Give me a little peek behind the

curve there.

So you can buy, forget about again, like I like to use analogies.

I gave you an analogy using a single bedroom apartment.

I'm not going to use an apartment again, although I can.

by the way, let's say, imagine we're doing a single patty burger index, right? You can buy a burger from a burger stand off of the street of New York, or you can walk into a high-end steakhouse and order a burger.

Believe you me, like those two burgers are going to cost very different, right? They're both burgers that, you know, and what we try to do is we try to, as much as possible, find the marginal price for a unit of a compute or burger in this case that is sort of have the same relative the same feature i wouldn't want to compare the price of a three you know a three piece you know a three patty burger with a single patty burger right but once i make those adjustments there are some adjustments i cannot reasonably make because it's capturing a different type of premium from product bundling or product differentiation so in the case of hyperscalers indeed you are observing correctly they charge regularly, consistently, at least two to three times, sometimes more, compared to a typical, a new cloud.

So what are you measuring on a day-by-day basis that allows you to say what the proprietary cost is doing at such a fine grain of resolution?

Let's talk about the Silicon Data indices and what's the difference between, let's say, a stock index or a bond index and the Silicon Data index that you guys are providing.

So thank you very much for that intro.

Indeed, I come from a traditional index background, traditional quant finance background, joined Silicon Data about four months ago to try to build an index for this brand new asset class.

We're trying to People are calling it an asset class, we certainly don't believe it, and we're trying to make it so.

And there is a lot of similarities and knowledge that can be borrowed from traditional asset classes, as you've got a new asset class, but there's also, I think, new situations that need to be handled slightly differently.

So think about you made an analogy with an equity index or stock index or even, you know, I can bring up maybe bonds or commodities later.

I think the biggest difference is, of course, the fact that this is not a very, I think, liquid and efficient market just yet.

What... kind of buyer or seller would use this index and have very little basis risk versus the index and their actual core business.

So

so the.

I think the one thing that's somewhat reassuring about all this is we are seeing a pretty decent degree of rapid proliferation of adoption, right? I remember a couple of years ago, even just like a year and change ago, I was talking to my buddy at a top law firm, which I should remain...

unnamed, but like top four or five law firm.

And he was like, oh, like AI, I don't, you know, it's useless to me.

Like it's hallucinating.

And I didn't know, like today I asked him the same thing.

And how much do you think is just sort of a function of, I don't know, where we are from a macro standpoint as a country?

So I would love to hear, before we kinda get into the nitty-gritty of what Silicon Data's doing, like, for those who aren't aware, can you briefly explain futures contracts simply for listeners who don't, who don't trade?

Yeah.

Futures actually it sounds complicated, and indeed, I mean, most average investor, retail investors probably don't use, you know, financial derivatives futures, but it's actually one of the most ancient, you know, financial instruments in the world, right? Um, the whole idea is that, uh, in the ancient days, suppose you are planting crops, right? And, uh, you-- let's say your cost of, you know, planting crops is a dollar, and, uh, currently cops, you know, crops sell for $2, you make a profit of a dollar.

But, uh, you worry that by the time you harvest, let's say, six months or a year from now, the, the price of, uh, corn, you know, may actually plummet to be, uh, uh, below your cost, and you may end up with a loss.

So if there exists a futures market which allows you to deliver, right, that bushel, you know, or, or the amount of, uh, crop at that point of harvest at the price that's currently, you know, trading at, right, say above your cost, you might be inclined to lock it in, right? Now, uh, why would that be such a contract? Why would someone offer you such an arrangement? Someone else must have a different point of view, right? And someone may actually believe that, "Oh, actually based on my expectation of weather and, and whatnot, by, by the time, you know, harvest comes around, uh, uh, corn actually will be more than $2.

You know, I can actually make even more by sell- you know, selling you to... buying it from you for two two and selling it on to whatever I think it's gonna be," right? So market is always full of people with different opinions.

Some people are bearish, some people are bullish, some people are risk averse.

... who might wanna hedge downside versus like AI labs and inference platforms who might wanna hedge upside and, and then just like anyone in between, retail, futures traders, et cetera?

So, so to give our audience a sense of, uh, uh, okay, so Steve described what they do, but how do I know it's any good? So I'm sure you, uh, there might be things you're not comfortable sharing, but give us a sense of, at a high level at least, what, what are the data sources?

So, uh, our data sources comprise a mixture, right, of, uh, publicly available list prices as well as, uh, you know, proprietary, like transaction price that we obtain from our, uh, data partners, cloud providers, uh, as well as our own sister exchange.

So Silicon Data has a sister company called Compute Exchange, right? So the, what the company does is in its name, right? The Compute Exchange, uh, basically allows people to, uh, transact compute, you know, if you have spare compute, you're looking for compute.

And, uh, we take the totality of the, all the data we, we observe and, uh, we, you know, sort of normalize it.

So just like if I give an analogy, if I might, uh, like suppose you wanted to create a single bedroom unfurnished apartment, you know, rental index in New York, right? How would you do it? You go about collecting, you know, rental agreements from across, you know, the city and, you know, and, uh, they come in different shape and form, right? Some are furnished, some aren't furnished, some rent for three, six, nine months, 12 months.

You know, they have, you know, all the idiosyncratic reasons why they're a little bit different, location and so on and so forth.

And you take all these attributes, and you basically then normalize them so that they are apples to apples comparable, and you consolidate it into a single index.

Yeah

[chuckles]

Yeah.

Indeed.

Yeah, indeed.

Uh, this, this, uh, index was sitting on the shelves, uh, you know, when I joined the company, and, uh, you know, for, like quite a couple of months, and then, you know, uh, and I think it was not getting a lot of attention.

I think I just noticed that people were sort of whoever noticed it was actually not interpreted even correctly too, or thought it was either a price index or total volume index.

I think the name of the index on terminal display has also got a little bit unfortunate.

... and yeah.

What does that mean? Is that token prices are coming down or that demand for tokens is coming down?

Uh, actually, it's kind of neither.

It's a little subtle, and that's why we wrote a little social media post and, uh, which led to this index, you know, getting picked up by, uh, I think Citadel Securities wrote a little note on it, and every other news agency started picking up on it, and now has become almost k- you know, unintentionally, un-unintentionally by us, this kind of AI bearish index.

We-- I actually love this opportunity to clarify a little bit.

Uh, what it is, is actually a little bit like in the macroeconomics where you have inflation measured by expenditure-weighted price index, right? The PCE way of measuring inflation, right? In other words, I'm going to allow consumers to rationally substitute between alternatives they can consume and, uh, uh, and instead of fixing a basket and say 20% this, 20% that, and whatever, I'm gonna allow them to choose however much they want to consume and see what the net price expenditure is, and that captures the true inflation.

So in this case, we're going to allow for the fact that people can choose different AI models, especially when in, in a agentic kind of API call, uh, uh, context.

Uh, now that you have a lot of these platforms that have routing, you know, uh, uh, mechanism that you can actually choose.

How do you think something like DeepSeek fits into the equation? Uh, I have a lot of friends increasingly telling me they're using DeepSeek because it's cheap, it's fast, and if you're using it for everyday stuff, it's pretty much indistinguishable from some of the US models.

[laughs]

Yes, in the bubble-- It was always going to be a bubble, right? Back in twenty nineteen three, when I first got quite bullish on AI investments when, was when everyone started talking about this concept of baby bubble, right? I think Bank of America or one of these, like, sell side shops had this chart where they compared all historical bubbles and they said, "Oh, there's a baby bubble brewing in AI." And, uh, they pick a few, I think that not even necessarily very good.

It was under the stocks of, uh, AI companies as attributed like the Mag Seven or whatever to say, "Oh, this is actually the beginning of, um, uh, of a AI baby bubble." In hindsight, that was actually just the market rebounding from the lows of, uh, late twenty twenty-two.

But regardless, there was a, a, a, a, you know, the AI, you know, with the introduction of ChatGPT at the end of twenty twenty-two was becoming to spread throughout the stock market, at least sort of the ones that are related via AI value-- picks and stock market value chain.

Now, the question is not what-- The de- I think the interesting debate is not whether or not AI w-was a bubble or will become a bubble.

It manifestly was, uh, and had become one, right? The question is, uh, how long is the bubble, right? How big could the bubble get, and at what stage are we? Uh, and I think, uh, that is the part that was very underestimated by people, uh, by especially by macro people.

Once they heard the word bubble, they started dismissing its, you know, significance.

So yeah, I'm curious, how do you think the Fed should even be thinking about this? Yeah.