Building a Company in Stealth | Travis Kalanick with a16z

And it’s not just the return of Travis; it’s also the return of this partnership, or what could have been in this relationship between Travis and Ben. I want to read a quote from Travis’s launch post:

“I’ve known Mark and Ben for a long time, and we got so close to partnering up at Uber in 2011. I, Travis, blame Marc. Ben blames himself, but let’s just say it was all on us—it was on all of us. And in 2017, Uber suffered the consequences of not having Marc on the board. If you know, you know.”

Travis, tell the story.

Okay, so this will be interesting because I think Ben and I have maybe discussed this once or twice.

Yeah, it’s very painful.

I’m not sure if there’s a difference of opinion on what happened, but I’ll tell you what I thought happened.

Yeah.

I’m in fundraising mode. This is our Series B in 2011. Uber started in June 2010, and I’m raising our Series B. Earlier that year, I did Series A with a VC I’ll remain nameless.

I was hardcore. I had gone through such tough times as an entrepreneur before Uber that, once I was at Uber and things were working, I was still acting like I wasn’t going to eat tomorrow. I would do everything right up to the line—be perfect. I had this whole fundraising process that I would execute to perfection, and that meant I would run an auction.

I met with the a16z guys and lots of other people. It was what I would call a winner-takes-all auction, meaning there was 1 major lead. They were going to set the price, and everybody else would fall in.

I would go to the 1 guy and tell him what the story was. They’d say, “Oh, yeah, we’re interested.” He’s leaning forward, and then they’d ask, “How much are you raising? What’s the price?”

I would say, “We don’t know. This is going to be 1 of the hottest deals in Silicon Valley this year. We don’t know where the price is going to be or where it’s going to go, but it’s at least this much.”

I called this the uncapped anchor. I’m never in a place where I’m negotiating with somebody where I’m here and they’re here and we meet in the middle. I’m always saying, “It’s at least this, and it can go up.” But I start low.

So it only goes up.

So everybody’s pumped and leaning forward. You can see the body language. You can read it. It’s good.

You go to the next guy, who literally comes to your office an hour later. The same thing happens. He’s fired up. I’m showing all the analytics, and it’s great. Then he asks, “What are you thinking about for the price?”

If the last one was at least 2 times our previous round, it just became 3 times. I’d say, “We don’t know. It’s going to be 1 of the top deals in Silicon Valley this year, but it will be at least 3 times our last round.” And he’d say, “Okay.”

Then you call the guy before and say, “Hey, look, dude. It went from 2 times to 3 times. We’ll see where it goes. I don’t know.” They start getting nervous.

You get all the way to the top, and then you go, “Going once, going twice.” It’s kind of funny because, in this particular case, Yuri was almost coming over the top. We got to $375 million pre-money with a16z. Yuri almost came in at $400 million, but he was like, “No, I can’t do it.” So I said, “Okay, going once, going twice, sold.”

Yeah.

It was a high-momentum deal, and that’s how you do it. At the time, you almost should feel a little shy about doing something like that. It’s a little bit crass, let’s just say, but that’s how I was wired back in the day: go all the way. Run the touchdown. Score the touchdown, even when it’s a little much.

We started working, and we were going to get a term sheet going—the whole thing. Then I got an email from Mark. It was something like, “Hey, Travis, let’s go to dinner.” I think it was the next night or something like that.

I said, “Send me the term sheet, and then, yeah, we can go to dinner.” He responded, “Let’s go to dinner.” I thought, “Oh, I didn’t like that.”

He came down south to this neck of the woods because I was in San Francisco. We went to a little sushi place. I don’t remember the name of it. He said, “Look, I know we talked about $375 million, but I went to the partnership, and they don’t agree. They just didn’t think it was the right number. The best we’re going to be able to do is $210 million.”

That’s how it went down for me.

Yeah. Well, that’s definitely different from the other side.

It would be interesting to hear your side. I’ve got more than the hour we have scheduled if we want to take the time.

Anyway, I had a high-momentum deal. It went all the way to the top, I auctioned it, cleared the market, and then the bottom came out from under it. I had to go back to everybody else and say, “$210 million is the new number.”

Yeah.

I had to tell them, “The guy who won backed out. I’m now back at $210 million. We’re starting the auction over.” It was super awkward and weird because people lose credibility in that moment.

I built it back up, and basically it ended up being Menlo Park. That’s how Shervin got involved in Uber. At some point, I told Shervin to stop negotiating against himself because he just kept driving the price up without anybody else. I said, “Dude, you won. It’s fine.” I didn’t want the same thing to happen again.

I’m very clear about the $375 million, the $210 million, and the experience I had, but there may be another side of that story.

Yeah. Well, let me just tell you what I actually remember, and then what I kind of remember.

I remember you came in for the pitch, and I think it was just you.

Yeah, that’s right.

Which was, by the way, unusual. Normally, particularly in those days, it was very rare for an entrepreneur to come in with no team members on a deal that size.

That was a big deal, by the way. That was a high-price deal at the time. Now it’s like a $75 million—

$75 million.

Yeah, you’d be like, “Dude, pre-seed, pre-seed.”

Pre-pre.

Right?

Yeah.

The pitch was super impressive. The other thing I recall pretty closely was that Travis had definitely read some of my blogs. I felt like, “Oh, we can win this. This is a deal.”

We had the discussion, and it was clear that we wanted to do the deal. At the time, I thought, “I should really do this one,” but I wasn’t the consumer guy. I also had a lot of board seats—

It was like 16 board seats. It was something crazy.

It was just—it was a lot. So I gave it to Mark and Jono Ferro to do.

Then all I remember is—I remember something different at this point, but I got it all from them, so I wasn’t there. There was something about the employee option pool or some [expletive]. I couldn’t—

But it was 1 of these things where I just assumed we were doing the deal. Then the next thing I knew, Shervin had the deal, and I was like, “Jeez.”

I regret—I can’t tell you how long, because that was the beginning of my torture over this thing. I thought, “I know we should do that deal. We lost the deal.”

Meanwhile, Scott Weisenthal did a deal for the other company in the space, Lyft. By the way, that got off to a very, very difficult start, largely because of Uber. In order to Dave and Catch[?], I ended up having to go on the board. I took over that board seat to help them through that thing.

And just so you understand, my experience with the Lyft thing happened, I believe, a year later—almost to the day, something like that.

Yeah.

The deal on Lyft was 210.

Was it really? I don’t know.

I know. I’m just saying, this is what the entrepreneur’s going to remember, dude. That’s what I’m saying.

Holy cow.

But anyway, because if a company is in a lot of trouble, that’s generally when I get called, given my history with that.

Oh man, I was working so hard on that.

And so—

I was working hard on that trouble.

Oh my God, it was a lot of trouble. It was a lot of trouble.

But anyway, I’m living with that. Then, by the way, I would see Travis now and again. I had him over to my house for a barbecue—just all kinds of stuff—because I knew who he was.

Yeah, we were kind of friends.

Yeah, yeah, yeah. There was always a lot of respect there because we knew who each other were. But he was winning.

And it was just like he wasn’t going to let me not hear it. It was horrible. I had to listen. I was like, “God damn it.”

There was also time—like, for a decade. There was also this other thing that would happen. This really started around 2014 or 2015: Emil and I would meet up with Mark and Ben at this super-undercover restaurant.

Ah, yes. Yes.

Right next to our office, which was in the same building at the time as Twitter—or, sorry, the same building as Square, right?

By the way, Emil is also a remarkably amazing guy and currently the CTO of the Department of War.

Yeah, 100%.

Yeah.

The best.

Yeah.

We would do these sessions. We’d just have dinner, break some bread, and talk shop. Emil and I would always walk out going, “Damn. It would be so great to have these guys involved.” Of course, it didn’t happen because they were at Lyft.

Yeah, yeah, yeah. You know what I—

And sometimes we might even talk about, “Hey, is there a thing that comes together?” But I was just so— that’s the way I was. There was no way—

—to deal with him other than if we gave him the company.

Yeah. And he could just fire everybody instantly.

But it was like that: there was mutual respect, with this really cool tension at the same time. That was just the extra texture on the Uber experience, at least for me.

Yeah, it was like the old Larry Bird–Magic Johnson thing. You’re a basketball guy. It was like, “Okay, we’ll meet each other. We hate you, but we wish you were on our team.”

This sucks.

So it’s the one that got away, but fast-forward, and now there’s a new chance at a partnership.

What I would say is, there’s one middle piece to this. Those who are obviously in the tech industry know what Uber’s 2017 was like. That wouldn’t have gone that way if Ben or Marc had been on the board.

Nope.

So when Travis and I talk about this, we’re like, “It was our fault.”

Yeah.

We screwed it up.

Yeah, yeah.

I mean, Uber would be considerably larger, more important, and more central as a company today. There’s no question. First of all, it would have won food. No question. Uber was dominating food at the time.

Yeah.

Dominating food at the time. And it would also have been—

DoorDash was—

—for sure a leader in autonomy.

DoorDash had 5% market share when I left.

Yeah.

And—

And, by the way—

—a tenth of a percent.

And, by the way, yeah—

—we’re not going home this weekend.

And by the way, Tony, God bless him, had a hell of a time getting his round done at the time. Right at that period, he was doing a round at DoorDash, and he had to really struggle to get it done. Congratulations to him for building what he built.

He got it done.

Yeah.

You know what? It’s really easy to do the what-ifs and the this and that. The thing that matters most is the tale of the tape. He got some stuff done.

He did. He did. He did.

You know, respect.

No, amazing. He survived it, which is, by the way, in entrepreneurship, surviving is a big part of it.

Yeah.

Yeah. Good stuff. Not that easy.

Our autonomy program was second to Waymo but catching up at the time.

And you had the network.

We had the network, too. We had the ferocity—or fierceness—that it takes to catch up in a certain technology space while building all these other things that we were doing. It was a lot. It was good times. It was a lot of fun.

Yeah, yeah, yeah.

Even though 2017 was super tough for me, and maybe you could say even for the company at large—certainly for the company at large—I loved every minute. Honestly, even when you come out of it, you’re like, “Yeah, it got a little weird at the end. It was tough.” I still— it was still great.

You still love some of that. You love it in retrospect.

Yeah, no, I get you. Trust me, you don’t have to remind me of that. I’m just saying that feeling that’s so horrible at the time, you know—

Yeah, there was a love affair there. That’s almost the way I would put it. A lot of people talk about, “Well, are you mad about this or that? Are you upset that that happened?” And I’m like, when you fall in love again, you don’t think about the ex very much.

Yeah.

That’s it. Then you can just be like, “It was a good time. She got a little crazy at the end.”

Yeah, that’s true. The crazy ex.

Yeah. So there we go.

How did you fall in love again? Or did it take you—how did you handle this interim period between—

We had a really interesting situation, I think. There were the headlines and the negativity in the headlines from that period of time, and you could really say the extreme wokeness that was coming in on Uber and trying to constrain it, if that makes sense.

There was an interim period of 7 or 8 months between when I left and when I started what’s now called Atoms. I had to sort of fight for my life because there were a number of lawsuits, a number of investigations, and all these things. I continued to stand by every decision I made at Uber, but there were a lot of times when you could make the argument that I just got way too close to the line.

I could show that there was no chalk on the shoe, but you would need a electromagnetic scanning microscope to see that there was no chalk on the shoe. You reverse-angle the slow-mo, and it’s just like—that’s the problem. It’s the edge of being a hard-scrabble small-startup kid brought to something that’s working and going super big.

Yeah.

You don’t realize that when you get big, there actually are different rules. And they’re not just the rules; there are the vibes of the rules.

The pirate becomes the navy. We talked about this at the firm a lot. When you’re the upstart and you’re battling the man, you can go hard in ways that, if you do it when you’re the big dog—

Okay.

Yeah, you’re viewed completely differently.

Yeah, and I hadn’t consumed that information.

It actually helped me because, when we started the firm, we were the upstart. I talked some crazy shit about the other VCs and this and that, and called them names. But at some point, I could feel it turning, and I remembered what happened to him. I was like—and we actually had a whole thing: “We’re not the pirates anymore. Don’t say that type of stuff. Be chill.”

Yeah, we had this thing where, against Lyft, we would recruit the drivers on Lyft to bring them over to Uber so that it was hard for Lyft to build supply. We would do it aggressively, and we had a program for it internally that we called “shoplifting.”

Okay.

At some point—

Just that name, by the way. Forget about all that. I know some of the techniques they used—very aggressive—but that name? You can’t use that name if you’re—

Yeah, it’s just not a thing. I remember David Drummond from Google was on our board, and he was like, “You know, Travis, that’s not a thing, dude.”

You’ve got to change it. And then I had antitrust training, which, again, at the time, I’m like this small-startup kid. I’m like, “Antitrust? What the hell is that?” Legal people were telling me whatever. They were like, “Look.” I’m like, “I can’t wait to have that problem.” And they’re like, “Look, here’s the thing: the way you call projects, it has to be appropriate for a 10-year-old basketball team.”

So it went from Shoplift to the North American Championship Series. We called it the NACS. [Laughter] And we had a whole very heavy-duty initiative called the NACS to make sure that the North American Championship Series was won by us. [Laughter]

That is hilarious. I didn't know about the North American Championship Series then. [Laughter] Yeah. Okay, so you go through this 8-month period. You're in this hell. You're dealing with the lawsuits, you're reflecting back, you're thinking about what's next, and you say, “Hey, let's build again.”

So, what happened was, I had a buddy who started a thing right near the end of my Uber tenure, and it was called CloudKitchens. He's a real estate savant.

Yeah.

That's just the thing. He was looking at real estate, like, “There's got to be an angle in tech and real estate. How do we do this?” And he got a property up. Actually, let me start even before this. We got Uber Eats up in 2015. We saw the first dark kitchens—I don't know, it was late 2015, early 2016. We saw pictures of commercial kitchens in Melbourne that were not restaurants.

Hm.

They were on Uber Eats.

Oh, wow. Wow.

The first dark kitchens were emerging. We were like, “This is crazy. That's so interesting.” And a buddy of mine who is this real estate savant caught wind of what's going on, and he had this idea: Why don't we have a multi-tenant approach to this? Thirty kitchens on a single property, and basically lease these kitchens, which are 200 square feet, as delivery-only locations.

Anyway, I didn't really know what was going on. I sort of knew what was going on, but not really. Then, in the fall, 3 months after I had left Uber, we met up, and he's telling me about it. He's like, “Yeah, you know, we got a few tenants. It's kind of working. I'm going to do a few more of these. I think I'm going to try to figure out how to do a few more.” I'm like, “You mean a few thousand, right?” [Laughter]

And so then we partnered up. I acquired it. I think they'd just done their Series A. So I acquired it, and then we just went to town. It's really interesting because sometimes you start a thing, and sometimes you're involved in a thing and then you come in. I think this is an interesting story about how Elon was involved—obviously involved—in Tesla from the beginning. He was an investor at first but then went all the way in. It's interesting how these ideas come. I like to say that sometimes you have ideas, and sometimes ideas come to you.

Right.

But if it's an idea that's meant to be your soulmate, you know it when you see it. Sometimes you just go for it. It's like a love affair; you just go for it. Some people have these long lists.

Because there's a real difference between that and what's known as the professional CEO, right? Nobody smart, other than a few weird political people, would ever consider Elon a professional CEO. [Laughter] It's like, “Oh, that's an idea.”

“That's—I didn't start the company, but it's mine. That's mine. It is me. I know how to do it. These guys did a great job starting it, but there's no way they can do it.” That's a very different animal.

Yeah.

So, you fall in love with something, and it's meant to be for you at that time, and you just go for it. For me, complexity is interesting. Things that are naturally not sexy on the surface are also weirdly interesting. Especially if you can see that it's sexy but people don't understand it yet. That's where the sparkle in your eye goes: “You don't see what I see, and I'm pretty sure I'm right.” That's the fun part, yeah.

But this is a very complex thing because we're buying property, we're doing construction, and we then have a sales team that's selling restaurants a delivery-only location. So they're expanding the restaurant chain. This isn't like I'm just getting into a vendor flow. We're actually making a strategic sale, which is, “You should expand. You should expand your business, and you should expand it in a new way,” which is 2 strategic moves at the same time.

Yeah. By the way, you like complexity.

Exactly, yeah. So then you're like, “Okay, we've woven technology through this whole real estate thing.” We have a software stack that, of course, will work in this facility but will also sell it to all your brick-and-mortars everywhere. And let's just get some robotics going at the same time, because labor's your biggest problem if you're a restaurateur. If you solve that, it's really big. I can get into the strategy of why that is, but just the brass tacks of the complexity of the project were interesting to me.

So it was more emergent or iterative than, “Hey, I've got this master vision for what it's all going to turn into”?

No, I think at that time, if I don't see how it's big and meaningful, then complexity for complexity's sake isn't a thing. There has to be—maybe some would say—a pot of gold at the end of the rainbow. It has to be something meaningful.

The meaningful thing about Adam’s food, or what folks know as CloudKitchens, is: Can you make the preparation and delivery of a meal—a quality meal—so efficient that it approaches the cost of going to the grocery store? If it does, you do to the kitchen what Uber did to the car. You go, “Okay, that's kind of a big deal.”

But then you say, “What do you need to do to get there?” Okay, well, you need to do e-commerce for online food delivery, which means I need warehouses, Amazon-style.

Yeah.

Except they're not just for logistics. It's not just picking, packing, sorting, and then putting it in a car. You also have to have manufacturing there, too—manufacturing, also known as a restaurant. People don't think of that, but you go to the Bureau of Labor Statistics: restaurants are manufacturers, okay? So you have to have manufacturing and logistics infrastructure in the same place. That's real estate.

Then you actually have to have automated production. That's the manufacturing, so robotic food robots, essentially—food robotics. And then you need to have robotic couriers. You do those 3 things, and then all of a sudden, you are getting super-high-quality meals, everything you could ever imagine. We call it the internet food court. And it gets to you at the price of going to the grocery store.

So you go, “Okay, nice. That's a nice story.” Then you go, “Well, in 20 years, will people be cooking food regularly?” And there will be a suite of robotics that basically will make higher-quality meals. They'll just be doing a better job than any of us could do.

I think most people, especially those in the robotics space, would be like, “Of course. It's obvious that that's going to happen.” And you're like, “In 20 years, are we going to have—I call them autonomous burritos—boxes on wheels that hold food at temperature and come to your house?” Yes. Everybody will say yes, absolutely. We see Waymo's cruising around already. We just need a much cheaper version of that.

So 20 years is the easy one. Then you go, “Well, what about 10? What about 7? What about 5?” And so that's when you're bending reality. You're bending it toward now. Everybody would agree that this is going to happen, and now it's about when and who. Then you say, “LFG.”

Sometimes you're a little bit early, and I would say on CloudKitchens, we were a bit early. But you build those bricks—all those things I talked about. We have many hundreds of facilities in 30 countries. Restaurants are doing better and better because the online delivery market keeps growing. So success is happening.

Yeah.

Right? Well, the restaurant world has high churn. Your churn starts coming down because volume's going up. We have applications that guarantee volume into these restaurants so that we can have successful tenants, which means I'm a successful real estate guy from a business model perspective.

The robotics start to work. We're getting a manufacturing line for our robots up in Q4. The pieces start to come together. You're like, “Production is now 50% cheaper.” Once we get the robotic couriers, you go from $12 a drop for each meal to $0.50 to $1 in distribution cost per meal.

Now you're taking $6 out in labor. You're taking, let's call it, $10 or $11 out on the courier. On the real estate, because you're going to get higher-value volume because you're driving the prices down, you probably save a few bucks. Let's call it $2 or $3 on occupancy per meal. All of a sudden, you've got an $8 or $10 meal that's delivered to you all-in.

Plus, you don't have to recruit drivers, you know, and get Shoplifted. [Laughter]

That's true.

That's right. You can't Shoplift a robotic courier.

Yeah. Yes.

So we need to stop using this term.

It's so wrong. Okay.

You have 100 facilities in 30 countries and thousands of employees over an 8-year period, and you're in stealth.

Yeah, dude. Yeah, that's just super-hard mode.

Yeah.

And the thing is, look, coming out of 2017, it was 150 articles a day of negativity. I wanted to be able to build without bringing that negativity into the system. And I wanted the team to be able to build without worrying about what The New York Times was writing tomorrow.

Yeah.

Simple as that. And it was the right thing to do.

Yeah.

The hard mode was that we had to recruit talent, and it was always from a recruiter who had a stealth thing on LinkedIn and in their signature. You had to go get customers with, like, “Stealth” in LinkedIn. You're thousands of people big; this is hard mode.

Yeah.

And it was interesting. We didn't know how long it was going to go.

And you have to trust that no one's going to leak.

You know what happens? Here's the cool part about stealth if you go that long. I don't recommend this for anybody. It's not the thing.

Special case.

Yeah, you don't really—I don't know. But eventually, there's this break in your narrative. There was this story arc, and it was going, going, going, and then it's discontinuous. It stops. There's this long period of time where nothing is filling in.

What happens then is the media doesn't want to cover a story because it's so hard to discuss a thing that has a gap of a few years before it. It's too hard to do the story and it's too hard to cover it. So you get to this place where it becomes self-fulfilling.

Yeah.

It's easier to be stealth over time because even if somebody tries to leak, nobody even knows what to do with it.

Yeah.

They're just like, “I don't even understand it.” By the way, we did crazy stuff. People know it as CloudKitchens in the U.S., but it's Cocina Secreta in Latin America. And we have different names that mean “magical kitchen,” essentially. We have Kitchen Valley in Korea.

Yeah.

We have Flash Kitchen and 3 other names in China. We have Foodstars in London. It goes, you know.

Yeah.

I think it's KitchenPark in the Middle East. We were going to go with Yalla Kitchens. It's a bunch of things like this, so it was very hard to connect the dots.

Yeah.

You're extra stealth.

Yeah.

Anyways.

It worked.

And so, at some point, an even bigger vision for this company emerges to digitize the physical world, where you're category-creating and getting to coin the term “industrial AI.” Talk about when this bigger vision starts to culminate.

So, look, the vision for digitizing the physical world for me started at Uber. Treating atoms like bits started at Uber. It was sci-fi at one point, I promise. You would touch glass and a car would come to you.

Yeah.

And you had satellite view. You were watching this car come to you. That was crazy. Now it's just like, whatever, okay? But it was crazy at the time. Everybody's first Uber experience was a magical moment.

Digitizing the physical world was the thing. This idea of what I would call an atoms-based computer was something I was working on in terms of a framework toward the end of Uber. A CPU manipulates bits, storage stores bits, and a network moves bits from point A to point B.

Okay, well, a CPU manipulates bits—what manipulates atoms? That's manufacturing. Storage stores bits—what stores atoms? That's real estate. A network moves bits from point A to point B—what moves atoms? That's transport or logistics.

Those are now your 3 core computing resources in an atoms-based computer. And you know what? Everything you learn in your computer science curriculum, everything you learn in engineering, it plays in the atoms world. Same frameworks. All the data structures, all the algorithms—it works exactly the same.

Routing algorithms.

Oh, totally. Totally. In fact, we use TCP sawtooth to basically manage how much demand to send to certain kitchens based on capacity and reliability, as an example.

When I'm pitching this to a tech person who wants to understand that 10,000-square-foot facility, I call it a 10,000-square-foot semiconductor. It's a 30-core processor: 30 kitchens. They're computing atoms, not bits. The corridors where the food moves—that's a network bus. The processing center is like an L1 cache or something similar.

You're taking this analogy very far.

The cold storage where you have ice cream and juice? That's precomputed items that sit on the edge. Those are your Akamai servers. The courier who comes and picks up the food, goes on the road, and brings the meal to you—he's a TCP packet. This goes forever.

That's excellent. Okay, so the vision started at Uber.

The point is that atoms-based computation is a thing. Honestly, it's always been a thing. Uber was a network for the physical world. It digitized transportation. And, by the way, it's almost done.

Why is it almost done? Because it's almost fully software. Waymos are cruising. It's almost fully software. But what about a CPU for the physical world and storage for the physical world? Basically, digitize manufacturing and digitize real estate.

There's just so much to do there. It's crazy. There's a huge amount of innovation there, not just with what Atoms is doing, but there's going to be lots of stuff even beyond what we're doing. It's a portfolio strategy, really, if you're investing.

It's very interesting because you kind of got to the conclusion that we've now all learned from AI, which is you can model just about anything with computation.

Mhm.

And you basically were ahead of that in modeling this physical world, even metaphorically, with computation, and now kind of manifesting that—

It's—

—you know, probably entirely with AI on top of your computation model.

It's funny because the framework started near the end of my Uber time, really, when we were getting deep into deep learning and heavy-duty ML. We would go after some of the best AI talent. We had an AI lab that was probably one of the best in the world as well.

I was always pitching that the physical world is more interesting for this stuff than the digital one. And look, I'm selling my book. That's what we do, right? But you have way, way more variables—axes of stochastic distribution.

Yeah.

Which then means you're not going to solve it with an algorithm. You're going to solve it by empirically understanding how it works.

Yeah. Right. So let's fast-forward to how the bigger vision for Atoms comes together in terms of the combination of the companies.

I keep stopping you, you know.

And then I just want to get to the partnership because it also ties together.

It's all good. So, okay. You've got to have your manufacturing logistics hub. That's the real estate. You've got to have your automated production. And you've got to have your autonomous burritos.

About a year ago, I started looking into autonomous burritos, and I went to China and checked out all of the autonomy guys. I was here in the U.S. talking to a lot of the autonomy guys. Word got out. They're like, “Travis is looking at autonomy again.” People started going, “What's going on?”

Yeah. Yeah.

And, you know, folks like Uber and other partners across different categories were like, “We're interested.” For one reason or another, you could even be a supply-chain guy in food. You're also saying, “I need to have, for business continuity and just peace of mind, another alternative for how this freight—how these trucks, heavy transport—are going to move. And I need an alternative. I'd love to have a partner.” This kind of stuff.

So, I partnered with a couple of folks. I actually created a new company to do it because funding autonomy through a sort of core food company just wasn't going to work at first. The reason why is that these are just different profiles. It's a real-estate company; this is heavy-duty software, with big money going after it.

It's like a computer company trying to build a network company off to the side. You have to buy a networking company.

So we just got started, and these partners wanted a pure play. We created that sort of near the end of last year. But what happened is, you've got to get—okay, well, now I need to get my team together. I need to get my guys from way back when, 2017, when we were running hardcore at autonomy.

We need to get that going because if you're going to be in the specialized robotics game, you have to have autonomy for yourself. You have to. Your robots are moving and acting in the physical world, and you cannot depend on only 1 company that has an existence proof now. Of course, Tesla's going to get there, you know, sometime soon—or maybe some amount of time.

We don't know exactly. You can't be dependent in that way. So, okay, get your team together. Eric Meyhofer, who ran ATG, is already running my food robotics division. That's Uber's Advanced Technology Group, working for us.

Anthony Levandowski, who is one of my top leads in autonomy and, back at Uber, was running Pronto, which was an autonomous mining company—let's call it off-road autonomy. But it's a workflow or category where it's much more than just, "This thing needs to move." It's, "How does it move? What is the context in which it's moving?" It's a very specialized thing.

I was the biggest investor in Pronto. I'm like, "Let's go." So I acquired Pronto, and now I'm in the mining business. But it's not like I didn't know about it—I was the biggest investor in it to begin with, for a good reason. It's one of the most beautiful and interesting autonomy categories or use cases.

Our mission statement on the mining business is "More productive mines to power Earth's industries." We can go to a gold mine CEO right now and say, "Would you like to get 20% more gold per year?" We don't get a no. Yeah, right? Now he says, "Prove it." He says, "Prove it." And we're like, "Let's go."

What's happened is that Pronto has been working on mining for 7 or 8 years on its own. So, separately, Pronto has been working on mining for 7 or 8 years on its own.

Also a great target for autonomy because maybe the worst job for humans there is mining, right?

And we have mining customers because we're operating in a bunch of mines around the world. Our mining customers have existential safety problems—or let's call them safety issues, safety concerns—where bad things can happen on any given day. You've got real people putting their lives on the line when they're at these mines.

When you automate certain parts of mining, you dramatically improve safety, and you do it in a real way. When you look at the work that's being done and how it's being done, it's a little bit heart-wrenching when you're exposed to it and you see how mines work. It's not the mines' fault; it's just the nature of the work. There's a massive safety upside that goes along with the productivity upside, which is really special.

Yeah.

So, what's happened is that Pronto is just now coming past human productivity. It's like any enterprise software company: You do some cool pilots with big companies, you've got stuff, and you're like, "Dude, I've got 8 seats at this 10,000-person company." And you're like, "Okay, maybe." Then all of a sudden, it's better.

In the autonomy world, it's about being better than human productivity. That means your mines and quarries are exactly where you hit the gas. It's go time.

We're going to see super-exponential growth. We're seeing it—we're in the middle of it right now—on the mining side, which is super special. These guys ran a very lean startup for a long time, and now I'm going to their customers, who are begging us to move faster and get more out there.

We have to do supply chain and manufacture kits. We have to install these kits on machines that are in the middle of the Amazon or on the border of Saudi Arabia and Iraq, in these crazy, wild places. We have to do it fast, and then do the change management at every one of these mines.

They're like, "It's time to go. Let's go," and they're pushing us. So, how we talk to them is, "Look, we were a lean startup, and we're going to get muscular."

Yeah.

That's how I talk to the team, too, because you don't want it to get fat.

Yeah.

It's like lean muscle.

Yes.

But we need some protein shakes, creatine, and things like that. It's about process and professionalizing a lot of this stuff. In some ways, the growth there is almost deterministic, as long as you do the professional things to get all that stuff in line.

What it means is that existing mines are going to be producing more, and they're going to be able to go much further and get more. New mines are going to emerge that weren't viable before. What that means for AI, and what that means for progress here and everywhere, is a super-big deal.

Yeah, and so—

To continue, I'll get to the partnership conversation.

[Laughter]

At what point does this become a conglomerate? Are you thinking from the beginning that this is a conglomerate, or at what point does this become all wrapped up in Atoms?

Well, the fundraise was part of it. What happened is that we started the fundraise, and Ben remembers this because I gave him a preview at first, in Vegas, actually. I was like the guy with a trench coat that had 20 watches. I'm like, "Which watch do you want? I've got all kinds of watches."

He's like, "Dude, I just want the whole freaking trench coat. I don't want to invest in mining or transport or food. I want the things that you're doing, and let's be partners for all the things you do." So, that was the impetus for bringing the companies together.

The creation of Atoms—

Funding itself, yeah. The creation of Atoms and the funding itself being at what we would call topco level.

Yeah.

Right. And so now it's just one entity. I mean, there are subentities all over the place, but it's one equity structure, which is great for the employees and great for me because managing—I honestly don't know how Elon does it, or did it. It's pretty wild, and I'm really glad I don't have that headache, to be honest.

Yeah. Yeah. Yeah, it's very complex.

Yeah.

Complicated alignment. But, you know, it's an interesting part because it's very obvious, when we talk about it here, how it should work. And I think the narrative of how these things get built is constantly mistold by the industry and by the press.

If you hear about how Facebook came about, how Google came about, or how Tesla came about, it's, "Oh, they had this idea, and the idea was to do this." But that's not really what happened. A lot of people had that idea. There are a lot of those ideas out there.

If you look at Tesla, there was a very famous documentary called Who Killed the Electric Car? that came out before Tesla. If you're an entrepreneur watching that, what you would see is that big oil, big auto, and big government were never going to let that happen. They would kill you.

[Snorts]

But then the reality is, oh, Elon Musk—that motherfucker's hard to kill. So, he by himself kind of decarbonized the American auto industry, which is really unbelievable in retrospect.

But then it gets told—if you listen to the politicians now, they're like, "Oh, he didn't build Tesla." He just struck a gold mine and started pulling the gold out. "All those people built it." But why didn't all those people build another Tesla? There's only one because there's only one non-fungible, very rare person in that equation who could do that.

And the same is true of SpaceX. That's why there's only one SpaceX. That's why Meta's Meta and Friendster's Friendster, and all that kind of thing.

That's a good one. I haven't heard that one in a long time.

Yeah.

[Laughter]

It's good.

And so, to build a company—a great company—it's always a great entrepreneur. When we look at it, there are very few entrepreneurs in the history of the Valley who built something like Uber. Uber basically survived and is still a very valuable company without him, which almost never happens without the founder. It usually turns to crap very fast.

That was what we liked: Yes, the idea is great, the ambition is amazing, and it's going to be amazing for the world if he pulls it off. But we want to invest in the guy who could pull it off. That's kind of what drove the whole thing.

And there's a fun thing. I realized this at Uber, too: I think, in many ways, maybe Bezos was the guy to really go there. The only constraint to your imagination is management capacity, because look at all the things Amazon started doing. They were earlier than everybody else in starting to do lots of things, right?

Yeah.

And so I'm like—

Oh, by the way, another good one: How many people had the idea for an online bookstore, an online retailer, or online anything? Here we go.

And how many Amazons?

Yeah. Totally. You know, by the way, Jeff—everybody who knows Jeff, everybody who's heard him talk, you're like, "Oh, he's very special. That's a special CEO right there." That's our business: investing in that kind of very special individual. They're super rare.

And so I think, in many ways, Bezos may have been the OG of our time. I always want to say "the OG" of something, but then you find out 100 years earlier somebody else did it. How did AWS happen? It literally had nothing to do with a bookstore, really.

Nothing.

I mean, there are some fun stories they tell, but this is just a whole new thing.

Yeah, but by the way, it wasn't like the guy who was running Amazon.

Yeah, I know. Totally. Yeah, that's my point. Yeah, yeah.

So once you get the foothold, the beachhead, whatever, and it's working, you can start going: Uber rides, then Uber Eats, then Uber autonomy, then Uber Freight, then Uber AI Labs. You can start letting your imagination, with constraints, start to do really interesting things. And so where you start matters.

Yeah. You're right. It then has to be like—you need knowledge. Right? The thing that Amazon did have on AWS is they had knowledge of what it meant to run these things at scale. They were the first to get there, or one of the first 5 to get to that kind of scale, and so they had that knowledge. Totally different business, totally different, but they had the knowledge about what the product should be.

Mm-hmm.

Because they were the first to get there, or one of the first 5 to get to that kind of scale, they had that knowledge. And because he's a customer of these things, right? At Uber, he all of a sudden knows what the network is. He knows how to do that. And although it's a totally different business, that knowledge is transferable if you know how to scale, execute, and expand management.

Yeah.

And sometimes you go to—sometimes those adjacent categories are way easier than whole new—

Oh, yeah.

—off-the-rails categories.

Yes.

And there's also—I mean, the amazing thing about AWS is it's one of the few expansions where nobody would think, “What I want to buy from Amazon is a compute cloud.” Nobody was thinking that at the time.

No.

You know, there's always, “Oh, I should sell them more stuff they want to buy from me.” But that's a whole 'nother thing. But if you have that level of Jeff skill—which, in terms of organizational design and cultural continuity, that guy is—

Peak Bezos. Hard to beat.

Very good.

Yeah. Where you start matters.

Yeah, don't let him live his best life, like—

Yeah.

He gets a lot of bad PR for that, but—

Well, I was wondering: you could have had a version of living your best life where you just say, “Hey, this is what I do.” Or was it, “I'm going to keep doing it forever”? Or was it, “Hey, I chased the promised land, I got really far, but I could have gotten so much further. I'm going to get there next”? What was the motivation to just get right back in the—

It's just I love being in the arena, fighting it. And it's not—

Yeah. By the way, he's one of the very rare guys who made as much money as he did and wanted to keep playing. I mean—

And start over, basically. Not totally, but almost completely, right? So you've got to really want it, because also remember, I did it stealth.

Yeah.

So what it means—here's the thing about stealth: it's—

Stealth and starting at the bottom, right? Doing it by hand, right? Like, not—not—

Like, I walked in and acquired this company.

Chairman.

I acquired CloudKitchens or partnered with a buddy of mine, Diego, right? There were 6 people in the company that I acquired. Okay? So it's so funny, because you've got to think: the last all-hands I did had 20,000 people.

The next all-hands I did in this only company I was in had 6. And I stood in front just like the same and I'm like, “Let's fucking go, guys. Let's do this.”

Charts, you know, like, here's what's happening. You know, let's roll. Let's do it. And these guys, who I had just walked in with, were like, “What the fuck just happened?” They went to work that morning at a 6-person startup, and an hour later, the founder of Uber is now the CEO.

Yeah, that is wild. And then, by the way, we've seen it so many times where somebody gets to the penthouse and they're like, “Oh, I want to do a new thing.” But they don't want to go all the way down to the ground floor and start at the bottom, putting the building together. Okay, now you're in a fucking shotgun house. They want to jump—

Yeah.

—all the way to the top, which never works.

You can't start at the top.

And yeah, what the bottom thing is also about—there's a humility thing to it.

Yeah.

By the way, if you can pull it off, that's a very significant cultural advantage, because as soon as the company feels like it's doing something for outside status, it's a dangerous—

Yeah.

—volatile shit situation.

You start making decisions based on external validation versus internal correctness.

Yeah, you lose true north.

Yeah. And that's part of the upside of stealth and part of what I was doing, because remember, I wanted to build without worrying about what The New York Times was saying.

Yeah.

So it starts with internal correctness versus external validation.

Yeah.

Once a company starts doing that, it's hard to reverse. You already see this with the big labs. People keep going, “Why do the big labs smack themselves in the face?” Well, it's because they need that external validation for their employees, because their employees are addicted to it. “I want to be a good guy. I want to be a good AI researcher.” So go ahead and tell them that we're going to take all the jobs. It's like, really? You want to say that? You don't even know if it's true, but you want to say that. Well, why are you doing that? Well, that external validation is so important to them, and it's so important not just to the people running it, but to all the people in the company. And so they almost have to do it, even though if you wanted—if that's what you think, then work on those problems. Work on, like, “Oh, you don't want it to reward-hack? Well, why don't you try and solve that?” As opposed to go running your mouth, because you care about the external validation more than the actual thing. And that's the—

Yeah, it's the issue.

And it's very easy to slip into that. Many companies slip into that, and it's—

Once it gets out of the bag, it's very hard to reverse it.

There's a human nature aspect to being proud of what you do, wanting your mom to be proud of what you do, and your community at large.

Yeah.

So there is a human nature piece, which I was malnourishing by being stealth. But it forced an emotional intelligence that's very powerful now that we're moving out of stealth.

Yeah.

And it will be interesting to see how the coming years go.

Yeah, that's going to be a very interesting cultural case study. We'll see what the company is like over time.

Yeah. Yeah.

We've talked about what books were to Amazon. Food is to Atoms, in the sense of where you start matters. Why was it important that you started with food, in terms of what's generalizable?

So here's the thing, right? What CloudKitchens was all about, the way I saw it—and this is what I brought to the party—was that it was digitized manufacturing and digitized real estate. People know it as CloudKitchens because that's what we'd sell restaurant customers on, but the name of the company was City Storage Systems.

Storage for the physical world. This is digitized real estate.

Yeah.

This is a food computer with a basis in real estate storage. Okay?

That means a network.

Yeah. So this was the name of the company from the beginning—or once I came in, okay? It was purposely boring because I was going stealth. Can you imagine selling something called City Storage Systems? People are like, “What is that?” They'll forget literally 20 minutes after you talk to them. They're like, “City CSS?” They're thinking HTML. They're confused. They don't know what's going on.

For me, the framework was already there. I already thought about this as digitized manufacturing in real estate, about an atoms-based computer, about a food computer. And that's why I was excited. The idea captured me because of that.

So it wasn't like, “Why did you start with food?” It was like I saw something different than maybe even the original team that was working on it saw. But I saw where it was going, in the way that I just described. And that's what the romance was about. And yes, there are lots of digitized manufacturing problems out there, like mining and a whole bunch of other things. But this captured me at the right time, you know?

I was sitting at the bar and somebody came up, and we had a great conversation. We went to Cabo the next weekend and then just put a ring on it. Like, “Let's go.”

That's serendipity, man.

Yeah.

That's a real thing.

So that's okay, because the thing is, if you believe that you can find something, make it work, just grind it out, make it work, will it into working—your imagination can keep going. For me, it's not as much about where you start. It's about making sure you're passionate about where you start. And then where you go, you have no idea.

And this is very much—I go back to the Bezos style—which is like, your imagination will just keep coming up with new, cool ideas.

And so it's not about where you start; it's about why you start and how you build a culture that allows you to keep going into new places over time and lets your imagination, when it's right, flourish.

Yeah. That's why we say it's not an idea; it's an idea maze. The more you learn about the idea and all the things, the more you realize you don't know. The thing that is so hard to understand from the outside is that you know so little when you start. I mean, you didn't even tell me. When you started, you didn't know 1% of the problems you were going to run into with CloudKitchens.

Yeah, for sure.

Because you've got to really get into it.

Does the guy know anything about real estate?

Yeah.

0.0.

Yeah, so the whole thing is a learning process. The best entrepreneurs can take those learnings and multiply the applications. That's just a very, very rare thing. The fact that Elon has taken SpaceX to Starlink to data centers in space is just—nobody would start there.

Or think about this: the first Roadster—

Yeah.

—is now Optimus.

Yeah.

Okay? Talk about—

Yes. Yes.

—letting the imagination flourish. It's a beautiful thing.

Yeah, understanding what the core principles are.

Yeah.

And applying them.

So, with that said, paint us more of a picture about the future of Atoms, or maybe say some of the guiding principles that will dictate which industries you enter next or how you think about where you go from there.

Yeah, look. I tell people I'm very focused right now because it's only food, mining, and transport. There are so many things I haven't done yet. I'm so constrained right now.

But it goes back to that thing I was saying before: you've got to find that beachhead, make it flourish, build management capacity to solve 98% of the problems that are going on, and then you have the room to go do a new thing. I've been accused of using the marathon analogy too much.

Yeah.

And I'm going to live up to that. At some point, you might be doing a business and it just starts working. I started seeing that on rides at Uber, where I had my leads in different geographies, and I'd call them up because we'd jam on hard ideas and crank and all this. We didn't have much to talk about.

Yeah, because it was working.

It just worked, you know? And they're like, “Travis, we got this. I don't know why you're calling me right now.”

So, actually, this brings up one of the things that he did best at Uber. Having had to deal with these guys, if you look at most companies—most of the really big tech companies that emerged in the 2000s and 2010s—the management teams were kind of professional middle managers, et cetera.

He had so many guys who you would have said, “No, that's a founder. That's a founder over there. That's a founder over there. That's an actual entrepreneur.” It's amazing that they're not building their own companies because they're working for Travis. But that enabled them to get to a multitude of ideas. You need a team like that, and there just aren't that many people who can build that kind of team.

It's very hard to attract and maintain that level of capability, because those guys, unless you're building something absolutely spectacular, aren't going to be in. Even if you are, you've got to be able to manage that.

Well, what happens is the management of it becomes an empowerment puzzle.

Yeah.

So—

Yes.

Okay.

Not unlike here, by the way—

What I was going to say.

So you've got—I've got, you know, yeah, I'm running Uber, let's say. I have 5 continents going at any given time. You have an engineering team that's multiplexed across, a finance team that's multiplexed across, product—there's a lot going on.

The only way it works is: okay, you have that entrepreneur there, but empowerment basically starts with alignment. What's the strategy? Are we aligned on it?

That's the challenge, right? Empowerment plus alignment.

So alignment on the front end, accountability on the back end.

Yeah.

And we would call this: “Leaders lead, builders build.”

Yeah.

But there's also—so you've got alignment, but alignment is multidimensional. It's alignment on, okay, what are the goals? What are the objectives? Also, what is the culture? What is behavior? What you guys are doing over there is going to affect us over here and us over there, and so forth.

We can't dilute the culture because we gave you autonomy. We've got to have that, and that combination is a very high level of management skill.

And that leader has to be good enough to do it. That leader then has to work with the CFO and the CFO's people, the head of product and the product team's people, engineering and engineering's people, to make sure that it works. Because if I'm a single point of failure, it ain't scaling across 24 time zones. There's no freaking way.

Yeah. I think that's the difference between what we did at Uber versus what we're seeing with a lot of other companies that are expanding internationally: they're buying things. I was like, “Well, why buy? We could just freaking do this. A city is a city. Let's go.”

And then if you buy it, guess what? You get another culture.

Yeah.

Okay, and now you've got another reputation. That reputation is going to seep back, and now you don't have a culture.

And by the way—

Because how are those guys getting away with that in Korea when we can't get away with that in the US? All those things.

And by the way, your product and your technology are a reflection of your culture. So what happens is the culture is reflected in the technology and in how even that technology is built—the product decisions that are made. Now you've got to fuse these products and technologies together, which is very hard. 2 years later, maybe.

Yeah, exactly.

That's why I didn't buy Lyft.

That was a very different culture.

Yeah.

I could tell.

This was the reason I couldn't do it.

Yeah.

Yeah, super interesting. So you're constrained now by 3 different industries. What is the criteria that's going to determine if and when you go into another?

Take him off script.

No, no. What is the criteria that's going to determine if and when you go into another?

Yeah, it's the beachhead thing, and it's the—I'll go back to the marathon analogy—if it's getting easy, you must make it hard again.

But you've got to get it to easy.

You've got to start. It's a vibe. It's a feeling where you're basically constantly creating problems. I'm a problem creator. Creating a problem might be like, “Hey, let's do Uber in China.”

That's creating a real freaking problem. And by the way, you don't have a full understanding of that problem when you create it. Why is that hard? Because if you create a problem, you have to solve it.

I have this framework I call the meta-problem, which is: the derivative of your problem creation with respect to time must be less than or equal to the derivative of your problem solving with respect to time. If it's not, you have a real problem.

Yeah. Then you're underwater. You can drown, and if you have multiples of those, you're really underwater.

That's right. When you create a problem, you're guessing at what your capacity is to solve problems. You're guessing 6, 12, or 18 months ahead with only a certain understanding of what the nature of the problem is. When you get it really wrong, you're underwater.

Then you must stop all problem creation so that you can get back above water. When you go from your neck to your waist in terms of how underwater you are with your problems, you're like, “Oh, I could get to here.” You create more problems. You let the problem-creation spigot open.

And by the way, when you talk about capacity, a lot of it in this case is his personal capacity. If he's creating a new, large problem, he's got to be able to not pay nearly the same level of attention to the other parts of his business. Those have to be working, because if he's solving problems over there and creates a big new problem, and that one gets in trouble, then—

It's a little bit of both. It's a little bit of both. I never felt like rideshare was solved, because what happened? “Oh, it's time for autonomy.”

Yeah.

Okay. All right. Let's go.

By the way, you were right about that.

Yeah.

Despite—

Yeah.

I'm not going to even mention names.

And so, if you had still been running Uber in this other universe, some of the vision that you have here with Atoms, could you have gotten to mining with Uber?

It's a continuation of the same vision.

Right? So, digitizing the physical world—the Atoms-based computation—was already set.

Yeah.

And so, if the future of food is about infrastructure, automated production, and automated logistics, that's the future of Uber Eats, too. Now, when you're an OEM and you make a food computer, it doesn't mean you have to make all the parts.

Yeah.

But the future is the same.

Yeah.

And so that's my point: once you have that framework, why is the framework important? It's because it tells you what the future's going to be. Yeah, it's correct, and it's also compelling. It's interesting once you start seeing things. By the way, I built, with my team, 500 cloud kitchen facilities around the world. You could just look at it as, “Why are you building kitchens?” Of course, I would go around all the time and tell people, “I'm just your local humble kitchen builder,” just sort of messing around. But when you're hiring, it's like, “Why does this matter?”

Yeah.

Oh, that's right. It's a food computer.

Yeah.

So it's a mix of telling you where the future's going, pointing the way toward where the interesting things are, and helping you describe this in a way to top-quality talent who can now see what you see and get excited about it.

Yeah.

There is another interesting thing that we went through when we were understanding the business. Unlike in a computer, where you have storage and that storage kind of becomes obsolete over time, in the physical storage world—in real estate—if you have the computer optimizing the storage, the storage becomes more valuable. So you build the food computer, and the storage is central to delivering the end service, particularly before there are OEMs and everybody knows there's a food computer available. That food computer multiplies the value of the real estate. It's weird. Imagine a computer that multiplied the value of its storage. I guess it's a little like AI: if you've got data, all of a sudden data becomes valuable in a way that it wasn't really before.

Yeah.

Well, or you think of it this way: we could go to a data center. That's real estate, right? How many TOPS are happening at a data center today versus 20 years ago?

Yeah, exactly.

And how much value is coming out of those TOPS, right? The value of a data center, in terms of what it provides society, is going up dramatically over time.

Yeah.

And there's a lot of happy data center owners right now because of it.

No doubt.

Yeah.

I know my friend Rob, right?

Yeah.

Shout-out to Rob.

Yeah, we'll get him on soon. So, speaking of painting this vision to talent and talking about where the future is going, talk more about the emerging category that Atoms sits within, or that you're calling industrial AI.

Yeah. So, look, I did a vision letter when we launched Atoms that outlined a vision for physical AI, where we sit within it, and even went into what a physical AI stack is. In some ways, it was philosophy about where the technology world—in the tech industry and in the physical world—is going to go. But what I didn't do was get to the brass tacks.

When you look at it, there's software and there are bits that are affecting what happens to atoms: how they're assembled, where they go, and what function they perform in the physical world. We wanted to create an easy way to understand what that means. I said, “Atoms' mission: physical automation to transform industries.” You're like, “Okay.” It kind of gets me there, but what does that mean exactly? Well, it means we take entire industries that are ripe for automation and AI and, because these are physical industries, completely transform them.

In a sense, you've had one idea. Uber and Atoms—Atoms is an evolution of Uber, in a sense. You kind of said, “Okay, we're the Cisco of the physical world.” Then you're like, “Well, what about the other parts of the physical world?” You extended that vision into the computer, and now you're adding the parts.

But the computer's around an industry.

Yes.

So we make computers for industries.

Yeah.

Those computers are—

Right, that's an important distinction. They're industry-specific computers, but generalized networks in some sense.

Yes. So you go, “Okay, why are we doing transport?” Well, because you've got to have wheelbase for robots. Okay, cool. But what it actually means is every industry is moving things. There's some movement of things, eventually, and you must have wheelbase as part of what they do if you want to automate the full stack of that industry, as an example. But transport itself is an industry, too.

Yeah.

But maybe a horizontal one versus a vertical.

Right.

Yeah. Peter Thiel, a decade ago, famously talked about how we've had so much progress in bits but not in atoms, and how, because of regulation or whatever reason, it was just easier in bits. Here we are, and we're finally—

Yeah.

Getting to it in a real way.

Well, and I think the entrepreneurs who are coming up since the Uber time are not scared of regulation the way the bits—

Yeah, well, that was the breakthrough, in a large sense, that you made.

Yeah, and so what it means—

They've been trying to regulate bits lately.

Yeah, totally.

But these go together, because once the bits start controlling atoms in the physical world, then the regulators are in the game big time. That's generally how regulation works: it's about our physical world, and that's how it's been for thousands of years, certainly hundreds. Once you get into the physical world, you're like, “Okay, regulation's a thing.” You say, “Okay, when in the past have we seen something like this?” If you start studying the Second Industrial Revolution, there are so many parallels between what the Carnegies, the Rockefellers, Fricks, and Fords had to deal with in that massive, rapid change at that moment.

Fords.

The Fords—what those guys had to deal with in that massive, rapid change at that moment.

Same, very similar political climate.

Very similar.

The New Dealers were not that different from the current ones in terms of their complete hatred of the industrialists.

Okay, so there's a guy, Frick, who I think was Carnegie's partner. This is about steel, how to process steel, and doing it at scale. I believe this story is accurate—possibly mythology, but I believe it's true. An anarchist broke into his office while he was working. This was in either the early 1900s or late 1800s. The anarchist broke into his office and shot him in the neck.

Yeah.

He got it stitched up and went back to work that day.

Wow.

Somebody's going to look it up. I believe it's true, but somebody should fact-check it. You get the vibe of what's going on and how those guys rolled, and think about some of the entrepreneurs today who roll that way. Not all of them do. The Bits guys don't necessarily roll this way.

Elon rolls it.

But the Atoms guys roll this way.

Definitely.

Right? And so, when you look at the Second Industrial Revolution, the characters, and the adversity they had to go through, they're also none of them is Jesus. They had their flaws, too.

Yeah, yeah.

But there's a very interesting parallel to what's happening—or about to happen—with the coming industrial age. For selling our own book, we would call it the Atoms Age.

Yeah.

There's a book that's very interesting on the building of the war machine by the industrialists, which won us World War II, no question, and the resistance they got. It's called *Freedom's Forge*. There was this guy Bill Knudsen, who was a complete badass. He figured out mass production at Ford, then got in an argument with Henry Ford and went to GM and built Chevy. Then he went to work for the government for $1 to industrialize the nation. The way he was fought by the press and everybody was incredible. On the most basic thing, he's like, “Well, if we tell them to shut down making cars—and cars, by the way, employed 20% of America at the time—to make tanks and other things, then we're going to have to give them some money. So can we give them cost plus 7%?” That seemed very reasonable reading about it.

They're going to profit on the war?

Like Eleanor Roosevelt, everybody was just out to kill this guy. Somehow, that kind of resistance is very similar to what we're seeing in data centers today and that kind of thing. I think the governor of New York just banned data centers in the state, which—

There's a lot of parallels politically, which I would just call resistance to change.

Yeah. And nature resists very fast change.

Very fast change. Yeah.

That's correct. So, I think of it this way: the most important truth-seeking is the seeking of valuable unknown truths. But if you're good at finding valuable unknown truths—

You're creating change.

Then you know things other people don't know, which means you can do things other people can't do, and the better you are at it, the faster the change comes. And the faster the change comes, the more nature will do what it does. It's called resistance to change.

Yes.

And so you have to figure out how to bring big-time progress to overwhelm resistance to change. You also have to build trust as best as you can so that you have more advocates versus adversaries. And this is the dynamic.

Hard to do, by the way.

Very hard. Everybody loves progress; it's change they can't stand.

Yeah, those two happen to go together.

Duh.

Yeah. On the industrial AI thing, my whole vision note was physical AI, and this theoretical thing was very academic and theoretical. I wanted to get something practical—brass tacks. Industrial AI is basically this stack of software, sensors, robotics, and machinery that automates an industry, and we take it 1 industry at a time.

It just became much more real and clear what it is. You go, “Physical AI,” and you're like, “Oh, so you're doing a humanoid?” I'm like, “No, no, we're the non-humanoid guys.” I don't have a problem with them, but that's not what we do. You're like, “Oh, do you do some military thing?” You're like, [snorts] “No, that's not my thing either.” This felt very clean about what we're about and how we're going after it.

By the way, a really interesting point on this whole thing is that one of the things we completely forgot how to do is manufacture. We were the best in the world by far; we taught the world how to do it. Now we've basically forgotten how to do it as a country, other than Elon. If you need somebody who knows how to manufacture, you literally have to get them from Elon.

Yeah.

And we're going to have to relearn that. Actually, the beginnings of that are literally data centers. That is a manufacturing problem to a large extent, with the cooling and the power. You've got to build your own power and all that kind of thing. And there's so much resistance to that from the exact people who were harping and harping for years about how all the manufacturing jobs were going away.

Mhm.

Well, they're back. [laughter] But we don't want them, and so that's how tough this story and this kind of narrative is.

Yeah.

Thank God there are still enough places that are welcoming data centers.

Yeah.

It's not like—the federal system is a wonderful thing because you have this competition of some kind.

Yeah. It's really interesting how that works out.

We'd be Europe.

Yeah. Yeah. So we were talking offline. You were saying that when you started Uber, you worked with people who would choose to work with you. Now you're in a different position; you have a bit more choice. You've been building in stealth, and now this is your coming-out story. Talk to us about what kind of team you're building here and how folks can help.

Well, look, we just multiplied the number of industries we're going after by 3. That's kind of a thing, and each of them is a multitrillion-dollar opportunity. I had some awesome guys. This is kind of a funny thing to say, but when you make somebody hundreds of millions of dollars, it's hard to hire them again.

I've never had great success hiring rich people.

So there you go. There's the crew—Emil Michael is now undersecretary of defense, but he was my right hand on doing just epic dealmaking and strategic moves. I need the next Emil Michael, if that makes sense: a head of corp dev and biz dev who can build an inspiring team. As an example, we're in the market for a general counsel.

We're looking for somebody who can ride shotgun. You know, I like to go to the line; we should be a few inches off the line. That's a good thing to do. And then, of course, I need a CEO of my mining business. I need to build technical leadership across the board, from autonomy to heavy-duty robotics.

We have lots of folks coming in all the time, but it's those epic leaders who can inspire their team to go above and beyond and really make magic. No matter where you are as an entrepreneur, the talent game is the game. And again, when you multiply the industries by 3, we're back in the flow.

Because we were in stealth for so long, we weren't in the flow; we were cold-calling everybody. I think one of the beautiful things about us being back out there is that we can now have a place for people to understand who we are and what we're about. I can start putting teams together that can be out there in a very public way and let people know about the great work we're doing.

And if you're joining now, you're joining with both the benefit of having all the lessons you've learned and things you've proven out with CloudKitchens and, of course, Uber for almost 2 decades, but early enough that there's the upside of enormous opportunity going forward—

Yeah.

—and the ability to do something super meaningful. I mean, save kids from working in the mines, basically making it so that everybody can have great food instead of eating super-unhealthy, poisonous stuff that gives people their time back—

Yeah. Yeah.

—while also saving money.

And saving money.

Yeah. It's a hell of a mission.

Yeah, so we're having a lot of fun. It's good times.