Jacob Rintamaki on how robotics and AI are transforming data center maintenance and operations

about Console. Console builds AI agents that automate 70% of IT, HR, and finance support, giving employees instant resolution, access request, and password resets. Uh, let's bring in Jacob to the TBP Ultra Drum. Jacob, how are you doing?

Hello. How are you guys doing?

We're doing fantastically.

Great to have you here.

Is your 2026 off to a good start? I think it is. Yeah. Since to be honest, like just some backstory for why I'm here. Uh I about a month ago, a friend of mine was like, "Oh, you have some interesting thoughts about robotics, about kind of the data center market. You should like write something up." But it's kind of like a small piece.

And then it it took a month instead of like, oh, a couple days it spiraled. We had a whole section uh Dwar released a piece, so we have a section on that. We have a section about uh kind of the future. We have short stories. We have like a whole like financial model we've built out. Um it really spiraled and no I'm having a great January cuz it's now it's no longer raining in San Francisco. It's actually sunny today. I think this is a good omen believer in omens in 2026. Omens are in maxing.

That's a great omen maxing. Yeah. I think Tyler he speaks a lot about this.

Yes. Yes. I need to be on the You need to be open to omens to find them when they're sitting through.

I think so. I think that you do have to be open to

strange things happening. Uh since I think that's kind of like, you know, the talk my book for a little bit, uh like that's kind of the point of this essay is that not only is this going to happen a lot faster than I think people are realize, it's going to be a lot stranger. Like one thing I've heard people bring up is this idea that as you could call it like the bifurcation theory. This idea that as we get whether it's on AI or whether on robotics, we're going to get this like robot produced quantities of goods, maybe it's images, it's media, maybe it's clothing, maybe it's food, maybe it's these other sorts of opportunities. You have people like Ari Manuel kind of betting on the idea that there's going to be this split. there's going to be the live world and that you're going to have like the human rare quality stuff and then like the infinite robot slot.

But I think that's you can add a little more nuance to it. Like for example, I think that the idea that human only fashion is probably it's going to be very niche and the reason why is simple. If we look at something like shine and we say, "Okay, we can cut the costs another 10 times and the quality is something you would wear at the Met Gala." Most people, frankly, aren't going to be able to tell the difference in their tailor. Maybe we could get like the men's wear guy on, maybe he could tell. But I feel like most people, they're not going to be able to tell the difference. They're just going to like click the buy button and go.

But there is something where I think that like in healthcare within an industry, you can split it up into subportions. like the logistics of actually bringing medicine to people in like an urban hospital or someone like in NYU uh that's probably going to be done by robots but the actual care for people like people want if your child is sick and about to die unless if you know this doctor is about to do an operation you don't want to see a robot looking over your child you want to have somebody to be able to kind of absorb that fear to absorb these emotions and if you take that and you say okay wow that was like maybe slightly different than what I would have expected and apply that to literally every part of our society in the next decade. Yeah, it's going to be a lot. It's going to take you a month. It's going to be hard. And I'm glad that people starting to see this. Um that's kind of why I wrote it.

Where are we in terms of humanoid robots or or would you would you even use that term? Is it more just useful to think about physical instantiation of artificial intelligence just robotics broadly? because maybe they won't look like humanoids, but it's still going to be transformational. Are we like uh you know, you you see these stats where I think Amazon employs a million robots already. Do you see it more as like a continuous process? Will there be an iPhone moment, a chat GBT moment? Like what are you waiting for?

Okay, so I think there's a couple of ways we can look at this. Um one of them is a really interesting split. It's between manipulation and navigation. Mhm.

So navigation has been a much easier problem. And the reason it's an easier problem is that we've been able to use classical methods. We don't need to collect a lot of data for you know a warehouse robot to kind of like move around and move the boxes around. Even self-driving like it's not self-driving wasn't the hard problem. Getting it reliable enough so it doesn't kill people was the hard problem. Uh but manipulation is difficult. How do you encode this? Is like an idea that Entre Karpathy kind of spoke about in software 2.0. You can't really you can write a program to say, "Oh, I want a robot to go forward and it's going to go to left. It's going to go to right." How do you encode this idea of I want it to pick up this goldfish? You know, I'm not sponsored by them. Maybe someday.

And like I think that uh you know, that's kind of a weird thing about the robot future is that um I've written about robot holidays. I think there's going to be probably some weird religions. There's going to be all sorts of cults. It's going to be, you know, there's uh um I also wrote I

what is what is getting oneshotted or or having l psychosis like like humanoid robot psychosis that feels

well I mean I think that like Isaac Azimoff he wrote about this back in the 60s. You have robo psychologist like they literally wrote about like this robot going insane cuz the posatronic brain it got modified the three laws. you kind of, you know, I think if you kind of remove the rule of like, wow, maybe we shouldn't kill people, that would be bad.

Yeah.

Uh what about uh there we we we were talking uh reading a post earlier about uh Tesla with Optimus like doing a bunch of supplier selection seeming seems like they're gearing up to actually make a lot of these things. And I think some people are somewhat surprised by that because everything we've seen so far from humanoids has been like kind of like cool cool demo but not like oh I need this in my home today or I need this at my warehouse today.

Yeah. I mean I think there's a couple of things. One is that on the home versus enterprise enterprise is going to be a lot more valuable. Um not only because home is distributed you have safety stuff the costs. There's also this weird tax thing in the one big beautiful bill. It's called the 100% bonus depreciation clause. CFOs, you should listen to this. Uh you can depreciate whatever asset you want. It's uncapped uh in a year instead of 5 to 7 under this as long as it's kind of in the US. So you can do it for data centers, robotics. Like I think there's just a lot of things that are going to push for enterprise adoption first. But then within enterprise, I think the supply chain, why people are going to get side, you're just going to get blindsided by this is that a couple of things. One, um, if you're focusing on how bad America is at manufacturing, you're focusing, it's important, but you're focusing on the wrong thing. You need to focus on how good Asia is at making stuff. in particular, uh, Pyong in, uh, Malaysia, um, Bakni in Vietnam, um, Taiwan, um, underrated, uh, South Korea, Japan, obviously, China, Shenzhen, um, Wong, you lots of places. So, Asia is going to be really important. You can also repurpose other industries. I talked about this with rare rare earths processing since rare earths rare earths aren't rare you know as Trump said uh but like the actual uh processing of rare earths in the same way like standard oil became a monopoly off of processing China has like a processing monopoly there you can just repurpose the oil and gas industry for uh processing rare earths which I haven't spoken about people have spoken about it you know for finding the rare earths but not reprocessing and in the same way you can repurpose the automotive industry uh for like humanoid robots like you already have Honda, you have Toyota, these are already people that are integrating it. You have the same actuators, you have the same battery management systems, you have everything else. It's something where the auto OEMs were in a secular slump right now. So if you're not if you believe that's kind of it's like the subrini thesis in some sense of that, you know, auto is in this sort of like secular slump and then as we kind of like in order to break this slump, we're also going to have robots. So, it's going to be this double whammy of people are expecting, oh, you know, this sort of dying industry, it's stagnating. You know, profits are being competed away. Tesla is kind of it has to sell the energy story, the AI story, the robot story. The cars aren't, you know, doing it as well. BYD, like I think this is a Dan Wang point about uh China. Um, it's weirdly more capitalist than us. Like the Tian idea of capitalism is bad in some sense because you're competing. You're constantly you're trying to compete everything away. Um, so in some sense, China is more capitalist than us. Everyone's competing. That means that the auto industry is in a slump and it's looking for something to save it. I think robots are probably going to be it. And even China right now, like Unitry, there's kind of it's whispers of that, oh, are they going to delay the IPO this year or stuff because the hardware is getting ahead of the software. So, it's something where I'm like, I don't see the hardware and you can just rely on if you have like, you know, uh, expensive precision reducers or some of these other things like you don't need rare earths to make a motor. You don't need inspect expensive precision reducers to make an actuator. You can rely on cheap stuff. The thing that's going to like do my like, you know, surgery or stuff is not the thing that's going to stock shelves. And we just have like ooi or like how people are collecting data now with the gloves. you can 3D print stuff that's like a pretty good embodiment match. Like I'm just incredibly bullish. Like a few years ago, you could have asked me, I would said no. But now it's like no, it's a sign that this is going to be big. It's going to happen fast. You had mentioned before, John, about you know, humanoid versus other form factors. The key word to think about is payback times to borrow something from finance and that

the faster your payback time like the more that creditors are going to want to lend to your business. Since it's not just going to be venture capital, it's going to be the entire private credit industry. It's going to be private equity, it's going to be RAS, it's going to be all of these other things. Um, so you got to think, okay, how do I drive my payback time down? Yeah,

capex is one component. You get rid of the legs, you have a wheelbase or you just even get rid of the mobile stuff. You just have arms on an assembly line assembling, you know, AI servers or GPUs and you just crank the infinite money printer. But even more so, it's not even the capex. I feel like a lot of investors and a lot of people, they focus on the bomb, they focus on the capex. They should be focusing kind of on the opex on OEE on all of these metrics of saying, "Wow, if this thing is going to run like industrially rated robot arms already exist, it's fine. This thing is going to run for 30,000 hours and I I built a calculator on the website so you can play around with it yourself. This thing is running 30,000 hours and every day it's like 22 hours with some maintenance and it's like pretty close to a human or stuff. That is going to be the dominant term for like whether I'm going to get a good return on investment. what my I like my blended levered IR is going to be all of these other sorts of things like there's just a lot that goes into it and um I that's why like I want people to read this like I don't want this to kind of be like oh I like this and then we had like a discussion once it's like no I really tried to talk to as many people as I

put your butt down and study

yeah basically it's like it's going to be big it's going to be a big deal and I I try it to be fun I don't want it to be homework like I feel like the idea of like I tried to write it in a way that's fun and not, you know, cuz you could just go down and and read some textbook or some like schizophrenic supply chain reading or whatever. It's like that's just not fun. Um but I tried to make this fun. Um

what's your what's your thesis on land? I've had this idea that

Oh, it's in there. So yeah, I I I uh I was trying to find it, but but basically I've had this idea that like

if uh if you can build things, you know, 10 100 times faster, let's say you could build a a factory or multif family apartment building, um you can potentially start getting a return much faster and it's cheaper to, you know, make infrastructure, things like that that but yet physics, the laws of physics will presumably still apply, which means that it might be more convenient to have a factory like next to a city versus some somewhere random. So, I've had this idea that like it's possible there's like a situational in uh awareness style like real estate fund that could start thinking about how the value of land will change as you have advanced robotics that can uh kind of accelerate progress and and just build things uh quite a bit faster. But how are you thinking about land in the context of u advanced robotics?

Yeah. So, I'm going to back up a second. I think that the four factors of production classically are land, labor, capital, entrepreneurship, and all of those are covered in this, but we're going to focus on the land for a second. Um, you can split it up into what is it going to look like from a consumer and then an enterprise perspective because consumers and enterprises, they're going to be looking at this differently. From a consumer side, I think that like the originals will appreciate in the sense that like long Europe, long like historical sites, long existing cities of that it's going to be very hard. Like the idea of, oh, we're going to get cheap housing because construction is now done by robots. I think for Austin, for the South, we're going to have a lot of that, but I don't think that's going to solve San Francisco or New York's political wos. Um, and then from the enterprise side, I think you're correct. I think that in some sense the final offshoring may be the greatest reshoring we'll ever have in America of that you go from being constrained by not having this like tacet knowledge of manufacturing or enough people or it's all expensive to now you have the energy resources, you have natural gas, you have IP. Um, that's going to be very important. And yeah, maybe we don't have to put it right next to San Francisco, but you know, you go out west, you go down south, you could just kind of let things start to rip. I know that taxes in particular seems very primed from a regulatory perspective. I would also say I think that like for the idea of land taxes or land appreciation or stuff that feels somewhat deep into the curve like I think this is why like George's um has always been like this great idea in theory but not in practice. And this is kind of what Doresh and Philip talked about in their essay is that there's so much of the gains initially are just going to come from capital. Who owns the factories? Who owns the energy? The data centers that by the time you get to the point where oh I am just going to be Leopold. I'm going to raise a trillion dollars. I'm going to buy Iowa. Um like at that point like you should have just been buying like Stargate 2. You should have been going to New Taipei. You should have been with Jensen. You should have been in all these other things. An interesting thing of land though is that uh I'm totally a big believer in the whole orbital data center stuff. Um not just because of you know being friends with a lot of teams kind of at the the SpaceX XI some of these other places but also like I think that the thing about orbital data centers I want to keep this short since you could go on a whole rant is that it trades off one problem which is cooling. Um, it makes cooling a lot harder in space because you can only do radiative for all of these other problems. You don't have permitting, you don't have land, you don't have a lot of this other stuff. Maybe you'll need to do some radiation hard and things. But I do think that the idea of space data centers like is not going to be crazy in some sense. Like I also think because I spent a lot of time saying if you're a robotics company really be focused on the AI like infrastructure buildout as a market. It's really good. Like every robotics company that wants to be big, you know, is secretly a Dyson sphere company. They just don't know it yet. Like if you think about in the end of what can we do with robots, like everyone's thinking so short term of like if I had 100 million laborers, a billion, 100 billion, a trillion, like what would they do at a certain point? You can only make so many like mocha lattes or stuff. It's like you got to start, you know, building the Dyson sphere. We got to start cranking

Dyson sphere before 2100 or after 2100. Yeah. And if you enjoyed this conversation so far, this is like basically 80 pages of this.

Awesome.

Um, except worse.

What's your what's your prediction for a Dyson sphere timeline? 2100 before or after?

What do you count as a Dyson sphere? Cuz I feel like if I'm sorry to be like a pedantic like [ __ ] about this, but like what what do you count? What do you count as? Is it like Oh,

it doesn't need to it doesn't need to capture 100%. Let's call it

like a meaningful percent. a meaningful let's call it over 10% and definitely like a structure that is uh like like gravitationally locked to the sun not the earth so it's not just like a bunch of random starink sc like like scattered about in the in the stars it's it's truly like around the sun

before 2100 um but then I think that it but the problem you know we can hit the gong on that one uh

pages

we're going to pull this when it happens. We're going to pull this clip up.

I don't know. But I think though for like the whole Dyson sphere stuff, um it feels very much I don't know if you've ever heard of like the parable of the back half of a chess board that you know this king when they ask you a kind of a checker board, one grain of rice on the first one, two on the second, four on the third. By the time you get the 32, by the time you get the 64, like it just, you know, swallows the earth. I think it's something very similar here. like this idea of robots building robots which is again in the essay people should read it like that's the point of this like read the essay but um Fanic has been doing they're like this robot arm company uh they've been having robots building robots for 25 years now like that's older than me which is astonishing um but it's something where granted it's in a very limited narrow capacity of that you you have to scope everything for safety and whatnot but it's the idea of like oh robots building robots that's some like sci-fi [ __ ] like some of us don't have to live in Berkeley or whatever but you know I do think that it is going to be more plausible than not particularly because you know talking about the AI data center builder stuff the ODMs that are making all the AI servers all the GPUs everything all the chillers everything that's going into the data centers or even from it as a prefab perspective are also the people that could do consumer electronics they're the people that could actually build the robots so if you're servicing these customers or stuff you're servicing these these industries, it's like, wow, you know, they're going to want to start building it themselves because it lowers their cost. It increases their margins. It's going to be something where the market forces are going to be very powerful here. And that's also why like we spend a lot of time on the social sections of this of thinking about taxes, about I think sovereign wealth fund initially like between Bessant and Lutnik um there's a lot of really interesting ideas that you can start off with so you don't drive away the business community. But I do think that there's a lot of factors pointing to like it's going to go from 2025 people maybe they saw a physical intelligence video or something to 2027 it's like holy [ __ ] this is actually happening like guys I know that you wanted a uh a robot horse John or something

yeah we have the dog I need the horse I don't know I feel like the robot dog's a little played out the robot horse that's that's it for 2026 nobody has that A ste that's a very good name for

you would think they could just scale it up just scale the dog up

make a bigger dog they kind of have I mean the MIT cheetah stuff they they kind of did scale it up but it was like uh but it's not the same it looks lame it has to look hores

but also have to look cool I feel like that's also part of this essay was like um like between clog code and other stuff like I got a lot of compliments on what it looked like or so on I think that the marginal cost of making a good website's going to zero. Same like physical labor. Marshall cost of physical labor is going to zero. And my website as well. Like, might as well. Like, you got to make a good website now.

Yeah.

I'm totally being clapped out. So, I'm going to leave.

No, no, no. You did. I was clapping for your website. It looks fantastic.

We making a beautiful uh making a beautiful website for an essay is the new meta. You just you just maybe the AI 2027 people maybe they get some new PDF. Is this available as a PDF?

Yes, it is. It's available as a P. And also, if you want the secret, here's how you make a good

here's how you make a good uh

you want to make a good PF. Okay.

So, if we start with the thesis that like Claude Code like is God and does everything. Um then

of course kind of well obviously I mean have you seen the new anthropic price 350? I mean uh

price the the market capitalization not the not the share price. Yeah, I mean I think it's still grossly undervalued. Like add another zero or something at this point. Like um sorry Morgan Stanley, sorry Michael. We're gonna need to add that one. But uh I think that like uh he

Yeah. Well, I mean I I think he's busy with XI and all of that stuff. But uh they also had their recent round. Um but I I do think that if you treat like what in my life can is secretly a code generation problem. I think that making really beautiful PDFs is that because latte, which is mostly used by like math nerds um to render um math research papers is a programmable way of creating documents and PDFs. Like if you wonder how Leopold, I never I didn't do the footnotes here cuz my side notes and footnotes were completely crazy and it just like cluttered up the side and looked on aesthetic. It looked awful. Um and the website was the better medium for this anyways. But I do think that the idea of saying, "Oh, I have this really great piece of writing. I'm going to put in the cloud code. It's going to render it in latte. You can get an Overleaf account. It's free. Overleaf.com. Um, and then sign up. You can just like make these documents. You can have it be very expressive for someone. And then you can make like a really nice website to go along with it. Since it's not just about the raw content, it's also how you say it and then how it's presented. In the same way of that, oh, you know, I see John and Jordy. Um, they're going down the tartine, which I thought was an SF thing. I was told by Michael it's now like it would start in LA, which is a bunch of [ __ ] Like it really like that's some I

open people go to. This is not like you guys ordering the um

we invented Open AI near me.

Yeah, we exported it to you.

We exported it to San Francisco.

First time that's happened. Um, I think that, uh, but the thing though is that, you know, if you had like John and Troy, if you both had like, oh, we're going to, you know, write about technology and business, which I mean, I I'm a daily active reader of John's newsletter, so I think that, you know, we can pump that. Um,

let's go.

TBR.com.

Maybe I should turn it into a 90page PDF every day. Just call it.

I think you could render it as a website or you could you could spice it up. I feel like it's 2026. You got to not same old. You gota you got I wasn't at the whole Lulu stuff is like you got to be real. You got to be You said you can't be like all these other people. I mean,

yeah, there's opportunity.

Well, we've talked about doing a daily print newsletter somewhat like a newspaper.

Yeah.

Oh, I think that like um the only problem though is that it's kind of like Well, I mean like Colossus does it better. So, it's like

Colossus is evergreen. That's Evergreen. This is more just you you literally throw it out at the end. You're not collecting. Yeah. not like you're not keeping most people are not keeping

and the key would be it would be a beneficiary of robotics because it needs to be it needs to be printed and delivered within like 2 hours of the news breaking so it's like uncannily fast and that's a unique product but who knows maybe it's too sci-fi

maybe I I feel like that's a little too

make it a newspaper they write this thing it gets delivered every day make it make sense

yeah it doesn't work out can we interest you in like some New York Times like games inside. Can we actually go an app? We can

Yeah, you should add some games to this.

What's your game strategy vibe code or something? Sh's building RTS. What are you doing for whole thing? Like he wanted some games or others have fun. I was like, "Oh, you I'm just keep economics and this." I mean, it's a slider, but that's not fun.

No, we we'll add games in the second version of this.

Build an entire Stellar Stellaris or Hearts of Iron 4 spin-off game. Dwarf Fortressto. This was like Schulta was doing over Breaker. He was honestly also I have to say like shout out to him. He was a very I already knew he was going to be a very thoughtful editor. He was like even more thoughtful and I think that his quote tweet in addition to others are starting to bring attention from people that

you know I'm not going to be like the class or say who's reading or who's like following now or stuff. It's like oh wow a lot of people look up to him like I'm for him. Like he's he's the goat.

He is the goat.

He is the goat. Let's give it up for Shalto. One of the greatest ever doing.

Yeah.

Well, thank you so much for coming on the show. Always good to hang out anytime you want to help. Great to see you. Great to see you. Oh, thank you so much.

And uh I'm glad your 2026 is going well. Dropping start

and just try to, you know, use the tool. Try to get one of these out every day. This one took you whatever a couple months. Try to get one.

It took me a month, but I feel like Yeah, we got to you got to get down an hour. Every minute you got a new essay. Yes, faster than I can read.

A thoughtful essay every minute forever.

Thank you so much for coming on the show. Hope you have a great rest of your day and we will talk soon.

Goodbye.

Well, uh, is there any breaking news that we have yet to cover? Or is that where we should conclude today's story, today's show? Um,

uh, Reggie James built the situation monitor.

Amazing. Thank you.

Global activity monitor. has a TVPN integration, an Intel feed, tech, finance, politics, news feed, stocks, crypto.

So funny,

prediction, markets, tech, layoffs, tracker, AI race news. Is the Fed

tracker? Oh, wait. It keeps going. Wow.

It's just a bunch of it.

Is the Fed printer on? This is amazing.

This is great. Turn this into a like a Chrome.

Wait, it has today's episode live streaming on it. This is amazing. You can refresh everything. This is very cool. I like uh I like a

Yeah, we need this on the projector all day all day long.

This is this is great. Yeah. Uh

this is cool that it has a map. You can click around and see new stories.

This is really cool. I I I love a project like this. This is awesome. Uh and and in other news, the team got to the bottom of the large 3D holographic volutric display. It apparently exists. It's a 3D display with no glasses uh required. Uh much bigger, almost the size of the table. Not quite. Looks a little bit bigger, but uh pretty cool. We'll have to get a demo of that soon, but that looks fun. Uh anyway, thank you so much for tuning in today. We will see you tomorrow at 11.

We'll be back tomorrow. Big surprise, I know, but we'll be back tomorrow.

We're coming back. Uh we got a bunch of great guests for you folks and we will see you tomorrow. Leave us five stars on Apple Podcast, Spotify, subscribe to the newsletter.com, and we will see you tomorrow. Goodbye.

Cheers, folks. We love you.