Sequoia's Shaun Maguire: hardware is coming back, AI video gets its ChatGPT moment within 12 months, robotics in 3 years

that with that. But I can tell you about my pillow.

It's just ads because

it just can't do anything.

We don't have any other

except uh it'll take a shot at it and serve you.

I think we have uh we have Shawn Magcguire in the reream waiting room. Uh let's bring him in. If he's ready to go, let's bring in Shawn Magcguire. Sean, how are you doing?

Team, how are we doing? What's going on?

Fantastic.

Great to see you.

You look great. I believe we the first time we met in person was 2 years ago at Hillen Valley, right?

I think that's correct. This is uh the only location where I'll look as dapper as you gentlemen.

Yes. I put on a tie just for you. Uh it's a good day.

Just for me.

How is Hillen Valley this year? What's changed? What are the focal topics? What are you focused on?

It's bigger.

Bigger and better.

Okay.

Uh Jacob Delian, Christian deserve a lot of credit. They built something that definitely exceeded my wildest expectations for them. I'm proud of them. They crushed it.

And I got to say, you just had Casey Handmer on. Yeah.

Casey is a Galaxy Brain.

Yeah.

We did our PhD.

You guys were at Caltech together, right?

Not not just at Caltech. I mean, his now wife was my housemate for a couple years. So, anyways, he's a smart dude. He taught me a lot.

Yeah.

Uh he's also a little eccentric. He loves doing night hikes like literally when full moon. So, I used to go hike up hike up mountains with him.

That's fun.

I buy that.

Makes sense. He's like, there's actually no rules that say you can't hike at night.

Yeah.

Yeah. That was a It was a good indoctrination for me and thinking out of the box.

Yeah. Wait, aren't you a surfer?

No, I surfer.

I grew up in a town where everyone surfs, but I was a computer nerd.

So, yes, I can surf.

Okay.

But like I can probably surf better than YOU GUYS, BUT I SHOT SEAN'S FIRE. But I'm But I'm horrible compared to now.

We have a surf off.

Now we have to surf.

We have to surf. Jordy takes a lot. I cannot surf. So, but I think Jord is very serious. But, uh,

also like 65 and the fake news is coming out today.

68.

68. Let's Let's keep counting.

Uh, okay. I want to get into I want to get into mass drivers and uh Elon, we we watched his presentation on Saturday.

I'm sure you did as well. Uh walk us through

how you processed it. You came out with an essay that you said you edited for the first time.

Um so yeah, walk us through your reaction and then how you've been processing uh the general response.

Yeah. Well, on the editing essay, which is obviously not the point, I I I started my UC uh college admissions essays that had an electronic deadline two hours before they were due and single shot shipped it. My mom was standing over my shoulder losing her minds. We talked about it yesterday. It was pretty funny. That's amazing. Um, look, Master Drive on the Moon, I know it sounds crazy, but I think it's kind of the inevitable evolution or the inevitable next step. It's not the final state,

but it's the next step. And if you have

like if you have Starship and if you have Optimus

with those two things, like if you have a lot of those things,

I think build that's basically having humans, you know, dig like synthetic humans you could put anywhere. The moon is a place. There's a lot of, you know, raw materials there. And so if you can build a fab on Earth and the supply chain, you can build it on the moon with Optimus. And so I mean, look, I I think we're about to have both Starship and Optimus. So I think the mass drive on the moon is kind of inevitable just due to economics.

Yeah.

And I I think it I saw your amazing question, John, around when it will happen. Yeah.

I can't tell you like I

would guess not within 15 years, but I I think it will happen within 25.

Yeah. Yeah. That's sort of where I got. And I think it's it's almost uncharacteristic in Silicon Valley for founders to actually think in decades and propose anything more than 10 years. Think in decades. Think in decades.

But no one actually says this is what I'm going to do in 20 years. Not like that. No, not like that. Yeah. No, no, no. It's true.

That's too ambitious.

Yeah. And and I think I I don't know. I I came away from it being like being like Yeah. It is crazy. Yeah. It is decades away. But what better north star to have than something that is actually life's work level mission that can go on that's not just the same thing that's not just the next iteration of what you've already built and you've done successfully. It's something completely net new that will take an incredible force of will. And so I don't know it is it is interesting to me like how much can that philosophy scale to other founders? I imagine that if a series B founder came into your office and said, "Okay, I'm laying out this like 40-year timeline," you might be like, "Okay, like let's focus on just keeping the company alive for the next two years." Um, but maybe we need more of that. I don't know how how do you think about it in terms of like just the founder psyche where we are in this like in this like weird moment of like AI doing everything. Who knows what the next few years will look like that like how do you how do you think about this scaling to just like the founders that you work with broadly? As you guys know, I jump around a lot. And so, just going back to the mass driver. Sure. I'll just say one thing that I think people don't understand about it that I was I I was trying to make this point. I didn't make it very clearly.

I think the single hardest thing we're about to have done, which is Starship and Optimus.

Yeah.

I think that what you actually build on the moon is not very hard. Like the mass driver itself is not very hard. building fabs and the entire supply chain to do that is a lot of work, but we've done it on Earth. And so it's like I think that we're about to having all the raw ingredients put together and then it's just going to take a lot of time and effort to do it. But I I think people are probably overestimating how much future difficulty is required and underestimating how much past difficulty we've already kind of surpassed.

Yeah.

So anyway, sorry, just had to make that point.

Yeah. Um, in terms of where we are, I wrote a hardware manifesto about three and a half years ago that I sent internally at Sequoa

and I sent it probably to like 100 people in Silicon Valley, you know, around then, but I never posted it publicly because felt like there was too much alpha in it and kind of the

It was kind of like a personal thing.

Yeah. You know, I was like trying to make money for my firm and my friends. Sure. Um any anyways the core argument I made was a few things but like I was trying to argue that in the case of Sequoia 50year-old firm at the time now 54 years old for the first 25 years of the firm we made almost all of our money in hardware

for the last 25 years we made almost all of our money in software

I was trying to argue that I think the next 25 years we're going to make our money in hardware again

like something like AI is unbelievable one of the biggest revolutions ever. But like literally in this essay I said that I think a lot of the money in AI will be made at the hardware layer. Um and as we as we enter kind of a world where digital intelligence as abundance before physical intelligence like my personal forecast is I think the abundance of digital intelligence is about 10 years ahead of physical intelligence. not in terms of like the state-of-the-art of what it can do, but in terms of the full roll out like the, you know, the total deployment or as Daria would say the diffusion rate.

Sure.

Um, and so I think there's like this 10-year moment where hardware is not commodified at all and software kind of will be. And a couple of the other points I made, one is that if you just think logically, every software revolution is preceded by a hardware revolution. like to have the iOS, you know, app store that enabled Uber and Door Dash and all these great companies, you needed to have the iPhone. The iPhone took 20 years of, you know, you needed Qualcomm, you needed Broadcom, you needed, you know, fiber layouts, you needed 4G, you needed, you know, CPUs getting smaller and, you know, battery touchscreen, you need battery density getting better, etc. This true for any software revolution. um literally by definition is preceded by a hardware revolution.

Mhm.

And this AI revolution like we're seeing what it can do from the software layer, but it's still limited by hardware. I think that's like at least 10 more years.

And then when you go beyond that like I I think we're kind of next point is I think we're entering like a phase transition where the hardware we were doing for a long time was kind of all following Moors law. it like it was all like the branching out of this decision in the mid1 1950s to go all in on the silicon supply chain and that has created magic. There's still a couple orders of magnitude of juice to squeeze but you know we hit fundamental physics law like limits dinard scaling things like that that I think this tech tree is kind of branching into humanoid robots into silicon photonics into orbital data centers um you know all these new hardware areas where there's going to be 20 plus years of progress we get to make and there's going to be you know just incredible businesses is built on the back of this

with and a lot of and a lot of dumpster fires as well.

Yeah. Are you seeing hardware companies pivot into AI? We've we we've talked to Blake Schaw at Boom about moving into uh I think natural gas turbine production. And I'm I'm wondering about some of your other space investments that might say, "Hey, maybe we want to get into the data center boom in space." like where else are you seeing energy reallocate towards other products into being more aligned with the AI boom? You know, I like pushing people's buttons. And, you know, I I remember the good old days when crypto and AI were friends.

And and I'll and I'll just remind you guys that a lot of these, you know, incredibly

groundbreaking AI data center companies like Crusoe. I know you had Chase on this on the show earlier today. Yeah.

You know, or we like

both of those were Bitcoin miners in the early days.

Yep.

Uh and pivoted to AI. But I I actually view this differently. I think that like AI and crypto have a shared history. And I'm not gonna I've articulated that online before. I'm not going to go into that whole thing. But I I think that yeah, we will see anyone that was an energy company is becoming an AI company. But I think that's naturally. I think that's natural. I think that this it's not like a hard pivot. I would frame it the other way that AI is becoming an energy

industry. I think that in Silicon Valley like it's it's almost this Capernac principle where people used to believe that the world revolved around the sun um or sorry revolved around the earth and then you know realize that it revolves around the sun at least our solar system. Um, I think that there's a Silicon Valley parallel where people think that all of technology, all of industry revolves around Silicon Valley, but it's actually kind of the opposite. Silicon Valley is earth and there's these much bigger forces you know and bodies which includes energy and you know chemicals industry these giant supply chains semiconductor industry which has been around for a very long time. It was pretty advanced and people in Silicon Valley forgot about it even though you know we played a big role in building it.

Like producing mirrors for the lithography machines that's something that is you're never going to see talked about in Silicon Valley until now. Yeah. Or or or making chemicals, you know, to make ultra pure, you know, wafers. Yeah.

And anyways, I just I don't like the language of other industries pivoting into AI.

I actually think it's the opposite. I think it's like Silicon Valley is waking up from its toddlerhood and realizing that there were these giant industries that we underestimated and they're actually really freaking good at what they do

and we can all benefit by working together. And so I just it's a different twist on how you even asked the question but yes there are many industries merging together and hardware is coming back just to the the root of every the way you think about every business in the valley.

Uh last question for me prediction around a chat GPT moment for robotics. Do you think we're on a 2year timeline one year fiveyear? Where do you sit? So, I actually think we're going to see this moment in video, like in AI video before we see it in robotics.

My forecast, like I may be wrong, it's just me forecasting. I think that within the next 12 months, we're going to see the chat GPT moment for AI video. I think that right now we're almost in like the GPT3 era where for the people that are really insiders, they know that the models have gotten really good, but they haven't quite been packaged in that form factor that just breaks out in the consumer psyche massively and just have this like extreme deployment. Um, so I I think I would say chatbt video moment in the next 12 months and then the robotics moment I think will be a little bit longer than that. is 100% coming, but my guess would be it's three years. Personally, this is my own personal forecast.

Yeah. Yeah. No, I like it. I think that's about right.

And and let me just say that like almost in the definition of the question is it's like a cons from a consumer perspective, I think that things like optimists will be doing very useful things inside factories, but that's not the chachi moment. That's not like when the public is like, "Oh, wow. this thing is changing my life. I want to use this thing. It's it's ubiquitous. I think that's probably a few years behind like useful things in private.

Yeah, that makes sense.