people are earning money.

[01:04:02] Audrow Nash: That is such a clever thing. I did not understand that. Yeah, you're right. Because you're getting these, tractors are mowing. If you can automate it, you can have them mowing around the clock and then you're getting so many hours of operation. You're testing your systems

[01:04:18] Ben Alfi: Maturing those algorithms that are not really still on the decision making side, but they're on the maturity side. So you have a real life laboratory that is running. And not just a simulation that we're also doing.

[01:04:37] Audrow Nash: And also providing value too, while it's learning. I really like that.

[01:04:41] Ben Alfi: Yeah.

[01:04:43] Audrow Nash: What's the, space applications? That sounds so interesting. what do you imagine for that? I have no idea what farming in space or mobility in space. I guess if you're on Mars and we wanna drive things around from here to there, we probably want autonomous systems to do it.

Is that what you mean? Or

what kind of

[01:05:00] Ben Alfi: Yes And I think these are the application that will be sensor based before we are doing a global positioning system in each and every, planet. And the idea is, how do I, can, navigate and monitor, and there will be those implements that are space implements that are needed.

to do whatever is needed over there. So there will be always the relationship between different type of vehicles that are running, different type of implements that are running, and operating system that can suit them all and integrate everything. It can be either not just air vehicles, but also air, not just ground vehicles, but also air and ground altogether under one operating system.

And the idea that the core infrastructure, what we have built and invested a lot on the infrastructure is to have the ability to grow and to adapt to those areas and not to be just stuck and pinpointed on one certain vehicle, one certain implement, one certain environment.

[01:06:12] Audrow Nash: Yeah. That would be limiting for sure. And so having it be flexible for all these vehicles opens up a lot. And then, yeah, you have such diverse applications or domains, space to mining, to farming. That's very cool. Let's see. So with space, just one more space question, then we'll get back to more practical things or more, sooner things I suppose.

do you imagine it'll be on the moon or on Mars or where would you think would be like the first application

[01:06:44] Ben Alfi: I think it will be in a place with an atmosphere. Whether it's an atmosphere done by humans, a dome like, or a place that has some kind of an atmosphere.

[01:06:58] Audrow Nash: Gotcha. Okay. Very cool. I can't wait until that future is here. I hope it's not too long.

[01:07:04] Ben Alfi: I hope, I'm guessing it will happen in your lifetime.

[01:07:11] Audrow Nash: Let's see.

[01:07:14] Data as a Service + Fellowship

[01:07:14] Audrow Nash: So, I wanted to talk about the data 'cause we haven't really talked too much about the data that you're getting and then the uses for it. So we've gone up the whole. Application stack. You have these actuators and sensors on tractors, and then you have a cloud that connects it to farmers.

And then you've mentioned the ability to use that data in flexible ways to maybe assess, I don't know, help them with insurance or help 'em assess crop health over time or whatever it might be. tell me a bit about the data and how you expose it and just how, that whole process of adding value from the data that you're capturing works.

[01:07:55] Ben Alfi: first and foremost, we see ourselves as, we are in charge of whatever is moving in the farm, and while it's moving to collect data and to distribute it to whoever needs it. As long as the grower approves it, and without infringing privacy. the idea is that I don't see Bluewhite as the company that needs to do agronomic evaluation.

But, think of it, if I can send whatever an agronomist needs, in order to evaluate what should be the spring next, the next spray event, or is it the time to start harvesting, or any, or Can I predict yield? It is through the cloud, through the data lake that we are, we have we can process to wherever on demand, depending on, the demands too.

Whoever third party is there, so we can go to ROS, say, Hey, okay, who's doing your yield prediction? Okay, somebody in the farm or a company that you're working with, and we can fill them up with the information. For those data companies to gather the information, this is what, really kills those companies because they're spending so much effort, time, money on the operation cost to collect this data and we can help. And this, the same data can help to so many applications. The same data can help, carbon. People are talking about carbon savings and, weather saving and all that, all those things.

[01:09:35] Audrow Nash: That's really cool. I really like how you guys are drawing lines of what you'll do and what you'll let other people. This is an opportunity and you guys could, maybe it becomes part of your vertical in a sense where you start helping people with the analysis of the data. but that's a future decision.

To make, but for now, it's like there's a line drawn and you say, it's not our core competence. We're gonna collect all it. We know it's valuable. but these other companies, like other businesses, can be founded around analyzing the data that you guys are generating. And so it's another, it's a, it's clever to me where you are deciding you are gonna focus and what you're gonna allow other people to scoop up for value in what you're already doing.

[01:10:26] Ben Alfi: First of all, I think it's very important in general, not to bully and not to be so aggressive and to think that we can, be, ruling all the world. Second, we are, a value based company. Our values are fellowship, love of the land, and innovation. When we talk about fellowship, it's a, this is an exact example of what we're talking.

We are not coming to the agriculture business who's been working for more than a century in a certain way and say, Okay, we don't need dealership. We don't need growers. We don't need tractor companies. We don't need the data company. No, we're blending in. We want to be enablers to all. We want everybody to say, Hey, Bluewhite is helping me.

It was on the operation cost. It's not my cost. It's about the people and the people who work there, their health care, their safety, their ability to work without their having their back ache or any other hazardous event. So this is how I see it, and if you are focused and you are, in a way, you are transparent in where you are, this is why we have so many great connections with the data companies like TELUS and others that are working with us, and communication companies that we talked about, we have from Israel or from the US, we work with Intel and others, growers, big, and also OEMs and also dealerships.

Maybe it's too authentic, maybe I'm too optimistic, but this is how I want to do

[01:12:06] Audrow Nash: way to do it. It strikes me as, even, like even if it wasn't ethical, your position on this, I think it would be pragmatic. and what I mean is that you can't do everything. And if you try to do all of these other areas to monopolize it, you'll spread the company thin. You need to raise even more money to do even less good across a more areas, So I think focusing on a core competence and then getting everything running, and then you can say, okay, maybe you can really work with these data companies or like in the future, but, for now, just saying, Hey, you have access to this. It's all free. You, or maybe not free. They can pay for it even. but it still is cheaper than them flying drones over the farm to try to gather their data.

This kind of thing.

[01:12:57] Ben Alfi: Exactly, we need to be humble, yet assertive. On autonomous vehicles we understand, and how to go from 0 to 100, or from 100 to 10, 000, we understand in this area. We've done it before. To be in places with no GPS, we understand, this is our happy place. But, We need to know what we know and what we don't know.

[01:13:21] Blending into the agruculture culture

[01:13:21] Ben Alfi: I had no idea about agriculture and my co founder is an amazing, he grew in a farm and he was with tractor all his life. And do what you know how to do, learn from others, more things and decide your boundaries.

[01:13:39] Audrow Nash: Yeah, definitely. So tell me more about, so you said your co-founder has a lot of experience with tractors and with

[01:13:48] Ben Alfi: just grew on the farm, and there's nothing better than that.

[01:13:53] Audrow Nash: Yeah, for sure. Is that a big way of, because what strikes me is that you guys are really trying to, and you, said it earlier, you're trying to fit into the ecosystem.

You're not trying to just like bully everyone and push 'em over and say, we do it all differently. Now you're trying to say, Hey, I can make this one pain point easier, but everything else, and work with, we work with too Tell me a bit about that, making yourself fit well in the ecosystem and maybe the role of your co-founder in fitting well in the ecosystem.

[01:14:29] Ben Alfi: First of all, again, it comes with the values and attitude. This is one. Second is the ability to be as transparent as possible. When you are transparent, when you are saying, okay, this is what I know how to do, and this is what I don't know, it creates openness. And, the ecosystem, what I like about the agriculture ecosystem, that there is no logical reason to be in agriculture, unless you are passionate about the mission. Unless you are passionate and you understand that you are part of a greater cause. A greater cause of making food available to the world. And you also understand that it's about trust. And trust is critical. these values resonate together with the goals of myself and Yair, who founded the company with me, and together we understood that these values, these, this is what will enable us to work correctly.

[01:15:44] Audrow Nash: Yeah, it makes a lot of sense. And yeah, I think you could probably even say that about robotics. So doubly true what you're doing, where there's not a reason to be in it unless you're passionate about the mission.

[01:15:55] Ben Alfi: Oh,

[01:15:56] Audrow Nash: so Robotics and agriculture,

[01:15:58] Ben Alfi: yeah. I'm doing it for, I'm doing it, I'm guessing almost 20 years dealing with unmanned systems and robotics. I love it. I think it's, I think it's part of the future. I love being part of the disruption. I love seeing the, transformation, from, talking about will build, will be adoption of robots there to discussion of where the hell are the robots?

I need them now,

[01:16:29] Audrow Nash: Yeah,

[01:16:30] Ben Alfi: and, also the understanding also people, okay, robots cannot do everything. It's not going to fly and give you coffee on the porch, and while it does it, it will also answer your telephone and send the email and do what, the balance and understanding of what it can do, what it cannot do, and also on the development side to start with.

Very repetitive, no missions and tasks, and just go step by step, do it with safety, don't alienate it, make it very simple to use, make it in a way, not just bipartisan for everybody, but also multi type of people that can run it, no matter how sophisticated are You these are the things that really creates at the bottom line.

These are more important, the blending in the business model, the usage of existing assets. This is much more important than another algorithm that is over there.

[01:17:50] Audrow Nash: For sure. And it's also just the pragmatic way to approach it too, I think. so I really like that one.

[01:17:58] Making systems safe

[01:17:58] Audrow Nash: So you've mentioned safety a few times. Tell me about making your equipment safe for this. And I wonder do you do safety certification work or match some sort of standards or how, do you

[01:18:12] Ben Alfi: Oh, yeah. We, spend a lot of time on that one. from the system architecture, from the models, from the ability to, to, record and learn. And what we have done also, we've integrated, because it's a new area, we've integrated a few standards from urban mobility, from a military standard, from agriculture standards, from machinery standards, all mixed together.

And in a way, we have internal safety commissions, we have external safety analysts to look at it. This is critical, this is backbone, and we are doing debriefs, we are collecting data on mishaps, not just something that happened, something that might have happened, and even, for example, so these are critical things of how to, go.

It starts from system engineering, it goes along with, system architecture. redundancies, and, going the approach from inside the envelope, not breaking the envelope. You cannot throw a tractor to a ditch and say, oh, it was too much. There are so

[01:19:38] Audrow Nash: yeah, for sure. Okay. And I get, I bet that, safety. Is a big part of the trust getting, farmers' trust, getting people's trust in general. You need safe systems. They need to trust you with the tractor that they have, and they need to trust your system to treat it well and not drive it into a ditch, as you say.

[01:20:06] Ben Alfi: still talking about Tesla accidents or Uber or Cruze or any others. and still much safer, but it's very hard to explain that it's safer, because, it's a technology and you're afraid and, you're a bit skeptic and you don't have control, you feel that you don't have control when it will happen.

take it slowly. at the end of the day, it is statistics, okay? And accidents will happen for sure. You need to See, and to reduce as much as possible, did you expect it? Did you do the correct risk mitigation? how was the risk analysis for that? how was, how a bigger surprise was it for you?

This is how you look at it. How, when it happened, did you try to cover up or were you open about it? And they published a debrief about it so everybody can learn. It's about that, okay? Transparency, this is, mistakes happen. We're doing lots of mistakes every day. The question is how do we behave when they happen?

[01:21:19] Audrow Nash: Yeah, definitely if you take ownership of it and say, okay, we understand it, we're trying to fix what's

[01:21:24] Ben Alfi: Yeah. had an assumption, the assumption was wrong, anything like that, I think it goes in everything in life, but when we're talking about robotics and, integration, and, dealing with adoption and, maturity. Safety must be at the top level.

[01:21:45] Audrow Nash: Yes, for sure.

[01:21:48] Recent Series C investment round

[01:21:48] Audrow Nash: now I wanted to make sure we had a chance to talk about your C round of funding. So first congrats on the, it was, if I remember correctly, was it 39 million for C round

[01:22:00] Ben Alfi: Yes, it was a 39 million. We raised up to now around 85 million. We are humbled that it's with not only existing investors who believe in us, but also new investors coming in. And, joining the gang on this crazy journey. And we have amazing workers, and amazing, customers and, that we are working with all of them.

They've also, send and expressed their, love and their, appreciation of what we're doing, and it also resonated with the investors. I think everybody understands how big the task is of creating this capability to the world. We have investors from Israel, from the U. S., from Mexico, from, Canada and coming in and saying, Hey, we want to be part of that journey.

So this is what is going on right now. We see this investment that it will enable us to take it to the next step of maturing the product in a way that it can be distributed worldwide and to grow with more and more capabilities as we talked along this discussion. until the next, I think the next either IPO or round, this is where we're going to, and stay tuned, I'm guessing.

[01:23:46] Audrow Nash: Yeah, for

sure

[01:23:48] Working with the community

[01:23:48] Ben Alfi: we are available to anyone, to be an entrepreneur in robotics, you need to be passionate, and you need to be a dreamer, and would love to help to anyone that is at this state, status. And whether it's personally or reach out to Bluewhite people and LinkedIn or any other media that you think.

And, remember that values come first. Doesn't matter what you do, make sure that values come first always. That's what I have from my side.

[01:24:28] Audrow Nash: Hell yeah. Yeah. It seems like you guys are a very value driven company, which is really nice 'cause it's very cool to see. I imagine you've been this way since the early days and seeing you as a larger company now who's in an exciting position to start scaling their work and the values are still driving you, which is very nice.

And like doing a debriefing when something goes wrong. I really like that idea. I feel like a lot of companies may just try to shove it under the rug when something goes wrong, but being transparent and, I don't know, trying to lift up the whole agriculture industry

[01:25:08] Ben Alfi: we have responsibility. to take this around, if we fail, the next investor won't invest in the other robotic company, who might be the successful one, right? So we cannot fail. And we need to be very mindful of what we're doing. We have responsibility on the agriculture. We have responsibility. It's not just a Take the money and run.

This is not why at this age, I've decided to spend every time of my life on, this, on what we're doing right now.

responsibility and, that's about it,

[01:25:51] Audrow Nash: I like that a lot. Yeah. 'cause you see a need and you are trying to help it. And you're right that there are big implications for any startup that's following because if the big Robotics startup goes bust, it makes if an investment in Robotics goes bad. It makes all future investments a little more difficult. and it may discourage people and it, you may, burn some farmers and if you burn some farmers, maybe they're hesitant to adopt Robotics technology in the future. So yeah, it's interesting to think of it as an ecosystem.

[01:26:27] Ben Alfi: Robotics, if you're in robotics in the next decade, think of yourself as a crusader.

[01:26:37] Audrow Nash: Love it.

[01:26:38] Ben Alfi: this is how you, and think of it that you are responsible. To how fast will the world adopt those capabilities? And you, if, and if you have a hunch that it's a wrong path, what you're trying to do, you are correct, go to a better one and, make it easier on yourself, make it easier on the ecosystem, develop what is needed.

But, be, It's a warrior type environment where you need to do it and I believe I see amazing companies here And I see also amazing successful companies that I wish we can be just like them.

[01:27:24] Audrow Nash: Hell yeah. And, I think the next years for Robotics are gonna be exciting and difficult and I, see it as, very necessary. Like I, I've been focusing on the labor shortages, for a while, and I really feel like Robotics is a way that we live in the future while labor shortages are happening and populations are aging and this kind of thing, and we maintain our quality of life.

I feel like Robotics has an important. It will be important for us to figure out how robots can get in and help

[01:28:06] Ben Alfi: I think it's

you're correct. It's labor shortage and labor transformation

[01:28:11] Audrow Nash: Mm-Hmm

[01:28:12] Ben Alfi: need to do this job anymore. Okay? We don't need to. There's no reason. If you want to work out, just do work out. don't take stones on your back and climb a hill. Just for work. I just wanted to say really, first of all, thank you for having me.

And it was great that you had this show. I've been talking about robotics, talking about from the technology, but not just the technology, how to implement it. To the world, I listened to some of your podcasts before and I learned from this a lot. I hope more and more people will come to the show and share, not just the happiness, but also the struggles.

And there are a huge amount of struggles in order to make it. And really, just to say thank you for having me here.

[01:29:12] Audrow Nash: Hell yeah. it's been absolutely wonderful, speaking with you and learning more about Bluewhite.

[01:29:18] Ben Alfi: Cheers.

[01:29:20] Audrow Nash: thank you. Alright, bye everyone.

[01:29:22] Ben Alfi: See you everybody. Bye bye.

​

[01:29:26] Episode outro

[01:29:26] Audrow Nash: You made it! What'd you think? Blue White is doing a great job, aren't they? What other robotics companies have you seen that are doing a good job fitting into an existing ecosystem, like agriculture? I'm curious to know in the comments or on X. If you like this interview You'll probably like the weekly spaces I'm hosting on X.

I often have a guest, and we do a short interview where you get to ask your questions. They've been a lot of fun, and there's a great community on X. If you're interested, there are Thursdays at 9pm Eastern, 6pm Pacific. Just look for at Audrow on X. That's all for now. Happy building!