Unchained - Bits + Bips: Is AI CapEx a Bubble? And Is Inflation Already Dead?

Episode Date: February 18, 2026

The Mag 7 have committed over $700 billion to AI infrastructure, but the companies building the models may never capture the value. Thank you to our sponsors: Adaptive Security Fuse: The Energy Network The BLS just quietly revised away 862,000 jobs, and real-time inflation trackers now peg price growth below 1%, less than half of what official figures report.  If the Fed is steering monetary policy with stale data, investors need to ask what else the models are getting wrong.  At the same time, the Mag 7 have committed more than $700 billion to AI infrastructure, with Anthropic alone projecting $1 trillion in revenue within five years. Is that conviction or the early stages of a debt cycle nobody is pricing?  And then there is the institutional side of crypto: BlackRock's BUIDL fund just landed on Uniswap with $2.4 billion in assets, Apollo acquired $90 million in Morpho tokens, and AI agents are already settling micropayments in stablecoins.  Austin Campbell, Ram Ahluwalia, and Christopher Perkins sit down with Truflation’s CEO Stefan Rust to ask whether the numbers we trust are telling us the truth. Hosts: ⁠Ram Ahluwalia⁠, CFA, CEO and Founder of Lumida ⁠Austin Campbell⁠, NYU Stern professor and founder and managing partner of Zero Knowledge Consulting ⁠Christopher Perkins⁠, Managing Partner and President of CoinFund Guest: Stefan Rust, Founder and CEO of Truflation Links: Unchained:  BlackRock Just Chose Uniswap. The Market Didn’t Care. Here’s Why. Apollo Moves Into DeFi Lending With Morpho Token Deal UNI Spikes on BlackRock DeFi Move, Then Gives It All Back Macro: NBC: U.S. had almost no job growth in 2025 PBS: Inflation measure falls to nearly five-year low as gas prices fall and housing costs cool Crowdfund Insider: Secretary Of The Treasury Scott Bessent Calls Out Truflation's Inflation Numbers At Senate Banking Hearing AI CapEx: Amazon, Google And Others Are Pouring $700 Billion Into AI CapEx, Top Analyst Explains Why This Makes It 'Hard' To Bet Against Nvidia CIO: Data center capex to hit $1.7 trillion by 2030 due to AI boom Reuters: OpenClaw founder Steinberger joins OpenAI, open-source bot becomes foundation Learn more about your ad choices. Visit megaphone.fm/adchoices

There are echoes of 2008 when you had debt that wasn't worth par that was being issued at par, and you had equity securities that were highly inflated. So I don't think it's great for Mac 7. We're seeing at the higher end, elder generation, not looking for a new job anymore, just sticking it out. We're seeing on the younger generation, people graduating, struggling to actually get a first job. I think the first thing is you're seeing an absolute manifestation that, this error of regulatory risk is over. Under the Gensler error, there's no way BlackRock or Apollo would have come into the space.

Hey, everybody. Welcome to bits and bips, where we explore how crypto and macro collide one basis point at a time. I'm your host, Austin Campbell, high scholar of zero knowledge group, here with my co-hosts, as usual, Ram Alawalia, maister of wealth, leader of Wibna, and Chris Perkins. Chris, I think you still haven't named your thing yet. So I'm going to continue to poke you on that until it's done. Soon. Soon.

All right. Fair enough. We'll see. And today we're joined by a guest I'm personally very excited about. Stefan Rust, CEO of Truflation. So we're here to discuss the latest of the worlds of crypto and macro. But just remember, nothing we say here is investment advice. Check unchained crypto.com bids and bips for more disclosures.

And before we begin, a quick word from one of our sponsors. This episode is brought to you by Adaptive Security, the first cybersecurity company backed by OpenAI. As AI makes deep fakes and synthetic identities easier than ever, Adaptive helps companies test and strengthen their defenses. Learn more at Adaptivesecurity.com. The energy network is an intelligent, decentralized grid that coordinates smart devices to balance supply and demand. Energy Dollar is the native token of the network from one of Europe's fastest growing energy startups. Follow at Fuse Energy on X to find out more. Quick note before we get into today's episode. Bits and Bits Now has its dedicated feeds. We're spinning up from the unchained feed and moving

to a new podcast and YouTube channel. So if you want to keep up with our weekly live streams and macro meets crypto breakdowns, make sure to subscribe to Bits and Bips directly. We won't publish there March, but subscribe today so you can be ready for launch. Be sure to subscribe to the new feeds at Unchained Crypto.com slash bits and bibs. All right, before we begin on our topics today, we have somebody who may have a unique level of insight on one of them. So, Stefan, I don't know if all of our viewers are going to be familiar with Truflation yet. So could you take a brief moment and explain to us what that is? Yeah, thanks. Four years ago, when Trufflation was only meant to be transitory, I got together with two other co-founders and we looked under the hoods to see how

inflation was actually calculated. We saw it was done manually. It was done on a survey basis, a thousand households tracking the whole cost of living across the US. And they were only tracking some 60,000 items. And in a world of technology, internet, we, blockchain, Web3, we felt that it could do with a hell of a lot more transparency, more immutability, accessibility, and automation in that whole process. So we started aggregating today, maybe nearly 30 million items that we track every day across some hundred different data providers that we pull together around the world to then calculate what our interpretation is CPI.

It's not the same as the BLS. We do use our data to then try and replicate the BLS number, and we got that down to an accuracy of about 99.93%. And the main reason why we do that is because most of the Wall Street or the city or sort of financial institutions really like our forecasting of where we believe the BLS read is going to go, or as we call it, the Bureau of Lacking Statistics. we try to give that a real-time interpretation of it. All right.

Thank you. Inflation is a complicated topic in general, as somebody who's had some unfortunate exposure to it in the past. So let's talk about that exact topic right now. We'll call this segment, soft jobs, softer inflation question mark. And the overall setup here is pretty, like, complicated, which is one, 2025 payrolls were revised down to 181,000 jobs total from 584K initially reported. BLS subtracted 862,000 jobs and benchmark revisions from March 2024 to 2025.

This makes 2025 the weakest hiring year since 2020, and the weakest, we'll call it, non-recession year since the early 2000s. January, however, added 130,000 jobs versus 55, 5,000 expected. Unemployment is currently as the way we calculated at 4.3%. That's labor. Inflation is drifting towards the target. So headbine CPI fell to 2.4% year over year in January, down from 2.7. Core CPI was at 2.5, lowest since March 2021. Gas prices are down about 3% month over month and 7.5% year over year. used car prices are down slightly month over month. Tenure yields fell post-release as markets priced greater odds of a rate cut.

And where we've ended up here is consumer prices are still about 25% higher than five years ago. So affordability remains a major talking point. But wage growth is moderating as higher exclows. And the inflation narrative has become somewhat complicated. Treasury Secretary's got best referenced Truflation during his Senate testimony. And policy may already be restrictive relative to real-time inflation, but where is real-time inflation? So, Stefan, I want to start with you. What are you seeing currently in the inflation numbers? So, I mean, right off the bat, we're below 1%. According to Truflation and aggregating real-time

data, we're below 1% in aggregate. We track 12 different categories. And in each of the those categories, we vary significantly, depending on the movement. What we've seen, a slight uptick, you know, sort of in sort of from 0.68 to 0.9, we've really, I mean, not that it makes a difference to every single household, but from a mathematical standpoint and from a derivatives perspective and from a hedging perspective, these numbers have a significant impact. And we've seen this uptick, largely due to transportation costs spiking a little bit and picking up. More people traveling, the cost of airfares, public transportation have all jumped a little bit, and people are witnessing that, and that's also the echo that we're getting on the street.

Not only our numbers are showing that, but you're seeing that in real-time sentiment feeds on all social media. Does that inflation from labor and housing and shelter costs, or is it primarily goods and merchandise? It's also including housing. So we include housing costs in there. We have our own sort of interpretation of shelter where we extract utilities out of the actual shelter costs. The main reason why volatility in utilities is greater than the movement of your rent or your mortgage payments. But yeah.

Are labor costs in this also? In terms of social welfare costs associated with labor or in terms of employment costs, just wage inflation. Wage inflation is not. It's a separate item that we track separately. Got it, got it. So I think that's a key distinction to make because overall, capitalism delivers deflation. Prices of things drop over time. Anyone that's in the technology industry, seen this. We've seen waves and waves of that. But the big inflation costs have been around the cost of labor and, you know, the cost of housing. Yeah, I mean, we see a delta go, yeah. Yeah, no, I mean, we're definitely seeing a change in the labor market.

I mean, we're all aware of, you know, productivity gains. We're seeing at the higher end, the older generation, not looking for a new job anymore, just sticking it out. We're seeing on the younger generation, people graduating, struggling to actually get a first job. Universities haven't done a great job in preparing them for the real world. And that's been a challenge. So on those two front ends, we've seen the military come in and hire a lot of new, some 200,000 new recruits hired in the military. So those are things that we've seen. We've launched an employment index largely because we see the BLS is really falling behind.

and during the shutdown, they didn't have any data. They weren't tracking the data. They're tracking fewer numbers. Their budgets have been cut. So ultimately, the data that the Fed is using is getting worse and worse. And so we just felt that those were the two items that we really wanted to track and get good at. How do you get your numbers wrong by 400,000, Stefan? I mean, they had 400,000 ghost jobs last year.

Like, how does that happen? So the way we understand it is they look at the number of businesses created in a given year, and they take an average employee per business create it. And then they multiply that, and that's how they extrapolate and put a deviation into there. And that's how they come up with this number. And then in the end, they get the actual numbers, and then they have to revise it downwards. So essentially, if we're looking at that sort of thing, We're seeing that small businesses or businesses recently created or hire it, call it less than they would have expected per the bottle.

And you're seeing less businesses created under the previous administrations. It was very, and still today, it's still very heavy, regulatory, regulatory heavy in terms of launching a new business, compliance heavy. And so a lot of costs are needed up front to launch a new business. A lot of permits are needed. So how do we sort of reduce those friction? And I think we've started to feel it earlier in sort of last year now, where the regulatory burden in terms of going and reporting all the time for publicly listed companies has dropped significantly.

And we heard, anyway, a lot of the big public listed companies mentioning that they don't need to go to Washington all the time and justify what they're doing and spend a lot of legal costs constantly knocking on SEC's door or answering to SEC letters. Yeah, I mean, the tariffs were a big punishment to small business that couldn't absorb or shift or avoid cost burden quickly as compared to Walmart and Costco, which didn't see much inflation pickup and their cost of goods sold. Small businesses were in pain. How do you see just the outlook, though, right?

If I look at the price of oil, it was at 56, now it's 65. That's up 20%. Gasoline futures are going up. Markets are rallying around commodities, like all of the commodities. Energy stocks are going up. So four looking market indicators are saying that inflation is coming. Is that accurate? Or will that then flow into your data?

Because your data is looking at snapshot in real time of what just happened. Yeah, but what we very quickly realized was our customers or users of our data didn't really just want the actual real-time data sets. They wanted to be able to, where's it going? I mean, we had customers asking us 10 years out. What is inflation in 10 years' time? I mean, I don't know how we calculate that, but we get them a forecast. And ultimately, it's a directional thought piece more than anything else.

But we definitely see commodity prices going up, all the raw materials. for a lot of the, especially rare earth's energy, battery, silicone, a lot of those are moving upwards and drastically. Gold, silver, have been, you know, and then to your point, energy. And that's going to have an impact in everything, but we just believe the volume and the scale are going to bring down unit costs significantly. And with robotics and AI coming in, the productivity gains and the volume that's going to be created and have a impact on unit costs at the end, which is what the consumer bears is not going to be significant enough to really have a big impact on inflation. We actually see the inflation going downwards even further.

Tariffs, everybody thought tariffs was going to have a big impact. We didn't actually see much of an impact, and we actually saw more of the export side of the industry taking on a bigger load. So it was people trying to sell into the U.S. versus the importers carrying a lot of the burden. Overall, worldwide trade is at its highest level. The BIS, which is the Bank of International Settlement, the central bank of central bankers. They actually did a report and they highlighted that the fabric of global trade has never been stronger than it is right now, which is contrary to what we're reading and what we're hearing all the time. How to interpret that, you know, I think it's time we'll tell.

But right now, so far, we really haven't seen much of an impact of these tariffs overall. All right. Let me pick up you then. On something you said that I think will be interesting, which is if we think about what consumers actually bear in the form of inflation, probably the one that is produced the most like Adjita recently is housing. So I would like to ask you, especially given in the United States, some of the dispersion across individual housing markets. Because if you look at like, call it San Francisco versus like an Austin, Texas versus Chicago or New York City versus Miami, you're getting very different stories. So how do you think about inflation in the housing market? How are you computing that?

What components are in there? Because another question that may be a step beyond that is also commercial real estate. Yeah. So housing prices haven't moved much. on the face of it. Underlying, though, we're seeing a lot of rent subsidies kick in. So you see one to two, three months free rent come in, but that's not being tracked. And so we're trying to sort of go a bit deeper into the pricing and housing.

How is that being priced? We did a big report with Penn State University. I think it was about a year and a half ago, nearly two years ago now. And there we sort of tried to really, how could we aggregate better real-time information and extract the utility elements out of there? Now, how do we make sure that we can capture some of these subsidies and what does that look at? How big are they? Are they really that big or not? But people are trying to hold rent prices largely due to mortgages and valuations of the real estate.

Austin's been booming. Areas where, again, the regulatory framework where it's easier to build, you can be. bring supply onto the market have a significantly greater impact than more restrictive environments, New York, San Francisco, where the licensing goes through a lot more challenging times in the process of trying to build some new property. So what a shot out stuff. In the last week and a half or so, we had, we saw David Sacks mentioned trufflation as as an important reference.

You saw Secretary of the Treasury Besson talk about you guys. Can you talk about some of your growth in users and adoption? Because it feels like you're starting to develop the social consensus, a concept that we always focus on in the crypto space at least. And you're seizing, I mean, this is a white-hot company, a white-hot initiative. And it sounds like you're starting to really capture that social consensus and people are pivoting more towards triflation than BLS.

I mean, it's apples to oranges, obviously. But tell us about your user adoption, because it sounds like it's been pretty profound. Yeah, we hit a big spike when Truflation at 12% and the BLS were at 8% at the peak. We definitely went through the ringer. We were battle tested. We had a lot of critics, sort of, you know, try to battle test our calculations and methodology. But it was actually just a more accurate reflection of where we were. Since then, we've been.

in, yeah, bubbling around and sort of holding the 2%. We broke down below the 2%. We're now below 1%. And we've gotten a lot more traction. The growth over the last two weeks has been, yeah, I mean, the systems have all stayed up. So that's a nice sign, exactly. But yeah, we've really become,

we've got a lot more interest. Demand has picked up, both from the independent financial analysts, macro-analysts, macro-traders that have always been using Truflation. So they've already seen a 45-day lead-time indicator from Truflation. So you get, and we've done regression analyses around these associated, and you can download it from a website. We've open-sourced our calculations and samples, even the code,

so you can go and do your own simulations with the data. But a lot of the macro-traders and macro-analysts already are using trufflation. And so now it's just swinging over into more institutional hedge funds, big private equity traders that are starting to use this data more and more. And we're also beginning to see the interpretation where truflation is acting as a signal to drive investment decisions. So they're using truflation against interest rates or yield to identify the real yield based against truflation, and what is that impact in risk assets, in the dollar? What is that impact having on the market? And where is it going to go, especially if we have these pivot changes in the market? And so that's what we're seeing happening, which is really exciting. How do we now

financialize the data, as we call it, right? With AI coming in, we're already connected into all the AI tools. We have MCP for all the data. And by the way, all of this data resides on a brand new SQL built blockchain. So it's using standard database language, standard database technology, so that anybody can on ramp really quickly with new sets of data. You have a block explorer where you can trace

all the sources of data. So you can go and pick datasets, build your own inflation calculation if you wish to do so. And then ultimately, the output, equally easy to integrate into an application, into your AI, et cetera. And so we've tried to really make it decentralized, number one,

and number two is easy and simple, purpose built for data. Ladies and gentlemen, it's a crypto and macro. How more crypto and macro can you get? Sorry, Ron. Great, great combination. No, that was great. I think remember when true inflation first came out, the perception would be that we would see the true inflation numbers,

then they'd be way ahead and above what the CPI, was reporting. And it's actually well below, which is a narrative violation for a lot of people, especially people in digital assets. And most people that go out to a restaurant, they're seeing higher prices. When they try to get someone to mow their lawn, it's higher prices. You get a babysitter, it's higher prices. You get a nanny. It's higher prices. Now, a lot of that is because of services, right? The number an employer in the vast majority of states, the United States, is services and it's health care. We have a services-based economy rather than a goods and manufacturing-based economy, which I think is a good thing. It's more stability. It's more mature economy. So how relevant is the goods-based

inflation data in a world where it's the services costs that are going higher and impacting people's quality of life. So a couple of things there. So we break out goods and services as well. So you can hit a button and you get the good side and you get the services side and you can prepare the two as well. So we actually do break it out just to make it more visible and try to just slice and dice the data in any way we can. I think what a lot of people, the misconception on inflation is we only track inflation year on year. So what was it the same time one year ago? If we look at it five years back, I think, you know, we were saying earlier that it's 27% or 28% inflation over the last five years. So that's a significant increase. So everybody's costs has gone up over the last five

years. And people begin to feel that. You go to the grocery store 10 times a month. Every time you go to the grocery store, you're going to feel that pain in terms of any food item in there that's going to be a bit more expensive. We believe the consumer can't have. handle any more price rises. So it's really hard to come up with price rise. What's happening much more is, you know, the grocery stores are building their own brands and taking out all of the marketing costs associated. You know, you can see that with Costco with their Kirkland brand. You know, Sam's clubs got, you know, got their own brand that they're building out to try and bring down costs and leverage their distribution power to bring down Unicoths again. So

that you have a choice between, I know, one brand and another brand, right? And the home brand, domestic brand. What do you see around health care and just care in general? That's an increasing part of the GDP. It's a one-way direction. It's not benefiting from technology and the magic of capitalism. Do you have data that you can share there? We are just putting something together.

We don't have it right now. Healthcare is notoriously difficult to get data on. But we've sort of set out, I think it was in December, where we really started going deep into the healthcare and the education categories. Healthcare, yeah, I mean, it's highly tied to labor costs and the demand of skilled labor in specific categories as it gets more and more nuanced. And the wellness category is taking a lot of resources out of traditional health care and Medicare. And so that's sort of what we're seeing. And we'll share more on that going forward. So I want to rewind to something that you had said earlier about people

piling in to use more and more of the Truflation products. Because a theme that myself, Rahm and Chris have talked about before is trust in institutions overall and how that art is evolving over time. Do you see people primarily using Truflation in addition to the BLS as an augment or people deprecating the BLS data or just not trusting it using Truflation essentially as call it either the replacement or primary source of data. Because to me, those are two very different pathways. Two sides, I mean, two different user segments.

The macro traders that have been using us really early on just use Truflation. They don't even look at the BLS numbers anymore. The general traders today, I mean, there's still some $5 trillion tied to. to the BLS inflation number. If you don't look at the whole bond market, if you just look at tips, for example. So there is a big value tied to the Bureau of Lacking Statistics number. And so, you know, that's always going to be important.

And that's the stickiness that they have. Trueflation, unfortunately, not yet, doesn't have that stickiness. But we're working to build that out. But yeah, more and more people, especially younger generations, and new analysts coming into the market, they're looking for an edge, they're looking for something different, and they're beginning to use and interpret

with the truflation data. Do I buy Bitcoin or not? When do I buy Bitcoin? Truflation is acting as a source of truth. When do I buy dollar? When do I go long dollar? When do I short? Truflation is acting as that source of truth.

There are things like that, which we find, you know, risk, you know, yeah, anyway, yeah. So all of these, we have lots of, I mean, they're all coming to shine where we've seen all these new analysts come up with these new ideas, new reports, new interpretations, and building out regimes. Sorry, building out strategies using the truflation regime, which is what we are trying to put forward a lot more. Stephen, can you unpack a little bit more how you're using blockchain technology and your solution? I mean, we've seen how many projects where somebody comes up with a big idea. They launch a token.

They're like, oh, it'll be governed in the future. Everyone buy it. Then it blows up later. But what you've done is you've created something of real material utility. And now you're leveraging blockchain rails. Can you walk us through how you're using those rails and then how you intend to build off of that going forward?

So we very early on partnered with Chainlink. So Chainlink was sort of the layer of truth where we published all of our indexes to the Ethereum chain with Chainlink or through Chainlink, through ChainLink node. In parallel, we saw ahead of the time, you know, we saw AI's coming, you know, four years ago when we launched. We also saw that the purpose-built ledger technology wasn't suited for high-end and volumes of data that we needed to aggregate. We partnered with a company called Quill out of Austin, and we built and were heavily focused on building out of SQL. native chain. So it was a whole new chain using SQL. We needed to have a consensus algorithm to

verify all the data. We needed an explorer so people could find and track and trace the data. We needed it to be able to bridge ERC 20 onto that chain. So we had to build that out. And we needed to enable SQL smart contract development on chain itself. So if you wanted to do calculation of new types of indexes, that that would be transparent to certain participants. that wanted to have a look at that. Those are all things that we set out to do. We've been investing the last four years to do that. We brought Quill in-house.

And we have an experienced blockchain team, meets data team, that are working side by side on really trying to scale this. It's on test net right now. And we want to expose this to more third parties. And we believe that the prediction market space is going to be extremely interesting, especially related to finance and economics, right? How do we build prediction markets for the economy and for finance? And at what point in time does that then have a bigger influence over where interest rates should go?

What does the market really believe based on these binary options that are put in place? All right. So one more on this topic. After this, I'm about to go out and get dinner. and one of the spicier components of CPI recently over the past year or two has been food inflation. So I'm going to say that's one where knowing some people still on the street, there is a food fight about what is going on in that series. I would be remiss not to ask you, what are your views on what's been going on in food and if you have any of where it's going. So angst was up at $8. It's hit deflation and it's now at $3.

right so so eggs have come down so we're experiencing your breakfast should be getting cheaper although coffee prices have been going up because there hasn't been enough range um but yeah so i mean food uh obviously hypersensitive we all eat food a lot we go to the grocery stores buy and shop a lot of food um but food has been significantly coming down so we've seen that trend and yeah just don't see much movement in the ability to start pricing food high higher. At the high end, you can start charging maybe a bit more. I think a lot of wealthy individuals or ultra high net worth individuals, they can forward better food and will always pay higher for quality food or what could be considered quality food. But on the general bracket, I just don't

think food pricing can go up much more. So dinner should stay pretty consistent unless there becomes more tax or something else gets layered onto it. All right. Well, on the topic of paying for things, we're going to be back shortly to discuss even more people buying stuff. But until then, here's a quick word from one of our sponsors. We'll be right back. This episode is brought to you by Adaptive Security, the first cybersecurity company backed

by OpenAI. As AI becomes more capable, attackers no longer need to break into your systems. They just need a convincing imitation of someone you trust. That could mean a deep faked voice on a call, a synthetic coworker on Zoom, or phishing emails written by AI that are nearly impossible to distinguish from the real thing. Adaptive's platform is designed for this new reality. It runs deep fake, vishing, and AI-generated fishing simulations so your team can see exactly how these attacks work and practice responding before it happens for real. Their AI content creator also turns new threats or compliance updates into interactive multilingual training within minutes. You can learn more at adaptive security.com.

The world is about to see one of the largest infrastructure shifts of the century. New technologies are using more energy than ever before, but our legacy grids can't supply the demand. And we are barreling towards a global bottleneck. So Fuse is rebuilding it. The energy network is an intelligent, decentralized grid that coordinates smart devices to balance supply and demand. The network harmonizes existing infrastructure, increases grid capacity, and unlocks low-cost, clean energy.

Energy dollar is the native token of the network. The more electricity the world needs? The higher the demand for the energy network, the value of energy dollars may fluctuate. from one of Europe's fastest growing energy startups. Follow at Fuse Energy on X to find out more. All right, everybody. Welcome back.

And now we're going to talk about a topic near and dear to Ram's heart, which is AI CapEx. So there has been a lot of spending on CapEx and AI. It is huge. It is front-loaded. And the Mag 7 AI Cap-X commitments currently are being framed at roughly $700 billion. Amazon guided to about $200 billion in 2026,

tied to AI and data set are buildouts. Meta is somewhere in the 125 billion plus or minus 10 for 2026 CAPEX. And the returns on this are currently uncertain and somewhat laggy, which is to say hyperscalers are spending ahead of clearly disclosed unit economics for AI currently. and the build is causing some second order effects in power, in cooling, in networking, in storage. If you want a good example of a market where there is very significant inflation right now,

go look at RAM prices as a result of what's going on with AI. Even so, it's hard to fade some of these companies. FutureM CEO Daniel Newman is arguing Nvidia could capture 40 to 50% of the Mag 7's AI CAPEX, meaning it's monetizing to boom, even if the customers don't find the ROI or it takes time. And Open AI is hiring Peter Steinberger of AI agent fame with OpenClaw. Even so, there is the question of where the value capture is and is this a bubble? So for this one, Rom, I know this has been something you've been thinking about deeply. Can I start with you and just ask you, what is your view on this and who in this space is both,

most vulnerable and most durable. Well, first off, it was great framing. By the way, briefly on Peter Steinberger, we've talked before about how you're, excuse me, talking before about how you're going to see a unicorn company of one. That's the first guy. We don't know the acquisition price.

They acquired open source technology, whatever that even means. It's sold in the public. They really bought him. And I wouldn't be surprised if it was over a billion dollars. So it's happened. There's no. more important question to risk assets than

capital spending AI. That's why it matters. It matters for everything. It's the primary dominant theme in markets is the AI story. It goes back to the inflation dynamic. We just talked about a moment ago. So, you know, what we saw at the end of October is Sam Malman had this awful interview with a hedgeman manager, Brad Gersner, where he said, look, I got a trillion dollars in committed spend obligations. Where we come

from, we call that debt. but they only have $23 billion in revenue to meet those obligations. So they're betting on the future. They're betting on demand coming in. And what we saw this weekend is Darya or the CEO of Anthropic come out and say that they have a projection

of a trillion dollars in revenue in five years. By the way, very few countries get to a trillion dollars in GDP. But as U.S., there's China, I'm waiting for the third. Okay? So they have a spreadsheet saying,

we're at 10 billion in revenue now, we're going to be a trillion in five years. and they're making cap-back spending plans and signing into contracts with companies like Oracle, Microsoft, Amazon, which are all below their 20-day moving average. And now the market's saying, gee, are those numbers real?

My view is that those numbers aren't real. They're way ahead of their skis and that you'll probably see those contracts, those revenue performance obligations not be met. I don't think that debt is worth par value. I certainly think the equities below the face value also. So I think that's the number one thing to track in markets and see how this unfolds. Chris, I'll jump in.

Yeah, so the first thing is these numbers are humongous, right? $200 billion, that's larger than the market cap of Disney, right? That's just about the size of the Ethereum market cap. These are big, big numbers. I think I really want to ask Stefan, but it feels like compute. is going to be probably arguably the most, one of the most important commodities going forward. You're going to see it getting traded more and more. And I'm wondering if you're starting to look at it, Stefan, as one of the inputs to your models.

And how is that going to rise in prominence? I mean, I imagine the price of compute is going to really inform some of these inflationary models going forward. I love your take on that. And then I want to pivot back to the agents because I have a lot to say about that. But Stefan. I mean, compute, yeah. I mean, it's all down to compute. right? I mean, what we look at is

computes the new utility, right? So you've got energy, we need energy to feed the compute. Those two are going to go hand in hand. I mean, Ethereum is a good benchmark, right? You have gas fees that ultimately is what you pay for the compute, and you have multiple participants actually hosting the compute resources. And I think that's the interesting side of things and decentralization. is going to distribute a lot of the costs and the maintenance

associated with the compute resources that we need. And ultimately, the more participants you have, the bigger the distribution in an ecosystem hosting that environment, not necessarily in the data center-wise, but actually ultimately the actual physical maintenance of the compute resources, you know, that's going to bring down the costs again and allow us to do more compute that way. And, you know, I think decentralization is the way to go.

You know, there are a number of benchmarks. There's going to be a whole set of new indexes. They're going to be tracking. You're going to be able to trade and you can already trade future hash power, future gas prices. And you can track the rewards that these, you know, with some of these indexes. So there's going to be a whole new economy. As the Ethereum economy grows, more and more assets move onto that. economy, more and more gets traded across that economy. That becomes inflationary, deflationary,

and you want to track the costs associated with transactions on that platform or another platform. Ethereum grew out of the fact that Bitcoin was becoming so expensive to do a transaction. Solana grew because Ethereum became so expensive in order to do transactions. So there are competitive elements out there for each of these economies. And if you look at Ethereum's economy, Solana as an economy and even Bitcoin as an asset class, then you've also got different types of digital forms of economy that are going to grow that need tracking and need inflation monitoring and all the other macro data. Well, I want to jump on this because I think it relates to the agent point that we're going

to get to with Chris here as well, which is to say the following. One of the questions that I've had for a while about many of these systems, and I think AI is a great example of that, is the potential misalignment between where the money is being spent and where the value capture is ultimately going to occur. Because right now we have model companies, so take the open AIs of the world, incinerating huge amounts of money currently from a balance sheet perspective. And it is not clear to me that they are going to be the ultimate beneficiaries of the AI boom. Like models may turn out to be relatively easily swappable, replaceable and might not have a huge amount of sticking power, which I want to be clear is not a

refutation of the value of AI. It is a question of does the value of AI accumulate to the model companies? A little bit like the fat protocol thesis problem with crypto. I see even saying that turned out the lights for ROM. See power problems right now. But no, the question that I want to leg into there for the group, and Chris, I'll throw it over to you since you wanted to talk about agents is where is the value going to accrue? Because it's pretty clear that Peter did pretty well here with OpenClaw, but like how is this going to be shaped? And what are people going to use and where is the pricing power? Yeah, look, I think we're still, all this is we're focusing on the infrastructure stage, right? And picks and shovels are probably where to focus on is my guess. But of course,

like who really stands to benefits are those super apps? And, you know, Rom talked about the unicorn of one here. I mean, you can't, we've been talking about agents for a long time. I remember back in my little clip on my Microsoft thing back in the day, annoying as hell. But now they're here and they're providing real utility and they've captured the zeit guys. Like, you know, everyone was talking about, you know, Claudebot, open claw, molt bot, whatever you want to call it. I guess he got in a little bit of trouble with Claude, so he decided to do the deal with Sam over to OpenAI. But these agents are real. And I guess one of the things, as I'm putting my crypto hat on, there's a couple things I'm thinking about. Number one, you know, we talk about retail versus institutional and like

that story and that saga. Retail's got burned. They got hit by 1010, blah, blah. The institutions are marching forward. You know, we're waiting on this gap. You know, no one's talking about these agents that are muscling into this space as well. Are you guys tracking X402, which is part of the the payment protocol now that Coinbase launched. You know, this is tracking, it's already tracking like 600 million in payments this year, like from zero. Now, in the payments world, Austin, you know that that's zero. It's like to drop in the bucket.

But it's not nothing anymore because you're starting to figure this out. You got Ethereum's EIP 8004 that's helping agents, you know, provide credentials to each other. So this infrastructure is getting built, you know, and it's moving quicker than I thought it would move, which is so my point is is that you have retail you get institutional retail's out of crypto institution moving forward the bots are coming fast and I think that's going to help fuel

and accelerate into whatever weird funky market we're now in into more into bull market finally on the agents you know I find it so interesting that this is going to OpenAI and Sam Altman who also let's not forget is the founder

of WorldCoyne which has been kind of quiet but he predicted this agent economy. He predicted that we would need proof of humanity. Haven't really seen him out there yet with World. But he founded it, and I'm excited to see it all come together, to be honest with you. Thoughts? What do you think, I mean, agents and LLMs, I mean, what currency do you think they're going to want to take?

Are they going to take crypto assets? Are they going to take digital stable coins? What currency are you going to pay for? these agents. And when agents talk to agents, what currencies are they going to select to settle with each other? So I'll say looking out into the market right now, the bet pretty clearly appears to be U.S. dollar stable coins is the dominant method of payment. And I think that comes down to two different factors. One, Stefonda, what you were saying earlier, Bitcoin is not well designed to be used for huge amounts of micropayments by agents. Like, that's just not a great framework

for that because of the design of the chain. And look, whether you think that's good or bad, it is what it is. And it may be the case that Bitcoin is way more useful as a store of value than a medium of exchange because of that. So be it. But two, you know, if you look at what's going on with Coinbase right now, as Chris rightly pointed out, if you look at what tempo is up to, if you look at what Cloudflare has said about payments, if you look at what like Radius is building, which is the guys who did Project Hamilton

for the CBDC at the Fed, they're all looking at a gentic comment. is a thing that needs to be high throughput, small payments, because agents are going to be doing, like, vast numbers of micro-transactions compared to, like, you know, meatbags like us. And they're going to be using U.S. dollar stable coins, is the thesis there. Now, maybe you could use other forms of stable coins as well. I just think the market demand is not been there for them, right? We're looking at a world over 99% plus of stable coins or U.S. dollar stable coins. But the market answer so far, and again, it's early, appears to be.

stable coins with what everybody is building in this space. So curious, Rom, Chris, if you guys disagree on that or have heard otherwise. No, I think you're right. The end customer is what drives value. Their liabilities are denominated in dollars. Their assets are denominated in dollars. The policy is driving more dollar adoption. So it's going to be more U.S. dollar stable coin. The competition is just so incredibly intense. It's hard people to notice the coin-based transaction volume. It's not nothing, to Chris's point, but like all these payment companies are in bare markets because of the brutality of competition. It's hard to get a moat. I think it's the difficult category. I think I think we're still in a bare market for digital assets with these

occasional vicious rallies. Overall beta is being punished. And if this cap-back's splurge, by the number is two, we're talking like trillions. Like, that's the, that's the, that's the revenue target in 2030, right? Oracle alone has $400 billion in revenue performance obligates, one company. Microsoft's got another $400 billion in other revenue performance obligations. Then you have the value of the equity, and the value of the equity is going to be like a quantum stock if they can't pay their debt. So we're talking here, if you just sum up these number is like a significant number in the debt and equity ecosystem. And it's in private markets also.

And it's starting to impact public markets. Certainly, being public markets out of the IPO, but it's already touching companies like the hyperscalers, like a cancer, right? If you touch OpenAIA credit risk, your stock is in a bare market. It's hard for me to see what causes any change to that. When I saw that I interviewed this weekend from the see of Anthropic. I went from Sam Malton is believing his own bullshit way too much to

no, no, no. This is Silicon Valley hysteria. Well, let me go one step further on that as we think about the market impact, which is the Mag 7 have been some of the best performers in U.S. equity markets. And a lot of the driver of that has been the amount of CAPEX they had to do compared to the amount of cash flow that they had meant that they were. were shoveling huge amounts into share buybacks. If you look at what Google has done, if you look at what meta has done, if you look at what all of these companies have been doing, they basically been saying, we generate huge amounts of money through our current franchise and we're going to take that money and we're going to

buy back stock. And that produces a one way upward trend over time on stock. For every single one of the Mag 7, excluding Apple, actually, interesting at Apple is the one company that just kind of sat this one out. It was like, we'll just let the train go by. everybody else has spent huge amounts of their cash flow on this AI boom. And it's like they've all been buying back in Viti's shares instead of their own. So like one thing that I would be curious about maybe for the entire group is like,

what is the outlook on Ag 7 stocks here, maybe X Apple given this threat? Like, Rom, if the revenue's not there, these have to go this way, right? You saw me nodding my head. You said it exactly right. You know, the SMP bought back a trillion dollars of their stock last year, a trillion dollars. It's massive buybacks. It's an enormous bid. And now you're seeing revenue projections of a trillion dollars.

So to me, it's just insanity to expect that the CFOs, these publicly traded companies will say, let me give you my earnings, Open AI and Anthropics, so you can achieve your forecast for your private market venture investor. They're not going to do that. You're going to ask competition. And I, you know, there echoes of 2008. there are echoes of 2008 when you had debt that wasn't worth par that was being issued at par

and you had equity securities that were highly inflated. So I don't think it's great for Mag 7. I really don't. I think it's, but you get to position well. Like, Navidia is the, Nevidia and Apple are the two stocks above the 20-day moving average because they're not massive capex payers, right? It's an interesting place where we are there because now you have a stock like Microsoft, which is a 4 PE valuation multiple of 22 times, which is the same multiple as it was at the lows of the 22 bear market.

But don't you think that Microsoft is throwing off a lot of profit? They're generating a lot of cash. They've got a very strong balance sheet. They've got to find ways to allocate that money. You've got the same with Facebook. No. Meadow, you want to know? No, no.

No. So what you just said, what you just said was consensus view. That's exactly my point. So if you look at the free cash flow generation, it's going to zero and or negative for many of the max seven. Earnings, what you call earnings is from depreciation. It's about saying, hey, I'm going to amortize my expense over a period of time. The actual, any free cash order that is shoveling out the back door and giving it to

Alston's point to Navidia, right? If Navidia were to lower the costs by 60% and price like any other semiconductor company does, then there would be no issue here, right? The Mag 7 companies built 10,000 data centers from 2010 to 2020, and it wasn't a big deal. No one noticed because you buy from Intel and it's a competitively priced product. Navidia's pricing power is so strong that it hurts them and they have no choice. But yeah, they are no, their balance sheets are not clean.

Their balance sheets have, they used to have pristine clean balance sheets. Now Microsoft has $500 billion in debt. Amazon just issued $25 billion debt debt and Oracle's done that and issued equity. $25 billion debt in equity. They're all re-leveraging. Their balance sheets are getting dirtier. If you look back at the end. internet.com boom, right? When Bezos was going out there, raising money, he was losing money all the time for years on years.

I don't know how many years it's, you know, subsequently he lost, I think eight years in a row he was losing money, right? Or 10 years in a row or something. He achieved positive free cash flow. This is a common misconception. He wasn't issuing stock to the public and diluting shareholders because they were generating positive. free cash flow, like in the 2000s, earlier than people fall. He was a very good allocate. Yeah, he raised money, sure, but he wasn't burning money in these incinerators, which is what you're seeing now, with trillion dollar revenue forecasts.

And there's it. Yeah, I think the original Amazon story, so call it Amazon at the 2000s, reminds me a little bit more of something like internal capital allocation at a place like Berkshire Hathaway, which is to say you have a set of business. that are themselves positive cash flowing and you just take all of that money and internally reinvest it. So to an outsider, it looks like you're not making any money. But what's happening is every year the balloon that's not making money is bigger, that bigger, a bigger, a bigger, a bigger, a big, right. And that was sort of the original story. What I'm worried about here and actually,

Stefan, to your point, I think the two that are most interesting for this are Oracle, which was referenced at meta, are just shooting money out of a fire hose out of the enterprise. It's not into their own internal reinvestment. It's into Nvidia, right, to Rom's point. And that is money that's not going to go to buybacks, right? And interestingly, especially in the case of Oracle and meta, they are potentially right for disruption in this space. I'm a little bit more like, call it NID on Microsoft because,

Rom, I think you would agree. They have some business lines that are going to be very hard to disrupt from like the cloud side at my office computing compared to some of the others. And to me, like, okay, I'm going to say something that will momentarily take us into left field here and go back to Chris's point about WorldCoyne. To me, it may be that what we're learning is the most valuable thing is actually the direct relationship that you could have with a two-legged human being over time at terms of who the end consumer is and where the value like is going to be. And so we're seeing all of these giants firing money, you know, out of a hose. but I'll remind everybody is we're talking about like the attention economy at general and where people are spending time and we're talking about open-cloth, you know, all of these sorts of things. There's a guy who just bought this tiny little fintech that nobody was paying attention to named Mr. Beast, who is the number one person on YouTube with like what hundreds upon hundreds.

He has more subscribers than the United States as humans, right? Like let's just say that out loud. That is true. And this guy just bought a fintech app. I like his odds at monetizing that thing because his cost of acquisition is essentially negative. I think all of this is going to be teaching the world a pretty ugly lesson in like unit economics and customer acquisition costs. So we can't have an episode if we don't bring up a debt, right? And of course, Tom Lee, BM&R, you know, did the Mr. B.Sdeal that came under scrutiny. I like it, though, because I think you're right. It's about people.

It's about distribution no matter what. Just bring it back to inflation. Sorry, Chris, to interrupt, right? But think about the credit card facility and the payment rails that we have today. Every time you do a transaction, there's at least 3% being taken off the table. Every single transaction, which is inflationary. Because that does not add any value. It brings one margin.

It brings lots of value to Visa and MasterCard, or the POS providers. but ultimately that's inflationary. That's where we can save with stable coins going over blockchain rails. That's going to be or fintech rails, peer to peer, database to database charge. That's going to be significant benefit to the economy and be deflationary as well. That's the thing about agent. No, to your point, I love to brought it up. The thing that agents do best is they optimize.

They are extremely good at optimization. And so I think it will lead in certain cases to consolidation as they apply those payments to blockchain rails. You know, I don't know if you have been watching this, but Salada, base how to head start looks like Solana is now starting to kick into gear on the payment side. Why? Maybe it's because those agents are optimizing for cost and TPS. And so I think it's going to really force the issue as you have a certain application, whether it's a payment or something else. What's the best ecosystem? What's the best route to get there?

and how do I over-optimized for low fees and high throughput? So that should result in some consolidation, I think, in the crypto space. And it will be very interesting to watch how they optimize. And then how those projects react. You know, is it a race for TPS? Is it a raise for low fees? What does it all mean? Here's something else that's deflationary.

You're a year away from these agents like OpenClaw monitoring your desktop behavior. They see your mouse click. They see you run a workflow, they see you analyze data, they see you make a decision that could be true for investing, radiology, accounting, or process, wherever it is. And those agents will learn and then be able to form those tasks. We're one to two years from this happening. That is significantly deflationary and disruptive. You're a year to two years away, for example, from being able to hire a crypto AI Asian trader. that sits in on your CIA calls and calls out risks and opportunities.

We're getting close to that. I'm debating that one inside my head because on one hand, Rom, I agree with you. On the other hand, one of the things that I worry about is sort of call it the burgeoning, like, agent ecosystem, is what is the quality of these things as we get beyond, like, the surface level? And I guess the way I would say it is I'm pretty well situated to think about this, having traded some pretty bizarre products on the street. So, you know, I have in my back pocket, I can always just go torture the AIs about stable value wraps and see if they continue lying to me about things. And the answer was I literally did that yesterday and they continue lying to me about things. I worry that one of the things that's going to happen with AI right now and that will really differentiate like winners and losers, is the people who understand the flaws of these and get very good at implementing them.

And the people who, like, my wife sent me this thing, and called it like slop cannons, right, where it is to say people who don't know what they're doing, you get their hands on AI, you just create ever larger amounts of things that make no sense that they're shotgunning into the world. I wonder if what we're really going to have here is, like, the ultimate scissors statement of, like, people driving good from bad as they encounter these things. Like, I don't want to be the CEO on an Ermex call taking, like, malform of the AI. driven calls. I think AI agent, CFO, AI agent, investor relations team.

Yeah, that's coming. It probably is. All right. So I want to start this last segment for which we only have a few minutes by just bringing up institutions are also buying it on these rails right now. Stefani, we're talking about payments. We're talking about moving to different worlds. So BlackRock has.

has its 2.4 billion tokenized treasury fund Biddle, tradable via Uniswops Labs, Uniswap X. Apollo just signed a deal to acquire 90 million of Morpho tokens and we'll be building into that ecosystem. And yet the tokens are lacking in value capture and what seems to be moving in here are actually traditional companies building onto these rails. Chris, I know we've talked about Tempo at Stripe previously here.

and bet on that story too. So, Stefan, you talked about the potential for technology like this to be very deflationary, right, if we bring payments costs down. And I would point out that's global, not just for the United States of America. That's payments costs globally. What do we make of this trend of like institutions moving in, but token prices like sucking wind? I mean, everything's sucking wind right now, right? I mean, even inflation. I think right now is the time to invest, right?

I mean, you buy when the prices are depressed, when it's low. And then when it picks up, hopefully your selection or your research and your choices are going to pay off. I mean, Morpho is an established brand in Defi. So is Uniswap? I mean, they're all pretty established. I think a lot of the institutions are struggling to get their heads around defy and crypto. They don't know what a – I mean, how many people working at Bank of America know what a vault is? You know, they think it's actually a physical vault in a cellar versus an on-chain vault, right?

How is that governed? How is that secured? What does that look like? How does that yield? You know, who are the providers of liquidity to that vault? All of those, I think those are – they're looking for expertise. And you can't acquire a doubt because they're generally decentralized. They're people working and contributing to that protocol. But they understand being paid in tokens. Curve.

One of the biggest benefactors, I think, for stable coins to be able to swap your JPM coin with your Mercado Libre coin from Argentina with your Amazon coin, with your Shopify coin. How are you going to swap all of these around? Where is that exchange going to happen? They're going to happen on morpho. They're going to happen on all autonomously, you know, agentically, you know, so it's like atomically, right? So I think those are going to be really interesting protocols that are going to be large beneficiaries of stable coin growth. Chris, what are you seeing in the space right now?

Well, I think the first thing is you're seeing an absolute manifestation that this error of regulatory risk is over. Under the Gensler error, there's no way Black Rock or Apollo would have come into the space. It would have been very scary and they would have never done it. So like that that just proves the case that even in the absence of the Clarity Act, they're comfortable coming into this ecosystem and using this technology. That's incredibly, incredibly bullish, right? As far as price action right now, again, it's just probably so happens that these very sophisticated institutions were exit liquidity for some very early adopters, right? And that we'll hit a bottom on that and then and then things will stabilize. Stefan, you're right. You're supposed to buy in times like

now when prices are down, when fundamentals are dislocated from sentiment, this is the time. But that's not how most people act and that's not how most people behave and they're waiting for that price to pop up so they can buy even more. So it is what it is. But I think if people start following trufflation and the policy makers do the right thing around rates. Who knows? Maybe we'll have some that tailwinds back. Two observations. One is Ron DeSantis and AOC are both aligned that data center grid and utility investment shouldn't create higher prices for consumers.

So they're not AI forward. I think policy and regulation and not in my backyard risk is real. You know, you mentioned earlier, Stefan, about we got to reduce regulation and permitting costs. So it's still real and significant. They're still very significant. And, you know, the prize we should talk about last week was defy, and there still isn't great frameworks around this. So regulatory climate is better.

I think the Epstein files make it more likely that the White House loses the house, which makes the possibility of deregulation and creating a framework around Defi more challenging. So I think it's a pretty tough backdrop. I think you should really be cautious in this market climate. All right. Well, on that note, since we've already run a little bit long and we know Stefan is up late to join all of us, I want to say thank you to everybody here. And as always, actually, I want to start by saying this part before we do the quick outro here, which is on rom what you just said with the cautious backdrop thing let me ask you a quick question is this going to be climbing a wall of worry or is something going to break if no there's no

wall of worry everyone on the show has an opposite view than me right where's the wall of worry well i may be more similar to you but we're at least two for two all right so that makes sense to me because i always want to interrogate that view all right so thank you for joining us for this episode of Bips and Bips. We'll be back in one week to discuss more about how the worlds of crypto and macro are colliding. Until then, take care of everyone. Awesome. Good stuff, guys. Fun stuff. Good stuff, guys. Thank you. Thank you. Thanks. Thanks.