# Jensen Huang Just Hinted At AI’s Next Trillion-Dollar Market

https://www.youtube.com/watch?v=EzJqiHrpYDM

[00:00] Let's run with the scenario, Vincent,
[00:01] that if you know, these IPOs launch,
[00:03] the early investors get out, get liquid with their money,
[00:05] you see a large pullback in the market, like the S&P pulls back to what I don't know, 5,500 or something like that.
[00:13] Like you see a major like 20, 30% pullback as a result, despite these new companies coming in.
[00:18] Well, I'm just okay, well, you guys can pick whatever that pullback scenario is for me, Vincent.
[00:22] Vincent, what would you then be looking at?
[00:24] Let's say a lot of your picks and even the stuff that you own, Milk Road Pro, let's say those those pulled back as well.
[00:28] E- equal amounts, which would you What would you What sector would you be looking at to continue this trade a- in what you're describing as a very nice buying opportunity?
[00:41] Yeah, obviously still AI.
[00:43] No, that's what I mean, but which parts of AI?
[00:45] There's a which what what would you be looking at?
[00:48] [laughter]
[00:50] Um I really like the trade outside of the rack, right?
[00:55] So, at the moment we saw everything inside the rack.
[01:00] CPUs, packaging, memory, etc.
[01:05] Outside the rack.
[01:05] Inside the rack, is that what you're saying?
[01:06] What does that mean?
[01:09] Can you explain to the audience what that means?
[01:11] That basically means everything you need to build and produce the chips and deploy them within data centers.
[01:20] Okay.
[01:22] Got you.
[01:22] Outside the rack means everything you need to deploy the chips and the racks within data centers.
[01:32] And that's power to me, and power has two sides to it.
[01:34] One is time to power.
[01:37] So, basically, how fast can you bring energy to the data centers?
[01:43] And the market kind of realized that trade, right?
[01:45] That's That's Bloom Energy.
[01:47] Um that's battery uh so, best battery energy storage.
[01:52] Um that's that's grid hardware, all those kind of things.
[01:55] The second side is power density.
[01:57] And that's coming with the fact that
[02:01] those regs, the chips, get more and more power density.
[02:05] So, we're moving toward towards that 800-V DC architecture.
[02:08] And that basically means you need stuff like power semiconductors to to convert the energy you get from the grid into the the type of energy those regs need.
[02:21] Those higher power dense regs also means we're creating a lot more heat within the regs.
[02:25] So, we need materials and chemicals to cool them down to ensure the integrity of those of those regs and and and and and and chips is is is kept alive, basically, right?
[02:37] And those are the traits I'm I'm looking at now.
[02:41] When you say chemicals what are you looking at companies or are you looking at buying like some sort of commodity or what do you mean by that?
[02:49] No, I'm I'm I'm I'm talking about buying companies here.
[02:52] Companies, yeah.
[02:53] Two that are really interesting are Entigrity and Kineti, I think.
[02:56] That's a spin-off out of DuPont.
[02:58] That's a massive chemical company.
[03:00] And what they're
[03:02] Providing is chemicals and materials.
[03:06] That in really simple terms make sure the reg does not overheat, does not get destroyed by the by the by the by the extensive heat it creates when when running, right?
[03:18] That's basically the trait here.
[03:20] And why is that important?
[03:22] Because I tried to explain this just now.
[03:25] When the future architecture of Nvidia chips are are increasing in power density, right?
[03:33] Up to 800-V by next year.
[03:35] That's completely different versus what we're having in the chips today, Blackwell from Nvidia, right?
[03:44] The the power density is so much higher.
[03:46] So, the heat we're creating within those chips is so much bigger.
[03:48] And that's why we need way more materials to cool them down to make the material stronger, more robust to actually survive that.
[03:58] And those are trades that the market has not realized yet.
[04:01] We are at a very crucial point in time.
[04:03] Right now.
[04:03] AI stocks have ripped.
[04:05] They're going to keep ripping.
[04:06] Our analysts on the AI side are up like literally 100% or more on quite a few calls.
[04:11] And crypto is about to boom again, or at least it feels that way.
[04:16] If you want on the inside on what we're buying, what our analysts are doing, what's on their watch list, all of that is in Milk Road Pro.
[04:21] So join at the link below.
[04:23] You're listening to the Milk Road Show, which means you've got takes.
[04:25] Strong ones, I bet.
[04:27] But where do those takes actually go?
[04:28] Do you tweet them into the void?
[04:31] Argue them in the group chat, or do you try to express them by buying a stock?
[04:34] The thing is, stocks move on like 50 other things at once.
[04:36] And that's where Kalshi comes in.
[04:38] It's a CFTC regulated prediction market where you bet directly on outcomes like Bitcoin hitting 100K, Fed rate cuts, GameStop buying eBay with a clean yes or no.
[04:49] We did a full deep dive on why prediction markets might be crypto's third product market fit moment.
[04:52] Check the show notes for the full report and claim $10 free when you trade $10 on Kalshi.
[04:57] All right.
[04:58] By the way, I just want to go back to what LG said.
[04:59] I don't see a world where S&P is going to come down 25, 30% as a
[05:05] result of a pullback in open AI.
[05:07] Open AI and Anthropic could go back 30% or 25% or whatever.
[05:12] Uh but I think the S&P would go down like maybe 10% max.
[05:15] That really triggered you, man.
[05:15] I I saw you physically react when I said that.
[05:20] I saw your body like kind of like contort.
[05:23] So upsetting for me to say it.
[05:25] Hey Vincent, I So I was I wrote a post this morning and um it was something from Jensen Huang had said, which was like, okay, so we started using AI with kind of chat first, right?
[05:35] Then and we So we started to price that in with Nvidia and all the companies that that had to do with that.
[05:41] It was like GPUs.
[05:42] Then it went to like the agentic economy.
[05:44] We're like, okay, we need more inference.
[05:45] We you know, we need CPUs.
[05:48] So like different companies got repriced.
[05:50] Then he said the next thing is physical AI.
[05:52] Right, we've talked about this a bunch.
[05:53] This is like this is your humanoid robots.
[05:55] This is your cars that are all using AI.
[06:00] How is this different?
[06:00] Is it still the exact same data centers that they need to to to compute and do what they need
[06:06] to do or is this a different bottleneck that this creates?
[06:11] Like how do you think through what this means and and how things are going to change?
[06:14] Cuz like if you were ahead of the Agentech thing and you're like, "Oh I knew that they're going to need CPUs."
[06:18] And you figured that out last year.
[06:20] Like you had a hell of a trade, right?
[06:22] So when you're thinking physical AI and I know we're still a while from getting humanoid robots and probably even, you know, robo taxis from really scaling and and using a lot of of of output here,
[06:32] where do you see that changing the bottleneck?
[06:36] Yeah.
[06:36] So first in terms of timeline, I think those trades are a year from now earliest.
[06:47] Especially when it comes to humanoid robots, right?
[06:48] Because I wrote a report about this humanoid robots at scale.
[06:55] In my view, 2029, 2030 the earliest.
[06:59] Not really because we're seeing issues on the on the supply side of humanoids, actually more on the demand side.
[07:06] We're not uh ready as a society to
[07:09] Deploy humanoid robots.
[07:10] It's a massive issue.
[07:12] Actually where the company I'm working at, we're we're seeing societal pushback from our employees against AI and humanoid robots.
[07:20] And I think that's something we're completely misunderstanding.
[07:23] But back to your question and specifically on humanoids, I think there are massive bottlenecks coming.
[07:29] Actuators, sensors.
[07:32] And uh Infineon I just bought is not only about semiconductors, but actually about sensors and and and power semis within humanoid as well.
[07:44] That's kind of the longer term tail of of of of of that company as well.
[07:49] So, I think there are a lot of bottlenecks that you can think of also when it comes to to robotaxis.
[07:54] Um because if you look out into a future with robotaxis, yes, we maybe have less cars, but we need to redo the entire fleet that is out there, right?
[08:08] And then you can think about everything you need.
[08:10] in a robotaxi versus a normal car.
[08:15] And those are all bottlenecks that that will emerge once once kind of humanoid robotaxis are produced at scale.
[08:22] And yeah, maybe it's a it's a good time to think about them now.
[08:24] Me personally, I'm more focused still on kind of the the data center bottlenecks because kind of the whole retail investor market is more focused on that, but maybe 6 months, 12 months from now it's time.
[08:37] But don't you think that the data center bottlenecks are already priced pretty insanely?
[08:43] Inside direct, yes. That's why I'm not buying Micron or the CPUs or anything.
[08:51] On the power side, chemical side, no, because look at an Infineon chart, look at a Kinetix chart on like longer term, two to three, four, five years, they're up maybe 50, 60% over the last four, five months.
[09:06] Yes, that's a lot, but over the last five years, they're also up 60,
[09:11] 70% because they did not perform, right?
[09:14] So, to me there's much more money to be made within those trades.
[09:20] Um and it's also come back to the come comes back to the chart you were showing earlier.
[09:24] The the capex for compute, for data center, for power will not decline over the next couple of years.
[09:30] And also importantly, the rate of change is not slowing that much down, which keeps that trade alive alive in my opinion.
[09:41] Yeah, makes sense.
[09:43] How are you thinking about physical embodied AI bottlenecks?
[09:48] Yeah, I don't have a strong thesis on it yet.
[09:51] I I think that to me that is the most obvious um next like big use case that comes from AI.
[10:01] Um and I think we've talked about that a bunch in the show, so that's nothing new.
[10:04] But I do think um and I talked about this a while ago, I think we're going to see an upgrade cycle.
[10:09] Like you said, there's there might not be more
[10:12] cars in the world as a result of robot taxis, but I think we're going to change everyone's going to change their cars.
[10:17] Um so meaning if you own one or two cars today, you're going to trade those in and you're going to get a brand new car that drives itself.
[10:24] Um and or you're going to buy cars that can go in within some sort of fleet, whether it's Waymo or Uber or or it's it's Tesla or whatever.
[10:30] And so I think there's going to be a huge upgrade cycle.
[10:35] Like when we all had Blackberries and then everyone switched to iPhones.
[10:39] Exactly.
[10:39] a huge moment and so much opportunity in markets, not just for Apple, but for the things that were put inside of Apple phones, like inside of iPhones, right?
[10:48] Like massive change the world, like the glass, right?
[10:49] Like if you invested in the company that used the glass for these iPhones, like holy that was crazy, right?
[10:55] Opportunity.
[10:58] And that's what I think is going to happen here with both and you kind of touched on it, humanoid robots and robot taxis.
[11:01] Robot taxis probably first.
[11:03] Um because I think like the world's already starting to wrap their heads around it and I think it's already close to being there.
[11:09] And then humanoid robots second, but the thing is is like uh while I do agree with you it's going
[11:13] to take till 2028, 2029 for humanoid robots to start to scale.
[11:18] I think there is one area where it will scale.
[11:21] I don't know if it's by then, but first, which is just in factories.
[11:25] Like Amazon is going to scale the hell out of this.
[11:26] The moment Figure or Tesla or whatever figures out humanoid robots, Amazon's going to buy a bajillion of them.
[11:33] You know what I mean?
[11:34] Um like immediately.
[11:36] And and so will other certain companies that are probably more tech forward that will do this.
[11:40] And so I do think there will be that we won't have scale in our homes.
[11:44] I fully agree with that.
[11:46] They will not be walking around doing our groceries, but they will have scale inside of factories um once they get to that point.
[11:51] Now, I don't know when that's going to happen, but what I want to be is is positioned in those before anyone starts to even think about that.
[11:58] And like you do not hear anyone talking about this yet, right?
[12:02] Like I never have conversations around the bottlenecks around human-aid robots with anyone or even robotaxis today.
[12:07] It's you don't see that on Twitter.
[12:09] And that's the time like there's a bit, but not much.
[12:10] And that would be the time you want to be invested.
[12:15] within those bottlenecks,
[12:17] especially for humanoids, you can look at a Nabtesco harmonic drive systems, I think from Japan, or Schaeffler, a German company.
[12:27] They're also up 70, 80, 90% this year.
[12:31] And their angle to AI is mainly humanoid robots.
[12:36] So that bottleneck already kind of started to to accelerate.
[12:40] Um but I agree with you, the real upside, similar to power semiconductors or the chemical side of things, is coming down the road.
[12:54] Um but again, the the all money was already made also within those trades.
[13:00] Hm.
[13:02] Yeah.
[13:02] I think money is made in most things right now if it's AI.
[13:04] The question is are people pricing in like for me, my opinion is it's kind of like Nvidia when it first started to have its like huge run when AI became a thing, everyone was like, "Oh yeah, this is going to be cool.
[13:15] It's going to be big."
[13:16] priced, but it's just kept getting repriced for years, right?
[13:19] Cuz we're like, "Oh, actually it's way bigger than we thought.
[13:20] Oh my god, it's even bigger.
[13:23] And it's like I feel like that's That's you're going to have with robotaxis and humanoid robots is like I don't know that what's priced in today, even though they've moved a lot, is anywhere remotely close to like where this thing's actually going to go.
[13:35] So, it's going to be one of those things that year over year you're like, "Holy it's actually way bigger."
[13:38] And it just keeps growing.
[13:39] I don't know exactly what companies are going to be that.
[13:42] So, I'm not saying the ones you listed off are those and you should just buy and hold them for 10 years.
[13:45] But, that's what I want to figure out.
[13:47] Something I have on my mind as well and since we're Milk Road and we have our roots in crypto, right?
[13:53] We need to talk about that as well, I guess.
[13:56] Crypto is so dead.
[13:59] Especially if I never heard of Bankless before, but I saw it all over X and I checked those guys out.
[14:04] Now I know what they're doing.
[14:06] They were the biggest bulls on ETH ever.
[14:09] Oh, yeah.
[14:09] And now they're gone.
[14:12] I'm just curious.
[14:14] I'm I'm Sorry.
[14:16] But, I'm just curious whether this.
[14:20] Yes, the entire AI market, especially from retail side of things, sucked out the liquidity out of crypto.
[14:25] Fully agree.
[14:26] But, once we see a slight hiccup in that semiconductor compute trade, is liquidity flowing back to certain things like crypto where retail thinks there may be opportunity?
[14:41] Especially if we're going into a world where where AI agents, to make the AI connect, need some form of economic rails and infrastructure, which is crypto.
[14:51] Want to stay ahead of the biggest technological shift in history?
[14:54] Subscribe now to get insights straight from the sharpest minds in tech and finance.
[14:59] Quickly, you'll note this show's for educational purposes only.
[15:01] Nothing here is financial advice.
[15:02] Investing always carries risk.
[15:05] Never invest more than you can afford to lose.
[15:06] Thanks for tuning in.
[15:08] See you in the next one.