Gavin Baker - Nvidia v. Google, Scaling Laws, and the Economics of AI - [Invest Like the Best, EP.451]

Transcript from https://podmenti.com/t/bdcb6281f75550ef

Hello and welcome everyone. I'm Patrick O'Shaughnessy and this is Invest Like the Best.
This show is an open-ended exploration of markets, ideas, stories, and strategies that will help you better invest both your time and your money.
If you enjoy these conversations and want to go deeper, check out Colossus Review, our quarterly publication with in-depth profiles of the people shaping business and investing. You can find Colossus Review along with all of our podcasts at joincolossis.com.
Patrick O'Shaughnessy is the CEO of Positive Sum.
All opinions expressed by Patrick and podcast guests are solely their own opinions and do not reflect the opinion of Positive Some.
This podcast is for informational purposes only and should not be relied upon as a basis for investment decisions.
Clients of positive sum may maintain positions in the securities discussed in this podcast.
To learn more, visit psum.vc.
I will never forget.
When I first met Gavin Baker.
In two thousand and seventeen.
I find his interest in markets, his curiosity about the world to be as infectious as any investor that I've ever come across.
He is encyclopedic on what is going on in the world of technology today.
And I've had the good fortune
to host him every year or two on this podcast.
In this conversation, we talk about everything that interests Gavin. We talk about
NVIDIA, Google and its TPUs, the changing AI landscape, the math and business models around AI companies.
And everything in between.
We even discuss the crazy idea of data centers in space, which he communicates with his usual passion and logic.
In closing at the end of this conversation, because I've asked him my traditional closing question before, I asked him a different question, which led to a discussion of his entire investing origin story that I had never heard before.
Because Gavin is one of the most passionate thinkers and investors that I know, these conversations are always amongst my most favorite. I hope you enjoy this latest in the series of discussions with Gavin Baker.
I would love to talk about how you like in the nitty gritty process new things that come out in this AI world because it's happening so constantly.
I'm extremely interested in it and I find it very hard to keep up. And I, you know, I have a couple of blogs that I go read and friends that I call.
But like maybe let's take Gemini three as like a recent example.
When that comes out.
Take me into your office. Like what are you doing? How do you and your team process an update like that given how often these things are happening?
I mean, I think the first thing is you have to use it yourself.
And I would just say I'm amazed at how many famous and August investors
Are reaching
Really definitive conclusions about AI. Well, no, based on the free tier.
The free tier is like you're dealing with the 10 year old and you're making conclusions about the 10 year old's capabilities as an adult.
And you could just pay and I do think actually you do need to pay for the highest year.
Whether it's Jim My Ultra, Super Groc, whatever it is, you have to pay the two hundred dollar per month tiers.
Whereas those are like a fully fledged thirty, thirty five year old.
It's really hard to extrapolate from an eight or a ten year old to the thirty five year old. And yet a lot of people are doing that. And the second thing is.
There was a you know an insider post about open AI and they said to a
large degree open AI runs on Twitter Fox.
And I just think AI
Happens.
On X.
There have been some really memorable moments. Like there was a giant fight.
between the PyTorch team at Meta and the Jack's team at Google. Odd X.
And the leaders of each lab had to step in and publicly say
No one from my lab is allowed to say bad things about the other lab, and I respect them, and that is the end of that. Yeah.
The companies are all commenting on each other's posts, you know, the research papers come out.
There's a list of You know, if on planet Earth there's five hundred to a thousand people who really
really understand this and are the cutting edge of it.
Good number of them live in China. I just think you have to follow those people closely.
And I think there is incredible signal.
Everything in AI.
It's just downstream.
Of those people. Yeah.
Everything Andre Karpathi writes, you have to read it three times. Yeah.
Minimum. Yeah, it's incredible. And then I would say any time at one of those labs
The four labs that matter.
Open AI, Jim and I.
And Thropic and X AI.
Which are clearly the four leading labs. Anytime somebody from one of those labs goes on a podcast.
I just think it's so important to listen.
For me, one of the best use cases of AI.
Із то кіп одес.
Listen to a podcast.
And then if there are parts that I thought were interesting.
Just
Talk about it with AI. I think it's really important to have as little friction as possible. I'll bring it up. I can either
Press this button and pull up Grock, or I have this.
I know. It's like somebody said a
Uh on X, you know, like we
Imbued these rocks with
Crazy spells. And now we can summon.
Super intelligent genies.
On our phones over the air, you know? It's crazy. Crazy. Okay, so
So something like Gemini three comes out.
The public interpretation was
Oh, this is interesting. It seems to say something about
scaling laws and the pre-training stuff. What is your frame on like the state of general progress in frontier models in general? Like what are you watching most closely?
Yeah, well I do think Gemini three was very important'cause it showed us that scaling laws for pre-trading are intact.
They stated that unequivocally. And that's important because no one on planet Earth knows.
How or why scaling laws for pre-training work.
It's actually not a law. It's an empirical observation.
And it's an empirical observation that we've measured extremely precisely and has held for a long time.
But our understanding of scaling loss for pre-training, and maybe this is a little bit controversial with twenty percent of researchers, but probably not more than that.
It's kind of like the ancient British people's understanding of the sun, or the ancient Egyptians understanding of the sun.
They can measure it so precisely that the east west axis of the great pyramids are perfectly aligned with the equinoxes, and so are the east west axis of Stone Edge.
Perfect.
Measurement.
They didn't understand orbital mechanics. They had no idea how or why.
It rose in the east, sat in the west, and you know, moved across the horizon. There's the aliens that don't our God in a chariot. And so it's really important every time we get a confirmation of that.
So Gemini Three was very important in that way.
But I'd say I think there's been a big misunderstanding.
of maybe in the public equity investing community or the broader, more general community.
Based on the scaling laws of pre-training, there really should have been no progress in twenty four and twenty five.
And the reason for that is is
After FAI figured out how to get two hundred thousand hoppers coherent.
You had to wait for the next generation of chips.
'Cause you really can't get more than two hundred thousand hoppers coherent. And coherent just means you could just think of it as each GPU knows what every other GPU is thinking. They kind of are sharing memory, you know, they're connected, they scale up networks and scale out.
And they have to be coherent.
During the pre training process.
The reason we've had all this progress.
Maybe we could
Show like the arc A G I slide where you had
Zero to eight over four years, zero to eight percent intelligence.
And then you went from eight percent to ninety five percent in three months when the first reasoning model came out from Open AI.
We have these two new scaling laws of post training.
Which is just reinforcement learning with verified rewards.
Verified is such an important concept in AI.
Like one of Karpathi's great things was with software.
Anything you can specify you can automate.
With AI, anything you can verify, you can automate. It's such an important concept. And I think important distinction.
And then test time compute.
Ентос в хав імен' прогрес.
Since October twenty fourth through today.
was based entirely on these two new scaling laws.
And Gemini three was arguably the first test.
Since Hopper came out.
Oh
the scaling law for pre-training.
And it held, and that's great because
All these scaling laws are multiplicative, so now we're going to apply
These two new reinforcement learning verified rewards and test time compute to much better base models.
There's a lot of misunderstanding about Gemini three that I think is really important. So
The most important thing.
to conceptualize everything in AI has a struggle between Google and NVIDIA.
And Google has the TPU.
And NVIDIA has their GPUs and Google only has a GPU and they use a bunch of other chips for networking, you know, NVIDIA has the full stack.
Blackwell was delayed.
Blackwell was NVIDIA's next generation chip.
The first iteration of that was the Blackwell two hundred.
A lot of different SKUs were canceled.
And the reason for that is it was by far the most complex product transition we've ever gone through in technology.
Going from Hopper to Blackwell.
First you go from air cooled to liquid cooled.
The rack goes from weighing round numbers to thousand pounds.
To three thousand pounds.
goes from round numbers thirty kilowatts, which is thirty American homes, to 130 kilowatts, which is 130 American homes.
I analogize it to imagine if to get a new iPhone.
You had to
change all the outlets in your house to 220 volt.
Put in a Tesla power wall.
Put in a generator, put in solar panels. That's the power.
You know, put in a whole home humidification system.
And then reinforce the floor.
Because the floor can't handle this. So it was a huge product transition.
And then just the rack was so dense it was really hard for them to get the heat out.
So black wells have only really started to be
deployed in really skilled deployments.
Over the last three or four months.
Can you explain why it has been such an important thing that Blackwell was delayed?
This black hole so complicated and it was so hard.
For everyone to get these exquisitely complex racks.
Working.
Consistently.
Had reasoning not come along
There would have been
No ei progres from mid two taus twenty four.
Through essentially Gemini three.
There would have been none.
Everything would have stalled. And can you imagine what that would have meant to the markets?
For sure we would have lived in a very different environment. So reasoning kind of bridg this eighteen month gap.
Reasoning kind of saved AI because it let AI make progress.
Without Blackwell or the next generation of TPU, which were necessary.
For the scaling laws for pre trading to continue.
Google came out with uh TPU V six in two thousand twenty four and the TPU V seven in two thousand twenty five.
In semiconductor time.
Imagine like Hopper.
It's like a World War Two era airplane.
And it was by far the best World War II era airplane. It's a P fifty one Mustang.
With the Merlin engine.
And two years later in semiconductor time, that's like
You're an F four phantom. Okay. Because Blackwell was such a complicated product and so hard to ramp.
Google was training Gemini three.
on twenty four and twenty five era TPUs, which are like
Four phantoms.
Blackwell?
It's like an F thirty five.
It just took a really long time to get it going.
So I think Google for sure has this temporary advantage right now.
From pre-training perspective.
I think it's also important.
That they've been the lowest cost producer of tokens.
And this is really important because
AI is the first time in my career as a tech investor that being the low cost producers ever matter.
Apple is not worth trillions because they're a low cost producer of phones.
Microsoft is not worth trillions because they're a little low cost producer of software.
And Video's not worth trillions because they're the low cost producer of AI accelerators.
It's never mattered.
And this is really important because what Google has been doing
has the low cost producer.
is they have been sucking the economic oxygen.
out of the AI ecosystem, which is an extremely rational strategy for them.
And for anyone who's a low cost producer, let's make life really hard.
For our competitors.
So what happens now? I think it's has pretty profound implications. One, we'll see the first models trained on Blackwell.
in early two thousand twenty six.
I think the first black ball model will come from X AI.
And the reason for that is just according to Jensen.
No one builds data centers faster than ELA.
Jensen has said this on the record.
And even once you have the black wells.
It takes six to nine months to get them performing at the level of Hopper.
'Cause hopper's finely tuned. Everybody knows how to use it. The software's perfect for it. The engineers know all its quirks.
Everybody knows how to architect a Hopper data center at this point. And by the way, when Hopper came out
It took six to twelve months for it to really outperform MPER, which was generation before. So
If your Jensen or NVIDIA, you need to get.
has many GPUs deployed in one data center as fast as possible in a coherent cluster.
So you can work out the bugs. And so this is what XAI effectively does for NVIDIA, because they build the data centers the fastest.
They can deploy.
Blackwells that scale the fastest.
And they can help work with NVIDIA to work out the bugs for everyone else.
So because they're the fastest.
they'll have the first black well model.
We know that
Scaling loss for pre training are intact.
And this means the black wall models are gonna be amazing.
Blackwell is
I mean it's not enough thirty five first enough for Phantom.
But from my perspective it is
A better chip.
You know, maybe it's like an F thirty five versus a Raphael. And so now that we know pre-scaling loss holding, we know that these black well models are gonna be really good.
Based on the Raw specs, they should probably be better.
Then something even more important happens.
So the G B two hundred was really hard to get a coin.
The G B three hundred.
Is a great chip.
It is drop-in compatible in every way with those GB two hundred racks. Now you're not gonna replace the GP two hundreds. Yeah, but just any data powerwalls, yeah. Yeah, just any data center that can handle those. You can slot in the GP three hundreds.
And now everybody's good at making those racks and you know how to get the heat out. You know how to cool them.
You're gonna put those GP three hundreds in.
And then the companies that use the G B three hundreds.
They're going to be the low cost producer of tokens.
Particularly if you're vertically integrated. If you're paying a margin to someone else to make those tokens, you're probably not gonna be.
I think this has pretty profound implications.
I think it has to change Google's strategic calculus.
If you have a decisive cost advantage.
And you're Google and you have search and all these other businesses.
Why not run AI?
at a negative thirty percent margin.
It is by far the rational decision.
Take the economic oxygen out of the environment. You eventually make it hard for your competitors.
Who need funding unlike you.
to raise the capital they need.
And then on the other side of that.
Maybe have an extremely dominant share position. Well, all that calculus changes.
Once Google is no longer
the low cost producer, which I think will be the case.
The black wells are now being used for training.
And then when that model is trained, do you start shifting black well clusters over to inference? And then all these cost calculations and these dynamics change.
It's very interesting. Like during the strategic and economic calculations between the players, I've never seen anything like it.
Everyone understands their position on the board, what the prize is.
what play their opponents are running and it's really interesting to watch.
If Google changes its behaviour.
'Cause it's gonna be really painful for them as a higher cost producer to run that negative thirty percent margin.
it might start to impact their stock.
That has pretty profound implications for the economics of AI.
And then when Ruben comes out.
The gap is gonna expand significantly. Versus TPUs? Versus TPUs and and
All other ASICs.
No, I think training three is probably gonna be pretty good and training four are gonna be good. Why is that the case? Why won't TPU V eight V nine be every bit as good?
A couple of things. So one, for whatever reason, Google made more conservative decide decisions.
A part of that is so Google, let's say the TPU.
So there's front end and back end.
of semiconductor design.
And then there's dealing with Taiwan Sunday.
And
You can make an ASIC in a lot of ways. What Google does is they
Do mostly the front end for the TPU.
And then Broadcom does the back end and manages Taiwan everything.
It's a crude analogy, but the front end is like the architect of a house.
They design the house. The back end is the person who builds the house.
And imagine Taiwan semi.
It's like stamping out that house like Lenar or you know DR Horton. And
for doing those two ladder parts.
Broadcom earns a fifty to a fifty five percent gross margin. We don't know what on TP use.
Let's say in two thousand and twenty seven.
T PU I think it sits this estimate maybe somewhere around thirty billion again. Who knows? Thirty billion, I think is a reasonable estimate.
Fifty fifty five percent.
Gross margins.
So Google is paying Broadcom fifteen billion dollars.
That's a lot of money.
At a certain point it makes sense to bring a semiconductor program entirely in house.
So in other words, Apple does not have an ASIC partner.
For their chips.
They do the front end themselves, the back end, and they manage Taiwan's.
And the reason is they don't want to pay that fifty percent margin.
So at a certain point it becomes rational to re negotiate this, and just as perspective the entire
Apex.
of Broadcom's semiconductor division is round numbers five billion dollars.
So it'd be economically rational.
Now that Google's paying if it's thirty billion, we're paying'em fifteen.
Google can go to every person who works in Broadcom Simi, double their comp.
And make an extra five billion.
In two thousand twenty eight, let's just say it does fifty billion. Now it's twenty five billion. You could triple their comp, but by the way, you don't need them all. Yeah.
And of course they're not gonna do that'cause of competitive concerns.
Well, with TPU V eight and V nine, all of this is beginning to have an impact because Google is bringing in media tech.
This is maybe the first way.
You send a warning shot to Broadcom, we're really not happy.
About.
But they did bring media tech in and the Taiwanese basic companies have much lower gross margins. So this is kind of the first shot against the battle.
And then there's all this stuff people say, but Broadcom has the best 30s.
Broadcom has really good Certies and Certies is like an extremely foundational technology because it's how the chips
Communicate with each other.
You have to serialize and deserialize.
But there are other good Certies providers in the world.
A really good Cirties is maybe it's worth ten or fifteen billion a year, but it's probably worth twenty five billion a year.
So because of that friction.
And I think conservative design choices on the part of Google.
And maybe the reason they made those conservative design choices is because they were going to a bifurcated
Supply.
T PE was slowing down, I would say
As the GPUs are accelerating. This is the first.
competitive response of Lisa and Jensen to everybody saying we're gonna have our own ASIC.
is hey we're just gonna accelerate.
We're gonna do a GPU every year and you cannot keep up with us.
And then I think what everybody is learning is like, Oh wow.
That's so cool. You made your own accelerator, has an ASIC.
Well, what's the Nick gonna be? What's the CPU gonna be? What's the scale up switch gonna be?
What's the scale of protocol?
What's the scale out switch?
What kind of optics do you use?
What's the software that's gonna make all this work together?
And then it's like oh shit, I made this.
Tiny little chip. And you know, like
Whether it's admitted or not, I'm sure the GPU makers don't love it when their customers make ASICs to try and compete with them.
And like, whoops. What did I do? I thought this was easy. I you know.
It takes at least three generations to make a good ship.
Like the TPU V one, I mean it was an achievement and then they made it.
It was really not till TPU V three or V four.
that the TPU started to become like even vaguely competitive. Is that just a classic like learning by doing things.
And even if you've made
From my perspective, the best ASIC team at any semiconductor company is actually the Amazon ASIC team.
They're the first one to make the graviton CPU.
They have this Nitro. It's called SuperNick. They've been extremely innovative, really clever.
And like
Training and infantry one.
Maybe they're a little better than the TPU V one, but only a little.
Training two, you get a little better.
Trading three.
It's I think the first time it's like okay.
And then you know I think training for will probably be good.
I will be surprised if there are a lot of
A six.
Other than training and TP.
And by the way, M Tranium and TPU will both run.
on customer owned tooling at some point. We can debate when that will happen.
But
The economics of success that I just described.
Mini's inevitable, like no matter what the companies say.
Just the economics make it and reasoning from first principles make it.
Absolutely inevitable.
If I were to zoom all the way out on this stuff, I find these details unbelievably interesting. And it's like the grandest game that's ever been played. It's so crazy and so fun to follow.
Sometimes I forget to zoom out and say, Well, so what? Like, okay, so project this forward.
three generations past Ruben or whatever. What is like the global human dividend of all this crazy development where we keep making the loss lower on these pre-training scaling models, like who cares? Like
It's been a while since I've asked this thing something that I wasn't kind of blown away by the answer for me personally.
What are the next couple of things that all this crazy infrastructure war
Allows us to unlock.
because they're so successful.
If I were to posit like an event path, I think the black hole models are gonna be amazing.
the dramatic reduction in per token cost enabled by the GP three hundred.
And probably more the MI four fifty than the MI355.
will lead to these models being allowed to think for much longer.
Which means they're going to be able to
Do new things.
I was very impressed, J my three made me a restaurant reservation.
So first time
It's done something for me.
And I mean other than like go research something and teach me stuff.
If you can make a restaurant reservation, you're not that far from being able to make a hotel reservation.
And an airplane reservation.
and order me an Uber. And all of a sudden you got an assistant. Yeah. And you could just imagine everybody talks about that, but you can just imagine it's on your phone. I think that's pretty near term.
But some big companies that are very tech forward.
Fifty percent plus of customer support is already done by AI and that's a four hundred billion dollar industry.
And then if, you know, what AI is great about is persuasion, that's sales and customer support.
And so of the functions of a company, if you think about them, they're make stuff, sell stuff, and then support the customers.
So right now maybe you
In late twenty six, you're gonna be pretty good at two of them.
I do think it's gonna have a big impact on media. Like I think robotics, you know, we talked about the last time are gonna finally start to be real. You know, there's an explosion kind of exciting.
robotic startups, I do still think that the main battle is gonna be between
Teslas Optimus and the Chinese'cause you know, it's easy to make prototypes. It's hard to mass produce them.
But then it goes back to that what Andre Karpathy said about AI can automate anything that can be verified.
So any function.
where there's a right or wrong answer or a right or wrong outcome.
You can apply reinforcement learning.
And make the AI really good at that. What are your favorite examples of that?
So far.
Or theoretically.
I mean, just does the model balance. They'll be really good at making models.
Do all the books globally reconcile.
They'll be really good at accounting. Double entry bookkeeping and has to balance. There's a verifiable you got it right or wrong.
Support or sale. Did you make the sale or not?
That's just like Alpha Go.
Did you win or you lose?
Did the guy
Convert or not.
Did the customer ask for an escalation during customer support or not?
It's most important functions are important because they can be verified. So I think if
All of this
Starts to happen.
It starts to happen in twenty six.
They'll be in ROI on Blackwell and then all this will continue.
And then we'll have Ruben.
And then that'll be another big quantum of spin, Ruben and the MI four fifty and the TPU V nine.
And then I do think just the most interesting question.
is what are the economic returns to artificial superintelligence?
Because all of these companies in this great game, they've been in a prisoners dilemma.
They're terrified that if they slow down gone forever.
And their competitors don't.
It's an existential risk, and you know, Microsoft blinked.
For like six weeks.
earlier this year. Yeah. But I think they would say they regret that.
But with Blackwell and for sure with Rubin.
The economics are going to dominate the prisoners' dilemma.
from a decision making and spinning perspective just because the numbers are so big.
And this goes to kind of the ROI on AI question.
And the ROI on AI has
Empirically, factually.
Unambiguously been positive.
I just always find it strange that there's any debate about this.
Because the largest spiders on GPUs are public companies.
They report something called audited quarterly financials. And you can use those things to calculate something called a return on investing capital.
And if you do that calculation, the ROI C of the big public spenders on GPUs is higher than it was before they ramp spending.
And you could say, Well, part of that is, you know, op ex savings. Well
At some level.
That is part of what you expect the ROI to be from AI.
And then you say, well, a lot of us actually just applying GPUs, moving the big recommender systems that power the advertising.
And the recommendation systems from CPUs to GPUs, and you've had massive efficiency gains.
And that's why all the revenue growth at these companies has accelerated. But like so what? The ROI has been there.
But it is interesting, like every big internet company.
The people who are responsible for the revenue.
Are intensely annoyed.
at the amount of GPUs that are being given to the researchers.
It's a very linear equation. If you give me more GPUs, I will drive more revenue. Yeah.
Give me those GPUs. We'll have more revenue, more gross profit, and then we can spend money. So it's this constant fight at every company.
One of the factors in the prisoners dilemma.
is everybody has this like religious belief.
Then we're gonna get to ASI.
And at the end of the day, what do they all want?
Almost all of them want to live forever. And they think that ASI is going to help them with that.
Right. Good return. That's a good return. But we don't know.
And if as humans we have pushed the boundaries
Of physics, biology, and chemistry, the natural laws that govern the universe.
Then maybe the economic returns to ASI aren't that high.
I'm very curious about your favorite
sort of throw cold water on this stuff type takes that you think about sometimes.
One would be like the things that would cause I'm curious what you think the things that would cause this demand for compute to change, or even the trajectory of it to change. There's one
Really obvious.
Yeah.
And it is just edge AI. And it's connected to the economic returns to ASI.
In three years.
On a bigger and bulkier phone to fit the amount of D RAM necessary.
Yeah, and the battery won't probably last as long.
You will be able to probably run like a
pruned down version of something.
Like Gemini Five or Grok Four, Groc 4.1.
Or chat GPT at thirty, sixty tokens per second.
And then that's free.
And this is clearly Apple's strategy. It's just we're gonna be a distributor of AI.
And we're gonna make it privacy safe and run on the phone.
And then you can call one of the big models, you know, the the God models in the cloud.
Whatever you have a question. And if that happens, if like
thirty sixty tokens a second at a one fifteen.
IQ is good enough.
I think that's
A bear case.
Other than just
The scaling laws break.
But in terms of if we assume scaling laws continue.
And we now know they're gonna continue for pre training for at least one more generation.
And we're very early in the two new scaling laws.
for post training, mid training, RL VR, whatever people want to call it, and then test time computed inference. We're so early in those and we're getting so much better.
at helping the models hold more and more context in their minds.
as they do this test time compute. And that's really powerful because
Everybody's like, Well, how's the model gonna know this? Well eventually if you can hold enough context.
You can just hold every slack message.
And Outlook Message and Company Manual.
In in a company.
In your context.
And then you can compute the new task.
And compare it with your knowledge of the world, what you think, what the model thinks.
All this context and
But it may be that like just really, really long context windows are the solution to a lot of the current limitations.
And that's enabled by all these cool tricks like K V cache offload and stuff.
But I do think other than scaling loss slowing down, other than there being
low economic returns to ASI.
Edge AI is to me
By far the most plausible.
And scary spare case.
I like to visualize like different S curves you invested through the iPhone and I love to like
see the visual of the iPhone models as it sort of went from this clunky, bricky thing up to the what we have now. We're like each one's like a little bit, you know, obviously we've sort of petered out on its form factor.
If you picture something similar for the frontier models themselves.
Does it feel like it's at a certain part of that natural technology paradigm progression?
If you're paying for Gemini Ultra or Super Groc.
And you're getting the good AI.
It's hard to see differences. Like I have to go
Really
Deep.
On something like
Do you think PCI Express or Ethernet is a better protocol for scale up networking and why? Show me the scientific papers.
And if you shift between models and you ask a question like that where you know it really deeply, then you see differences.
I do play fantasy football. Winnings are donated to charity, but it is like
You know, these new models are

Who should I play?
They think in much more sophisticated ways.
If you're a historically good fantasy football player
And you're having a bad season. This is why. This is why.
Because you're not using it. Yeah. You know? And I think we'll see that in more and more domains.
But I do think they are already at a level.
Where unless
You are a true expert, or just have an intellect that is beyond mind. It's hard to see.
The progress. And that's why I do think we need to shift from getting more intelligent
To more useful.
Unless more intelligence starts leading to these massive scientific breakthroughs.
And we're curing cancer in twenty six and twenty seven. Yeah. I don't know that we're gonna be curing cancer, but I do think from almost an ROIS curve.
We need to kind of hand off from intelligence to usefulness.
And then
Usefulness will then have to hand off.
To scientific breakthrough.
Just that creates whole new industries.
What are the building blocks of usefulness in your mind?
Just being able to do things consistently and reliably.
And a lot of that is
keeping all the context. Like there's a lot of context if someone wants to
Planet trip for me. Like, you know, I've I've acquired these strange preferences. Like I follow that guy, Andrew Huberman. So I like to have an east facing balcony so I can get morning sun. You know, the AI has to remember, here's how I like to fly. Here are my preferences for that.
Being on a plane with Starlink is important to me. Okay, here are the resorts I've historically liked, here are the kinds of areas I've liked, here are the rooms that I would really like at each. That's a lot of context. And to keep all of that and kind of weight those.
It's a hard problem.
So I think context windows are a big part of it. You know, there's this meter task evaluation thing. How long it can work for. How long it can work for. And y you you could think of that as being related in some way to context.
Not precisely.
But that just Task length needs to keep expanding because
Booking a restaurant and booking
is economically useful.
Exactly.
That economically useful.
But booking me an entire vacation
And knowing the preferences of my parents, my sister, my niece, and my nephew.
That's a much harder problem and that's something that like a human
Might have been
three or four hours on optimizing that.
And then if you can do that, that's amazing.
But then again, I just think
It has to be good at sales.
And customer support.
Relatively soon.
And then after that, it has to be in I think it is already here.
I do think we're gonna see an a kind of an acceleration in the awesomeness of various products.
engineers are using AI to make products better and faster. We both invested in Fortel, the hearing company, which is just absolutely remarkable. Like something I never would have thought of. And we're gonna see I think something like that in every vertical, and that's AI.
being used for the most core function.
of any company which is designing the product.
And then it will be, you know, there's already lots of examples of AI being used to help manufacture the product.
and distribute it more efficiently, whether it's optimizing a supply chain.
Having a vision system, watch a production line.
A lot of stuff is happening.
The other thing I think is really interesting in this whole ROI part is Fortune 500 companies are always the last to adopt a new technology. They're conservative, they have lots of regulations, lots of lawyers.
Startups are always the first.
So let's think about the cloud.
Which was the last truly transformative new technology for enterprises.
Being able to have all of your compute in the cloud.
And use SAS. So it's always upgraded. It's always great, et cetera, et cetera. You can get it on every device. I think the first AWS reInvent, I think was in 2013.
And by two thousand and fourteen
Every startup on planet Earth ran on the cloud.
The idea that you would buy your own server and storage box and router was ridiculous. And that probably happened like even earlier. That that had probably already happened before the first reInvent.
the first big Fortune five hundred company started to standardized on it like maybe five years later.
You see that with AI. I'm sure you've seen this in your startups. And I think one reason VCs are more broadly bullish on AI than public market investors.
Is VCs see very real productivity gains. There's all these charts that for a given level of revenue.
A company today.
has significantly lower employees than a company of two years ago.
And the reason is the AI is doing a lot of the
Sales.
the support and helping to make the product.
I mean there is you know iconic has some charts, A six G Z, by the way, David George is a good friend, great guy.
Yeah, he has this model busters thing.
So there's very clear data that this is happening. So people who have a lens into the world of venture see this.
And I do think it was very important in the third quarter. This is the first quarter where we had Fortune five hundred companies.
outside of the tech industry give specific quantitative examples of AI driven uplift.
So CH Robinson went up something like 20%.
On next.
Should I tell people what Siege Robinson does? Yeah.
Like let's just say a truck goes from
Chicago to Denver.
And then the trucker lives in Chicago, so it's gonna go back from Denver to Chicago.
There's
An empty load. It's CH Robinson has all these relationships with these truckers and trucking companies.
And they match shippers.
demand with that empty load supply to make the trucking more efficient. You know, they're a free forwarder, you know, there's actually lots of companies like this.
But they're the biggest and most dominant.
So one of the most important things they do.
is they quote price and availability. So somebody, a customer calls them up and says, Hey, I urgently need
Three eighteen wheelers from
Chicago to Denver.
But in the past they said it would take them, you know, fifteen to forty five minutes.
And they only quoted sixty percent.
Of inbound requests.
With AI, they're quoting 100% and doing it in seconds.
And so they printed a great quarter and the stock went up twenty percent and it was because of AI driven productivity.
That's impacting the revenue line, the cost line, everything.
I was actually very worried about the idea that we might have this Blackwell ROI air gap.
Because we're spending so much money on Blackwell.
Those black wells are being used for training.
And there's no ROI on training. The training is you're making the model. The ROI comes from inference. So I was really worried that, you know, we're gonna have maybe this three quarter period where the capex is unimaginably high.
Those black holes are only being used for training. Right. R staying flat, eyes going up. Yeah, exactly. So R O Y C goes down.
And you could see like meta, meta they printed, you know, because meta has not been able to make a frontier model. Meta printed a quarter where ROIC declined, and that was not good for the stocks.
I was really worried about that. I do think that those data points are important in terms of suggesting that maybe we'll be able to navigate this potential air gap and ROIC. Yeah. It makes me wonder about in this market, I'm like everybody else, it's the 10 companies at the top that are all the market cap more than all of the attention.
There's four hundred and ninety other companies in the SP five hundred. You study those too. Like what do you think about that group? Like what is interesting to you about the group that now nobody seems to talk about and no one really seems to care about because they haven't driven returns and they're a smaller percent of the overall index. I think that people are gonna start to care if you have more and more companies print these CH Robinson like
quarters.
I think the companies that have historically been really well run
The reason they have a long track record of success.
You cannot succeed without using technology well.
And so if you have a kind of internal culture of experimentation and innovation, I think you will do well.
With AI.
I would bet on the best investment banks.
To be earlier and better adopters of AI than
Maybe some of the trailing banks just sometimes passed his prologue.
And I think it's likely to be in this case one strong opinion I have.
All these V Cs are setting up these holding companies and you know, we're gonna use AI to make
traditional business is better.
And they're really smart VCs and they're great track records.
But that's what private equity's been doing for fifty years.
You're just not gonna be
private equity at their game.
This is what Vista did in the early days, right? Yeah. And I do think this is actually private equity's maybe had a little bit of a tough run. Just multiples have gone up. Now private assets are more expensive. The cost of financing has gone up.
It's tough to take a company public.
Because the public valuation is thirty percent lower than the private valuation. So P's had a tough run.
I actually think these private equity firms are gonna be pretty good.
That's systematically applying AI.
We haven't spent much time talking about meta, anthropic, or open AI. And I'd love your impression on everything that's going on in this infrastructure side that we talked about. These are three really important players in this grand game.
How does all of this development that we've discussed so far impact those players specifically?
First thing let me just say about frontier models broadly. Yeah.
In two thousand twenty three and twenty four I was fond of
quoting Eric Fisher, yeah.
And Eric Fisher's statement, our friend. Brilliant man.
And Eric would always say foundation models are the fastest appreciating assets in history.
And I would say he was ninety percent right. I modified the statement. I said foundation models.
Without unique data.
And internet scale distribution are the fastest depreciating assets in history.
And reasoning fundamentally changed that in a really profound way.
So there was a loop.
Flywheel to quote Jeff Bezos, that it was at the heart.
Of every great internet company. And it was you made a good product.
You got users.
Those users using the product generated data that could be fed back into the product to make it better.
And that flywheel has been spinning at Netflix, at Amazon, at Meta.
Google.
For over a decade.
And that's an incredibly powerful flywheel.
And it's why those internet businesses were so tough to compete with.
It's why there are increasing returns to scale.
Everybody talks about network effects.
They were important for social networks.
I don't know to what extent Meta is a social network anymore. It's more like a content distribution.
But they just had increasing returns to scale because of that.
Five we all.
And that dynamic was not present.
In the pre reasoning world of AI.
You pre-trained a model?
You let it out in the world.
And it was what it was.
And it was actually pretty hard.
They would do RLHF, reinforcement learning with human feedback.
And you try and make the bot model better and maybe you'd get a sense from Twitter vibes that
People didn't like this and so you tweak it.
There are the little up and down arrows, but it was actually pretty hard to feed that back into the model.
With reasoning.
It's early.
But that flywheel started to spin.
And that is
Really profound.
For these frontier laps.
So one reasoning fundamentally changed the industry dynamics of frontier lines. Just explain why specifically that is, like what is going on. Because if a lot of people
Are asking a similar question.
They're consistently either liking or not liking the answer.
Then you can kind of
Use that.
like that as a verifiable reward, that's a good outcome. And then you can kind of feed
those good answers back into the model. And we're very ear at this flywheel spinning. Yeah. Got it. Like it's hard to do now.
But you can see it beginning to spit.
So this is important fact number one for all of those dynamics.
Second.
I think it's really important that Meta, you know, Mark Zuckerberg at the beginning of this year in January
Said.
I'm highly confident I'm gonna get the quote wrong.
That at some point in two thousand twenty five we're gonna have the best and most performant AI.
I don't know if he's in the top hundred.
So he was as wrong as it was possible to be.
And I think that is a really important fact.
Because it suggests that what these four companies have done is really hard to do.
Because Meta threw a lot of money at it.
And they failed.
Yan Lakun had to leave.
They had to have the famous billion dollar
For AI researchers.
By the way, Microsoft also failed.
They did not make such an unequivocal prediction.
But they bought inflection AI.
And there were a lot of comments from them that we anticipate our internal models quickly getting better and we're gonna run more and more of our AI on our internal models.
Nope.
Amazon. They bought a company called Adept AI.
They have their models called Nova.
I don't think they're in the top twenty.

So clearly it's much harder to do than people thought a year ago.
And there's many, many reasons for that. Like it's actually really hard to keep a big cluster of GPUs coherent.
A lot of these companies
We're used to running their infrastructure to optimize for costs.
Instead of
Performance. Complexity and performance. Complexity and keeping the GPUs
Running at high utilization rate.
In a big cluster.
Actually really hard.
And there are
wild variations in how well companies run GPUs.
If the most anybody because the laws of physics, you know, maybe you can get two or three hundred thousand black wells coherent. We'll see.
But if you have thirty percent uptime on that cluster and you're competing with somebody who has ninety percent uptime.
You're not even competing.
So one, there's a huge spectrum in how well people run GPUs.
two, then I think there is, you know, these AI researchers, they like to talk about taste.
I find it very funny. You know, oh why do you make so much money? I have very good taste. You know.
What taste means is you have a good intuitive sense.
The experiments to perform.
This is why you pay people a lot of money.
Because it actually turns out that as these models get bigger, you can no longer run an experiment on a thousand GPO cluster and replicate it on a hundred thousand GPs.
You need to run that experiment.
On fifty thousand GPUs and maybe it takes, you know, days.
And so there's a very high opportunity cost.
You have to have a really good team.
That can make the right decisions about which experiments to run on this.
And then you need to do all the reinforcement learning during post training well and the test time compute well.
It's really hard to do. And everybody thinks it's easy, but all those things, you know, I used to have the saying like
Yeah, I was a retail analyst long ago.
Pick any vertical in America.
If you can just run a thousand stores in fifty states
And have them clean.
Well lit.
Stocked with
Relevant goods.
At good prices.
And staffed by friendly employees who
are not stealing from you. You're gonna be a twenty billion dollar company, a thirty billion dollar company.
Like fifteen companies have been able to do that. It's really hard.
And it's the same thing, doing all of these things well.
is really hard.
And then reasoning
With this flywheel, this is beginning to create barriers to entry.
And what's even more important, every one of those labs
X AI.
Gemini.
Open AI and Anthropic.
They have a more advanced checkpoint.
Internally.
Of the model. Checkpoint is just kind of continuously working on these models and then you release kind of a checkpoint. And then the reason these models get fast. Yeah, the one they're using internally is for the better and they're using that model to train the next model.
And if you do not have that latest checkpoint, it's getting really hard to catch up.
Chinese open source is a gift from God to meta.
Because you can use Chinese open source.
That can be your checkpoint and you can use that.
As a way to kind of bootstrap this.
And that's what I'm sure they're trying to do and everybody else.
The big problem and the big a giant swing factor, I think China's made a terrible mistake with this rare earth thing.
So China because you know they have
Huawei is in and it's a decent chip versus something like the deprecated hop preserve, it looks okay.
And so they're trying to force Chinese open source to use their Chinese chips.
They're domestically designed ships.
Problem is Blackwell's gonna come out now.
And the gap.
between these American frontier labs
And Chinese open source is gonna blow out.
Because of Blackwell.
And actually deep seek in their most recent technical paper V three point two.
said one of the reasons we struggle to compete with the American frontier labs is we don't have enough compute.
That was their very politically correct. Still a little bit risky way of saying
Because China said we don't want the black wells. Right. And they're saying some white wells. That might be a big mistake.
So if you just kinda play this out.
These
four American labs are gonna start to widen their gap for Chinese open source, which then makes it harder for anyone else to catch up.
Because that gap is growing. So you can't use Chinese open source to bootstrap.
Then geopolitically.
China thought they had the leverage.
They're gonna realize, oh, whoopsie daisy, we do need the Black Wells.
And unfortunately the problem for them.
They're probably realized that in late twenty six.
And at that point, there's an enormous effort underway. DARPA has there's all sorts of really cool DARPA and DOD programs.
To incentivise really clever technological solutions for rare earths and then
There's a lot of rare earth deposits in countries that are very friendly to America.
that don't mind actually refining it in the
traditional way. So I think rare earths
gonna be solved way faster than anyone thinks. You know, they're obviously not that rare. They're just misnamed. They're rare because
They're really messy to refine.
And so geopolitic, I actually think black well's pretty significant.
And it's gonna give America a lot of leverage.
As this gap widens.
And then in the context of all of that, going back to the dynamics between these companies.
I say I will be out with the first black well model and then they'll be the first ones probably using black well for inference at scale.
And I think that's an important moment for them. And by the way, it is funny, like if you go on open router, you can just look.
They have dominant share. Now open routers, whatever it is, it's one percent of API tokens.
But it's an indication.
They process one point three five trillion tokens. Google did like eight or nine hundred billion. This is like whatever it is, last seven days or last month.
Anthropic was at seven hundred billion. Like XAI is doing really, really well, and the model is fantastic. I highly recommend it.
But you'll see X AI come out with this.
Open AI will come out faster. Open AI's issue that they're trying to solve with Stargate.
It's'cause they pay a margin to people for compute.
And maybe the people who run their computer not the best at running GPUs.
They are a high cost producer of tokens.
And I think this kind of explains a lot of their
Code red recently. Yeah, well the one point four trillion dollars it's been to commit, but it's
And I think that was just like hey, they know they're gonna need to raise a lot of money.
Particularly if Google keeps its current strategy of sucking the economic oxygen out of the room.
And you know, you go from one point four trillion rough vibes, code red, like pretty fast, you know. And the reason they have a code red is because of all these dynamics.
So then they'll come out with a model.
But he
Will not have fixed their
per token cost disadvantage yet relative to
Both XAI and Google and Anthropic at that point. Anthropic is a good company. They're burning dramatically less cash.
than open AI and growing faster. So I think you have to give Anthropic a lot of credit.
And a lot of that is their relationship with Google and Amazon for the TPUs and the trainings. And so Enthropic has been able to benefit from the same dynamics that Google has.
I think it's very indicative in this great game of chess.
You can look at Dario and Jensen maybe have taken a few public comments, you know, that were
Between them. A little bit of jousting.
Well Anthropic just signed the five billion dollar deal with NVIDIA.
That is because Dario is a smart man and he understands these dynamics about Blackwell and Rubid.
Relative to TPU.
So NVIDIA now goes from having two fighters, X AI and OpenAI.
Two three fighters.
That helps.
In this NVIDIA versus Google battle.
And then if meta can catch up.
That's really important.
I am sure. Nvidia is doing whatever they can to help meta.
You're running those GPUs this way.
Maybe we should twist the screw this way or turn the dial that way.
And then it will be also if Blackwell comes back to China, which it seems like it'll probably happen, that will also be very good because then
Chinese open source will be back.
I'm always so curious about the polls of things. One poll would be the other breakthroughs that you have your mind on, things in the data center that aren't chips that we've talked about before is one example.
I think the most important thing that's going to happen in the world.
In this world.
In the next
Three to four years is data centers in space.
And this has really profound implications.
For everyone building a power plant.
or a data center on planet Earth. Okay. And there is a giant gold rush into this. I haven't heard anything about this, so please. Yeah, you know it's like everybody thinks like, hey, AI is risky.
But you know what, I'm gonna build a data center. I'm gonna build a power plant that's gonna do a data center.
We will need that, but if you think about it from first principles, data centers should be in space.
What are the fundamental inputs to running a data center?
Their power
And they're cooling.
And then there are the chips.
If you think about it from a total cost perspective. Yeah.
And just the inputs to making the tokens come out of the magic machines.
In space.
You can keep a satellite in the sun twenty four hours a day.
And the sun is thirty percent more intense.
You can have the satellite always kind of catch the light.
The sun is thirty percent more intense, and this results in six times more irradiance in outer space.
than on planet earth.
You're getting a lot of solar energy.
Point number two, because you're in the sun twenty four hours a day, you don't need a battery.
This is a giant percentage of the cost.
So the lowest cost energy
Available in our solar system.
Is solar energy and space.
Second, for cooling.
In one of these racks.
A majority of the mass and the weight is cooling.
The cooling.
And these data centers is incredibly H VAC.
The CDUs, the liquid cruise. It's amazing to see.
In space, cooling is free.
You just put a radiator on the dark side of the satellite. It's fucking gold. As close to absolute zero as you could get.
So all that goes away.
And that is a vast amount of cost.
Okay.
Let's think about how these
Maybe each satellite is kind of a rock.
one way to think of it. Maybe some people make bigger satellites that are three racks.
Well, how are you going to connect those rocks? Well it's funny in the data center.
The rocks are
over a certain distance connected with fiber optics. And that just means a laser going through a cable.
The only thing faster than a laser going through a fiber optic cable.
Is a laser going through absolute vacuum. So if you can link these satellites.
In space together.
Using lasers.
You actually have a faster and more coherent network.
Than in any data center on Earth. Okay.
For training, that's gonna take a long time. Because it's so big.
Yeah, just'cause it's so big.
training will eventually happen. But then for inference, let's think about
The user experience. When I asked Rock about you and it gave the nice answer.
Here's what happened.
A radio wave traveled from my cell phone to a cell tap.
Then it hit the base station.
Went into a fiber optic cable.
Went to some sort of metro aggregation facility in New York.
Right, within like, you know, ten blocks of here.
It's a small little metro router. It's
Routed those packets.
to a big XAI data center somewhere.
The computation was done.
And it came back over the same path.
If the satellites
Can communicate directly with the phone.
And Starling has demonstrated direct to sell capability. You just go boom boom.
It's a much better lower cost.
User experience.
So in every way.
Data centers in space.
From a first principles perspective.
are superior to data centers on earth.
If we could teleport that into existence, I understand that th that portion
Why will that not happen? Is it launch cost? Is it launch availability? Is it capacity? We need a lot of the space starships. Like the starships are the only ones that can economically make that happen.
We need a lot of those starships.
Maybe China or Russia will be able to land a rocket. Blue Origin just landed a booster.
This is a big idea.
And I do think.
It's an entirely
new and different way to think about space ac.
And it is.
Interesting that
Elon posted or said in an interview.
That Tesla, SpaceX, and XAIG were converging, and they really are. So XAI
will be the intelligence module for Optimus made by Tesla with Tesla Vision.
has its perception system.
And then SpaceX.
We'll have the data centers in space.
That will power a lot of the AI, presumably for X AI.
And Tesla and the octopuses and a lot of other
companies and it's just interesting the way that they're converging.
And each one is kind of creating competitive advantage for the other.
If you're X A I.
It's really nice.
Do you have this built-in relationship with the optimist?
Tesla's a public company, so there's gonna be like I'm I cannot imagine the level of vetting that will go into that.
Intercompany agreement. And then you have a big advantage with
These data centers in space.
And then it's also nice if you're X AI.
That you have two companies with a lot of customers.
Who you can use to help build your customer support agents, your customer sales agents with.
kind of built in customers.
So they really are all kind of converging.
In a neat way. And I do think
It's gonna be a big moment when that first black well model comes out from X AI next year.
If I go to the other end of the spectrum.
And I think about something that seems to have been historically endemic to the human economic experience, that shortages are always followed by gluts in capital cycles.
What if in this case the shortage is compute, like Mark Chen now is on the record as saying they would consume 10x as much compute if you gave it to them in like a couple of weeks.
So like there seems to still be a massive shortage of compute, which is all the stuff we've talked about today.
But there also just seems to be this like iron law of history that gluts follow shortages.
What do you think about that concept as it relates to this there will eventually be a glut. AI is fundamentally different than the software, just in that every time you use AI takes compute.
In a way that traditional software just did not. I mean, it is true. Like I think every one of these companies could consume 10x more compute. Like what would happen is just
the two hundred dollar tier would get a lot better.
The free tier would get like the two hundred dollar tier. Google has started to monetize AI mode with ads.
And I think that will give everyone else permission.
to introduce ads into the free mode and then that is going to be an important source of ROI.
Open AI is tailor made to Yeah, absolutely all of them. And actions like, you know, hey
Here are your three vacations. Would you like me to book one?
And then they're for sure gonna collect a commission. Yeah.
There's many ways you can make money.
I think we went into great detail on maybe a prior podcast about how just inventory dynamics made
These inventory cycles inevitable.
And Semis.
The iron law of S means it's just that customer buffer inventories have to equal lead times.
And that's why you got these inventory cycles historically.
We haven't seen a true capacity cycle in semis.
Arguably since the late nineties, and that's because Taiwan Simi has been so good at aggregating.
Smoothing supply.
And
A big problem in the world right now is that Taiwan Cimi is not expanding capacity as fast as their customers want.
They're in the process of making a mistake just because
You do have Intel and with these fabs and they're not as good and it's really hard to work with their PDK.
But now you have this guy Lipu, who's a really good executive and really understands that business.
I mean, by the way, Patrick Elsinger, I think, was also a good executive.
And he put intel on the only strategy that could result in the success.
And I actually think it's shameful that the Intel board fired him when they did.
But Lee Boo's a good executive and now he's reaping the benefits of Patrick's strategy.
And Intel has all these empty fabs.
And eventually, given the shortages we have of compute, those fabs are going to be filled.
So I think Taiwan City's in the process of making mistakes, but they're just so
Paranoid about an overbelt.
And they're so skeptical.
They're the guys who met with Sam Altman.
And laughed and said, He's a podcast bro. He has no idea what he's talking about. You know? They're terrified of an overbelt.
So it may be the Taiwan Cime.
single handedly that their caution is the governor. I think governors are good. It's good that power's a governor. It's good that Taiwan Simony is a governor.
If Taiwan Cimi opens up
At the same time when data centers and space relieve all power constraints.
But that's like
I don't know, five, six years away.
The data centers in space are majority of deployed megawatts.
Yeah, I think you can get it overbuilt really fast, but just we have these two really powerful natural governors.
And I think that's good. Smoother and longer is good.
We haven't talked about the power other than alluding to it through the space thing. Haven't talked about power very much.
Power was like the most uninteresting topic.
Nothing really changed for like a really, really long time.
all of a sudden we're trying to figure out how to get like gigawatts here, there, and everywhere.
How do you think about are you interested in powers? I'm very interested in it. I do feel lucky
In a prior life, I was the sector leader for the telecom and utilities team. So
I do have some base level of knowledge.
Having watts as a constraint is really good for the most advanced compete players.
Because if Watts are the constraint.
The price you pay for computers irrelevant.
The TCO of your compute is absolutely irrelevant.
Because if you could get three X or four X or five X more token's
For what?
That is literally three or four X or five X more revenue.
If you're gonna build an advanced data sitter costs fift billion, a data sitter with the ASIC maybe costs thirty five billion.
With that fifty billion dollar data center.
Pubs out.
twenty five billion of revenue and you're the ASIC data center at thirty five billion.
Is only puppy out eight billion? Well, like you're pretty bumped.
So I do think
It's good for all of the most advanced
technologies of the data center, which is exciting to be as an investor.
So as long as power is a governor.
The best products are gonna win irrespective of price and have crazy pricing power.
That's the first implication that's really important to me.
Second, it is in
The only solutions to this
We just can't build nuclear fast enough in America.
As much as we would love to build nuclear quickly, we just can't.
It's just too hard.
Nepa, all these rules. Like it's just it's too hard.
Like a a rare ant.
that we could move and it could be in a better environment could totally delay the construction of a nuclear power plant. You know, one ant.
You know, that is America's frog.
It's crazy, actually. Humans need to come first. We need to have a human centric view of the world.
But the solutions are natural gas and solar.
Yeah, the great thing is the great thing about these AI data centers is apart from the ones that you're going to do inference on, you can locate them anywhere.
So I think you were going to see and you're this is why you're seeing all this activity at Abilene.
Because it's in the middle of the big natural gas basin.
And we have a lot of natural gas in America because of fracking.
We're gonna have a lot of natural gas for a long time. We can rap production really fast.
So I think this is gonna be solved. You know, you're gonna have power plants.
Fed by gas or solar.
And I think that's the solution. And you know, you're already
All these turbine manufacturers were reluctant to expand capacity, but Caterpillar just said we're gonna
Increase capacity by seventy five percent over the next few years. So like
The system on the power side is beginning to respond.
One of the reasons that I always so love talking to you is that you do as much in the top ten companies in the world as you do looking at
Brand new companies with entrepreneurs that are
twenty five years old trying to
do something amazing. And so you have this very broad sense of what's going on.
If I think about that second category of young enterprisists who now are like the first generation of AI native
Entrepreneurs.
What are you seeing in that group that's notable or surprising or interesting?
These young CEOs, they're just so impressive in all ways.
And they get more polished faster.
And I think the reason is is they're talking to the AI.
How should I deal with pitching this investor?
I'm meeting with Patrick O'Shaughnessy.
What what do you think the best ways I should pitch him are and it works. Do a deep research thing and it's good. Yeah. You know? Hey, I have this difficult HR situation.
How would you handle it?
And it's good at that.
We're struggling to sell our product. What changes would you make?
And it's really good at all of that.
Today.
And so and that goes to these
VCs are seeing massive AI productivity in all their companies.
It's'cause their companies are full of these twenty three, twenty four or
Even younger AI natives.
And they're impressive.
I've been so impressed.
With like young investment talent.
And it's just part of it, like your podcast is part of that. There's just
Very specific knowledge has became so accessible.
True podcasts and the internet.
Like kids come in and they're just I feel like they're where I was as an investor, like in my, you know, early thirties and they're 22, and I'm like, Oh my God. Yeah.
Like I have to run so fast to keep up. These kids who are growing up native in AI
They are just proficient with it in a way that I have tried really hard to become.
Can we talk about semis VC specifically and like what is interesting in that universe? One thing I just think is so cool about it and so underappreciated.
Is your average semiconductor?
Venture founder is like fifty years old.
Okay.
And Jensen and what's happened with NVIDIA and the market cap of NVIDIA.
has like single handedly ignited ignited Cductor Venture, but the way it's ignited, it's ignited in an awesome way.
That's like really good for actually NVIDIA and Google and everyone.
Is like let's just say
You are the best DSP architect in the world.
You had made
For the last twenty years.
Every two years,'cause that's what you have to do. Semiconductors.
It's like every two years you have to run a race.
And if you won the last race, you start like
A foot ahead.
Over time those compound.
It makes each race easier to win.
But maybe that person and his team, maybe he's the head of networking at a big
public company.
And he's making a lot of money and he has a good life.
And he's fifty years old.
And then because he sees these outcomes and the size of the markets of the data center, he's like, Wow, why don't I just go start my own company?
But the reason that's important
Is it you know, I forget the number, but I mean there are thousands of parts in a black well rack.
And there's thousands of parts in a TPU rack.
And and the Blackwell R
Maybe NVIDIA makes

Same thing in an AMD rack.
And they need all of those other parts to accelerate with them.
So they couldn't go to this one year cadence.
If
Everything was not.
Keeping up with them.
So I think it's
The fact that Semictor Venture has come back with a vengeance, you know, Silicon Valley stopped being Silicon Valley long ago.
My little firm maybe has done more semiconductor deals in the last seven years than the top ten VCs combined. You know?
But that's really, really important because now
You have an ecosist of companies.
Who can keep up. And then that ecosystem of these venture companies
Is putting pressure
On the public companies that are also need to part of this, if we're gonna go
To this annual cadence, which is just so hard.
But it's one reason I'm really skeptical of these A six.
That don't already have some
Degree of success.
So I do think that's a super, super important dynamic and one that's
Absolutely foundational and necessary for all of this to happen.
Because not even NVIDIA can do it alone. AMD can't do it alone.
Google can't do it alone.
You need
The people who make the transceivers, you need the people who make the wires, who make the back plates.
You make the lasers. They all have to accelerate with you.
And one thing that I think is very cool about AI as an investor is it's just it's the first time
Where every level of the stack
That I look at at least the most important competitors are public and private.
NVIDIA, they're very important.
Private competitors. Broadcom, important private competitors. Marvell, important private competitors, you know, Lumentum, Coherent, all these companies.
There's even like a wave of innovation in memory.
Which is really exciting to see because memory
It's such a gating factor, but something that could slow all this down and be a natural governor.
is if we get our first true D RAM cycle since the late nineties. Say more what that means?
If the price of D RAM
If like a D RAM wafer is valued at like
A five carat diamond.
In the nineties when you had these true capacity cycles before Taiwan Sumi kinda smoothed everything out and
D RAM became more of an oligopoly.
You would have these crazy shortages where the price would just go ten X.
Unimaginable.
relative to the last 25 years.
Where like a giant D RAM cycle, a good D ram cycle as the price starts start stops going down.
An epic cycle is maybe it goes up, you know, whatever it is, thirty, forty, fifty percent.
But I mean if it starts to go up by X's of percentages.
That's a whole different game. By the way, we should talk about SAS.
Yeah, let's talk about it. What do you think's gonna happen?
Well, I think that application task companies are making the exact same mistake.
The brick and mortar retailers did with e commerce. So brick and mortar retailers
particularly after the telecom bubble crashed. You know, they looked at Amazon and they said, Oh, it's losing money.
E commerce is going to be a low margin business. How can it ever be more efficient as a business? Right now, our customers.
pay to transport themselves to the store and then they pay to transport the goods help.
How can it ever be more efficient if we're
Sending shipments out.
To individual customers.
Amazon's vision, of course, well, eventually we're just gonna go down a street and drop off a package at every house.
And so they did not invest in e commerce. They clearly saw customer demand for it.
But they did not like the margin structure.
of e commerce.
That is the fundamental reason.
that essentially every brick and mortar retailer was really slow to invest in e commerce.
Now here we are and you know, Amazon has higher margins in their North American retail business than
a lot of retailers that are mass market retailers.
So margins can change.
And if there's
a fundamental transformative kind of new technology that customers are demanding and it's always a mistake not to embrace it.
That's exactly what the SaaS companies are doing.
They have their seventy, eighty, ninety percent gross margins.
And they are reluctant to accept AI gross margins.
The very nature of AI is, you know, software, you write it once.
And it's written very efficiently.
And then you can distribute it broadly at very low cost. And that's why it was a great business.
AI is the exact opposite, where you have to recompute the answer every time.
And so a good AI company might have gross margins of forty percent.
The crazy thing is because of those efficiency gains.
They're generating cash way earlier.
than SaaS companies did historically, but they're generating cash earlier.
Not because they have high gross margins, but because they have very few human employees.
And it's just tragic to watch all of these companies.
Like you want to have an agent.
It's never gonna succeed.
If you're not willing to run it at a sub thirty five percent gross margin.
Because that's what the AI natives are running it at. Maybe they're running it at forty.
So if you were trying to preserve an eighty percent gross margin structure
You are guaranteed that you will not succeed at AI.
Absolute guarantee.
And this is so crazy to me because one.
We have an existence proof.
For software investors be willing to tolerate
Gross margin pressure as long as gross profit dollars are okay. It's called the cloud.
People don't remember.
But with Adobe converted from odd premise.
to a SaaS model, not only did their margins implode, their actual revenues imploded too.
Because you went from charging up front.
to charge you over a period of years.
Microsoft it was less dramatic.
But you know Microsoft was a
Tough stock.
In the early
In the early days of the cloud trades they should because investors were like, Oh my God, you're an eighty percent gross margin business.
The cloud is the fifties and they're like, Well, it's gonna be gross profit dollar
creative, it probably will improve those margins over time.
Microsoft.
They bought GitHub.
They use GitHub.
As a distribution channel for co pilot for coding. That's become a giant business.
A giant business.
Now for sure.
It runs at much lower gross margins.
But there are so many SaaS companies.
Like I I can't think of a single application SaaS company.
That could not be running a successful agent strategy.
They have a giant advantage over these AI natives.
And that they have a cash generative business.
And I think there is room for someone to be a new
Kind of activist or constructivist.
And just go.
Two SaaS companies and say
Stop being so dumb.
All you have to do is say, Here are my AI revenues.
And here are my AI gross margins, and you know it's real AI because it's low gross margins. I'm gonna show you that.
And here's a venture competitor over here that's losing a lot of money.
So maybe I'll actually take my gross margins to zero for a while.
But I have this business.
that the venture funded company doesn't have.
And this is just such a like obvious
Playbook that you can run
Salesforce, service now, hub spot.
Git Lab, Atlassian.
All of them.
Could run this.
And the way that those companies could or should think about
the way to use agents is just to
Ask the question, okay, what are the core functions we do for the customer now? Like how can we further automate that with agents effectively?
If you're in CRM, well, what our customers do, they talk to their customers.
We're customer relationship management software and we do some customer support too.
So make an agent that can do that. Right. And sell that at ten to twenty percent to let that agent access all the data you have. Right. Cause what's happening right now, another agent.
Made by someone else.
Is accessing your systems to do this job. Pulling the data into their system.
And then you'll eventually be turned off.
It is just crazy. It is just cause
Oh wow, but we want to preserve our eighty percent gross margins.
This is a life or death decision.
It essentially everyone except Microsoft.
It's failing it.
Quote that Bibbo from that.
Nokia guy long ago. Like their platforms are burning. Burning platform. Yeah. Yeah there's a really nice platform right over there. And you can just hop to it.
And then you can put out the fire in your platform that's on fire.
And now you got two platforms and it's great. You know?
Your data centers and space thing makes me wonder if there are other kind of
Less disgust.
off the wall things that you're thinking about in the markets in general that we haven't talked about since it does feel like since
Twenty twenty kicked off and twenty twenty two punctured this. Kind of a series of rolling bubbles.
So in twenty twenty there is a bubble in like EV startup EVs that were not Tesla.
And that's for sure a bubble. And they all went down, you know, 99%.
And there was kind of a bubble in more speculative stocks.
Then we have the meme stocks, you know, game stop.
And now it feels like the rolling bubble is in nuclear and quantum.
And these are
fusion and SMR. It would be a transformative technology. It's amazing.
But sadly.
From my perspective, none of the public
Ways you can invest in this.
are really good expressions of this theme are likely to succeed or have any real fundamental support.
And same thing with quantum.
I've been looking at quantum for ten years. We have a really good understanding of quantum.
And the public quantum companies, again, are not the leaders.
From my perspective, the leaders in quantum.
Would be Google.
IBM and then uh Honeywell.
Quantum.
So the public ways you can invest in this theme, which probably is exciting.
Are not the best. So you have two really clear bubbles.
I also think quantum supremacy is very misunderstood.
People hear it and they think that it means that quantum computers are gonna be better than classical computers at everything. With quantum, you can do some calculations.
The classical computers.
Cannot do.
That's it.
That's gonna be really useful and exciting and awesome.
But it doesn't mean that quantum takes over the world.
I think the thought that I

related.
to markets that just AI
I have just been fascinated that for the last two years whatever
A I
Needs
To keep growing
And advancing.
It gets
Have you ever seen public opinion?
Change so fast.
In the United States on any issue has nuclear power.
Just happened like that.
Like that.
And like why did that happen right when AI needed it to happen?
Now we're running up on boundaries of power on earth.
All of a sudden
Data centers in space.
It's just a little
Strange to me.
That whenever there is something
A bottleneck that might slow it down.
Everything accelerates.
Rubin is gonna be such an easy, seamless transition relative to Blackwell, and Rubin's a great chip. And then AMD getting into the game with the MI four fifty.
Whatever AI needs, it gets.
You're a deep reader of sci-fi, so uh you're making me think of Kevin Kelly's great book, What Technology Wants. He calls it the technium, like the like the overall mass of technology that just like is supplied by humans. Absolutely. Yes, it just wants to grow more and more powerful, and now we're going into an in state.
I have a selfish closing question. Speaking of young people.
So my kids who are twelve and ten, but especially my son who's older
is developing an interest in what I do, which I think is quite natural.
And I'm gonna try to start asking my friends who are the most passionate about
entrepreneurship and investing, why they are so passionate about it and what about it is so
interesting and life giving to them.
How would you pitch
what you've done the career you built this part of the world
to a young person that's interested in this. I do believe at some level kind of investing is the search for truth. And if you find
Truth.
First.
And you're right about it being a truth. That's how you generate alpha.
It has to be a truth that other people have not yet seen.
You're searching for hidden truths.
The earliest thing.
I can remember is being interested in history.
looking at books with pictures of the Phoenicians and the Egyptians and the Greeks.
And the Romans.
I loved history.
I vividly remember I think in the second grade.
Because my dad drove me to school every day.
We went through the whole history of World War Two.
In one year.
And I love that.
And then that
translated into a real interest in current events very early. So like as a pretty young person, you know, I don't know if it was eighth grade or seventh grade or ninth grade.
I was reading the New York Times and the Washington Post.
And I would get so excited.
When the mail came, because it meant that maybe there was an economist or a newsweek or a time or US News, and I was really into current events.
You know,'cause current events is kind of like applied history and watching history happen.
Thinking about what might happen next.
I didn't know anything about investing. My parents were both attorneys. Like I was anytime I want an argument, I was super rewarded.
Like, you know, if I could make a reasonable argument why I should stay up late, my parents would be so proud and they'd let me stay up late, but it I had to beat them, you know.
That was the way I grew up. I was just kind of going through life and you know, I really love to ski and I love rock climbing.
And I go to college and rock climbing is by far the most important thing in my life.
I dedicate myself to completely. I climbed
I did all my homework at the gym. I got to the rock climbing gym at seven AM.
would skip a lot of classes to stay at the gym. I'd do my homework on like a big bouldering mat.
Every weekend I went and climbed somewhere with the Dartmouth Mountaineering Club.
It was super important to me. I'm not a good athlete, so I was never a very good climber, but I did dedicate myself
Entirely to rock climbing.
And as part of that, like on Clive Trips, the movie Rougers came out while I was in college. So we started playing poker.
I like to play chess.
I mean I was never that good at chess or poker. You know, never really dedicated myself to either.
And my plan
After two or three years of college.
was I was going to leave
I was a skipper in college. I of I was a housekeeper
Cleaned a lot of toilets.
It was shocking to me how people treated me.
And it is like
permanently impacted how I treat other people, you know.
It'd be
cleaning somebody's room and they'd be in it and they'd be reading the same book as you
And like, you know, you'd say, Oh wow, that's a great book, you know, I'm about where you are.
And A, they look at you like you're a space alien. Like you speak.
And then they get even more shocked you read. You know. So it like had a big impact on how I've like just treated everyone since then.
Like being nice is free.
But anyways, I was gonna be a ski bomb in the winters.
I was gonna work on a river in the summers, and that was how I was gonna support myself. And then I was gonna climb in the shoulder seasons.
I was gonna try and be a wildlife photographer and write the next great American novel. That was my plan.
This was like my plan of record.
I was really lucky my parents were
Very supportive of everything I wanted to do.
My parents had very strict parents. So of course they're extremely permissive with me.
So yeah, I'll probably end up being a straight parent, you know, just the cycle continues.
And my parents were lawyers, you know, they'd done
Reasonably well.
They both grew up in I'd say very economically disadvantaged circumstances.
Like my dad talks about
Like he remembers every person who bought him a beer.
You couldn't afford a beer.
He worked the whole way through college. He was there on a scholarship.
He had one pair of shoes all through high school.
And so they were super on board with this plan and I'd been very lucky they sent me to college.
And I didn't have to pay for college. So they said, you know, Gavin, we think this plan
Of being
Ski Bum, River Rafty Guide, Wildlife Photographer, Climbing the Shoulder Seasons, try to write a novel. We think it sounds like a great plan.
But you know, we've never asked you for anything.
Haven't encouraged you to study anything. We've supported you in everything you've wanted to do.
Will you please get one professional internship? Just one. And we don't care what it is.
The only internship I could get, this was at the end of my sophomore summer at Dartmouth.
It was an internship with Donaldson Lufkin and Jin Rat DLJ.
My job was to
Every time DLJ published a research report, it was in like the private wealth management division.
And I worked for the guy who ran the office.
And my job was whenever they
produced a piece of research, I would go through and look at which of his clients owned that stock.
Then I would mail it to the clients. So this day
we wrote on general electric. So I need to mail the GE report to these thirty people.
No, the Cisco report to these twenty people. And then I started like reading the reports and I was like, Oh my God.
This is like the most interesting
thing imaginable.
So investing, I kind of conceptualized it. It's a game of skill and chance, kind of like poker.
Yeah, there's obviously chance in investing.
Like if you're an investor in a company and a media hits their headquarters.
That's bad luck.
But like you own that outcome.
So there is chance that is irreducible.
But they're skill too.
But so that really appealed to me.
And the way you got an edge in this
Greatest game of skill and chance imaginable.
Was you had the most throw knowledge possible of history.
And you intersected that with the most
Accurate.
understanding of current events in the world.
to form a differential opinion on what was going to happen next in this game of skill and chance.
Which stock is mispriced in the perimutual system?
That is the stock market.
And that was
Like day three.
I went to the bookstore.
And I bought like the books that they had, which were Peter Lynch's books. I read those books in like two days. I'm a I'm a very fast reader.
Then I read all these books about Warren Buffett.
Then I read Market Wizards, then I read Warren Buffett's letters to his shareholders. This is like during my internship.
Then I read Warren Buffett's letters to your shareholders again. Then I taught myself accounting. There's this great book, why stocks go up and down. Then I went back to school. I changed my majors from English and history to history and economics.
And I never looked back in it.
consumed like I continued to really focus on climbing.
But like instead of like I would be in the gym and I would print out everything.
that the people on the motley fool wrote.
They had these fools.
They were early to talking about return on invested capital. Incremental ROIC is like a really important indicator.
And I would just read it and I would underline it and I'd read books and then I'd read the Wall Street Journal and then eventually
There was a computer terminal.
Read news about stocks. And it was the most important thing in my life and like I barely kept my grades up.
And yeah, that's how I got into it. History, current events, skill and chance.
And I am a competitive person.
And I've
Actually never been good at anything else. Okay. I got picked last for every sports team.
Like I love to ski.
I've literally spent a small fortune.
Private ski lessons I'm not that good of a skier.
Um I like to play Pig Pog.
All my friends could be.
I tried to get really good at chess and this was before the we actually had to play the games.
And my goal was to beat one of the people. I'm sure there's a park somewhere. Okay, well there's one at Carbridge, and I wanted to beat one of them.
Never beat one of them.
I've never been good at anything.
I thought I would be good at this.
And the idea of being good at something other than taking a test.
That was competitive.
was very appealing to me.
So I think that's been a really important thing too, and to this day.
This is the only thing I'm good at. I'd love to be good at something else. I'm just not. I think I'm gonna start asking this question of everybody. The ongoing education of Pearson Maeve.
Amazing place to close. I love talking about everything. Thank you so much for this. Great, man. Thank you, thank you, thank you.
If you enjoyed this episode, visit joincolossis.com where you'll find every episode of this podcast complete with hand edited transcripts. You can also subscribe to Colossus Review, our quarterly print, digital, and private audio publication featuring in-depth profiles of the founders, investors, and companies that we admire most.
Learn more at joincolossis dot com slash subscribe.