And we've kind of rediscovered HTML as a great way for agents to present information to us, right?

I agree.

It's really good.

It's really hard to share privately, which is kind of a bummer.

So you have to generate a PDF.

And there again, going back to my base programming skills, I have not figured out a way of generating a presentation that renders well as a PDF because it generates all the...

Well, anyway, whatever the margins are not properly... managed but this is an example so you know i think within two and a half minutes running on the local model i can get a sense of exactly what happened in the business right so revenue gross up 41 cash is where net dollar retention is an exceptional metric and you can see the bit they crushed um there are quotes at the bottom from the quarterly earnings reports uh and so you can just kind of go through exactly what's happened in the business I think you have a quote here from Dylan talking about how AI is really important for the business.

Do you use like 7B or even 2B for anything?

Unless, unless these vertical AI companies like, like y-you just look at them and, and, and I'm, I'm wondering how they get into the billions of dollars of, of runway-- run rate, which I think right now the industry is sort of pricing them that, that, that at least some of them need to hit that for the category to work out.

Yeah, you're right.

I mean, we've benchmarked the, you know, AI harnesses are now the fastest growing ones are trading between a hundred to a hundred and fifty times current ARR, which is an enormous multiple, particularly at that level of scale.

Um, sometimes it's subsidized with lower gross margins, and the idea there is primarily just let's get the product out as far as possible, and then ultimately we'll improve the economics.

In their favor, I will say we made a prediction at the end of twenty five that we-- twenty twenty-six would be the first year where employers would pay agents at market level for a person or more.

Mm

... to serve.

And to do that, they need the data acquisition.

That's the thing that we should be looking at and we should be thinking about it a little bit differently.

Yeah, it's a great question, right? I mean, we had the SaaSpocalypse.

Benioff went on his earnings call and said this SaaSpocalypse stuff is totally overblown.

They had an amazing quarter.

Let's put these things in historical context, right? So like up until about 2012, 2016, publicly traded software, you kind of expected a four to six X multiple.

That was what people expected when companies went public.

And then, you know, you had Zerp after 2018.

do you think the AI market can offset, like, I guess my point is, are you bullish enough on the expanded activity of agents, software, and all of this infrastructure around artificial intelligence, that if you correlated it to the number of people that per capita... output will dramatically increase.

Yeah

...

I mean, this data's unbelievably valuable.

And, uh, y- we can look at it with the model releases even from last week.

The share shifts meaningfully.

Like, Fable comes out, bunch of people move.

In fact, we published some analysis.

[laughs]

Are we seeing three unique companies go out and possibly return a lot of money to investors, and should not read into that about what it means for other companies that may want to find liquidity? Or is this more an indication that the exit market is finally, um, l- you know, de-icing itself and becoming a bit more, uh, amenable to the venture capital cycle?

I think we're gonna see broad liquidity.

I mean, uh, Reuters announced, I think this morning, that the SpaceX IPO was two and a half, maybe three times oversubscribed, which is a stunning number just considering that the sum total of those three offerings that you mentioned, SpaceX, OpenAI, and Anthropic, if they raise what they intended to, would be greater than the sum total of all IPOs dollars raised in the previous decade.

So it, it clearly is just huge demand, uh, for exposure to AI, uh, and space.

So I, I think that's, that's really telling.

And then you're starting to see some other S1s, right? Bending Spoons, uh, came out, which is, um, a holding company.

They bought AOL, which kind of blew my mind, and that business is doing incredibly well on the back of AI.

Yeah.

I mean, what's your reaction to this? Like, is this conceivably very dangerous? I mean, does Anthropic, they're going to have to get this model out at some point, right?

I don't know.

I think, well, I suspect there's an era where but you have to be approved to use certain models.

This model has a certain designation, which means it can be a weapon system and you don't deliver that to everybody.

There are certain laws around which weapons can be sold to which countries and which kinds of companies.

There may be a class of models that are so important and so powerful that only a handful of companies have them.

I think there's a big question here about Can these models be used only for defensive purposes rather than being built in products? Because, okay, the US government will have access to this model.

So is cybersecurity entirely immune to the SaaSpocalypse and the threat that AI can cannibalize their business?

I mean, I just wondered if you see them at all diversifying in the near term or the long term if you know how much um this really uh matters in the story people are

trying to diversify so there's a couple of different layers of the stack that are important there's the chip itself and then there's the software to run that chip and their nvidia has had a lock for a while with the technology called cuda which is super high performance and Meta has developed their own equivalent layer.

So they have their NTIA chips.

And then on top of that, they have a layer that's called Triton.

And Triton is getting close to the level of performance, but it's still quite early.

AMD has an equivalent layer.

I think this is one of the reasons that Nvidia has such dominant share.

What does this sort of signal to you about the custom silicone story? I

Why Qowa among all the other companies that are in the space?

Yeah, great question.

I mean, you look at the search ads market, it's about two hundred fifty billion in size.

You look at the social media ads market, it's two hundred sixty-five last year.

It was the first time social surpassed search.

And both of those, and most online advertising, is predicated on the idea that there's some data that you can use to target, uh, individuals.

And the incredible part about AI, as we all know, is the amount of information that we share and the length of the queries that are going into AI.

What do you think is gonna be different about how this is built out and all the ad tech around it?

What do you think investors are reacting to?

Well, yeah, the stock is down a lot, 10%.

I think you're seeing 72% of software stocks this morning are trading down.

And Microsoft is top three in terms of overall decrease.

There's a difference in the way that Meta and Microsoft were perceived.

Meta's core business is accelerating, is growing faster.

Azure is still relatively constant.

What do you think?

And based on the revenue estimates, Broadcom is actually trading higher, which I think might be what people were getting at.

I think that's right.

I mean, you have very elevated multiples.

And I mean, we know from all the previous hardware cycles, whether it's telecom or also hardware, hard disks, that any miss here, you have a very significant drawdown.

And because of the capex here, so intense.

So if there's any weakness whatsoever or any perceived sign of potential weakness in the future, I mean, you saw it with Oracle earlier this week where the RPO was absolutely massive, but the timing of that RPO remains unclear and will people actually end up using this? I think that's the major factor that people are trying to understand.

Okay, OpenAI is committed to spending a particular amount.

What do people make of his leadership and the way that he's running the company?

Aaron Wu.

Erin Wu.

Yeah, so she mentioned that there's this notion of remaining performance obligations or what Google calls backlog.

And those are customer commitments to spend credits on AI over some particular period of time.

And the combination of those two companies yesterday, 555 billion of RPO, most the average duration for Microsoft's RPO is two years.

And so Microsoft has a tremendous sense of confidence when those dollars will be spent.

uh and underpinning the capex spend that they're talking about the ad business is seasonal right and it's it's much more cyclical it's tied to the economy and so you may see at some point when if there's a slowdown you may see a bit more volatility there so to watch meta increase their spend and uh in terms of capex be on par with google and amazon and then for Zuckerberg to say next year will be significantly larger.

What is your view on that issue here of the people working at these companies and if they have the talent?

Yeah.

Right? They definitely want to see a little bit of excess and that's okay.

You look at the CPU spot versus contract markets and those, there's excess CPUs on AWS and Google and the margin structures there are really good.

I think one major question with the GPU market is you re- within the large hyperscalers, you do see great utilization and there's a question around the NeoClouds of what utilization are they seeing and what the unit economics are there.

But overall, I, I think they probably won't look to burn, uh, others.

Mm-hmm.

... or text-based...

Uh, we call them context databases.

Give an example, yeah.

Well, I think, um, you know, I mean, f- w- I've backed a few different companies here that, uh, we have an idea.

We validate that there's customer demand.

Com- founder gets off the ground, has one or two different customers, and then 18 months later is kind of in the same place.

And, and they're, um...

You wanna believe, right? We all, I think we all wanna believe we're all optimistic.

We're all investors in startups for that reason.

Yeah.

(laughs)

Because, uh, you know, the US venture capital asset class has grown from about eight to 250 billion during that time, and so there's a lot more capital.

Uh, and so we have to differentiate our capital somehow.

And the way that we differentiate is by depth, so we operate.

Uh, every Monday, we get together, we talk about different market maps and what, what are research projects.

And I'll give you an example.

We had this amazing intern last summer.

Um, where are you investing in blockchain, on-chain broadly?

Like mean or Mern or whatever.

Yeah, exactly.

I, I don't know.

I mean, I think, I mean, you have like Google building all the TPUs and you have Amazon working on the friendship chipsets.

Microsoft is working on some open the eyes.

We're working on some, um, We had this amazing intern over the summer who came in and produced a 10-page paper and he said, I don't think GPUs are the architecture of AI.

I think they'll be ASICs.

Do you think they'll invest in chips just for their Lama 4?

As we move towards a more exhaustive compute environment with inference booting up, advanced agentic ecosystems, where do you see the power dynamics between hyperscalers and these independent LLMs?

I don't know if it was a pun or not, but the word power is incredibly important.

So electricity is absolutely critical.

I was talking to a friend of mine who spends a lot of time and energy, and he was saying, I was asking him, compared to an electric car, how much does a GPU consume? This is about 100x.

So if you're building a data center, if you're building a power plant for a bunch of electric cars, typically around like five megawatts.

If you're building a power plant for significant data centers, five to 30 gigawatts.

And so the major constraint, one of the major constraints today is just energy.

I'm just really curious, like internally, like what is the operating system for that? What is the stack and what are the different programs and tools that you regularly use?

As we move towards a more exhaustive compute environment with inference booting up, advanced agentic ecosystems, where do you see the power dynamics between hyperscalers and these independent LLMs?

I don't know if it was a pun or not, but the word power is incredibly important.

So electricity is absolutely critical.

I was talking to a friend of mine who spends a lot of time and energy, and he was saying, I was asking him, compared to an electric car, how much does a GPU consume? This is about 100x.

So if you're building a data center, if you're building a power plant for a bunch of electric cars, typically around like five megawatts.

If you're building a power plant for significant data centers, five to 30 gigawatts.

And so the major constraint, one of the major constraints today is just energy.

I'm just really curious, like internally, like what is the operating system for that? What is the stack and what are the different programs and tools that you regularly use?