This analysis is absolutley brilliant in how it reframes fintech as part of a larger capital-labor shift. The comparison of crypto's $3T market cap to the UK's entire equity market really drives home how quickly synthetic value creation is outpacing traditional economies. What strikes me though is that the "robot money" thesis might underestimate how regulatory capture will shape ownership structures, since history shows capital concentration rarely happens without state blessing or atleast acquiescence. The real question becomes whether DAOs and decentralized ownership models can actually compete against the Mag7's entrenched positionbefore regulation crystalizes the current power dynamics.
Yep, this is also my fear. Full centralization backed by state power is several orders of magnitude more effective at this stage than our Web3 community
GREAT ANALYSIS LEX!! I love the main thesis also super funny that I shared the same initial first principles thinking re Cobb Douglas function when working through where the gains of AI would accrue too. Love it.
I am a fan of Lex’s thought provoking articles here. Adding my own comment here as it is germain to the argument.
As A.I. and robots replace labour then the value of money ceases to be the traditional ‘store of labour’ that it was previously… simply because people will increasingly no longer be employed.
In the current economic cycle (with high global debt, inflation and debased fiat currencies), for our ‘money’ to retain value, it MUST move to become a store of the title of real world assets (RWA) for it to continue as a medium of exchange. The key point here is that A.I. and robots cannot ‘make’ real estate.
Real estate is arguably the ONLY common thing that AI and robots can never ‘make’ and which is both finite and in increasing demand.
TPX Property Exchanges has very carefully over many hears created the end-to-end mechanism of ‘real estate money’ under a global SaaS model to meet this requirement.
Hi Lex, thank you for another great writeup. I'd love to get your next take on DePIN (seeing it as a foundational layer, quite literally) around specific investment opportunities.
15%, 3yr notes, at 65-70% LTV? risk tranche of 26%? Is the principal the current GPU value (deflated from purchase price)? How to assess risks (declining GPU rental prices?) and jump risk (TPUs adoption and efficiency vs GPUs)?
If you hover over APY, it shows different GPUs with different APYs... What are the tradeoffs?
obviously, this is just one stablecoin's approach. Albeit, doesn't seem like stablecoin's the most attractive/interesting aspect, imho.
Robot money itself is interesting of course, too. Rooting for Virtuals continued thought-leadership, not web2 Goog/OpenAI monopoly expansion... :)
I don't love packaging junk bonds as stablecoins, that's my main issue with it as ppl don't really understand what they are buying. It's not a bad risk for a large PE shop. But other than that ...
Haha. We're all sensitive to different types of risk, aren't we.
50% LTV on high utilization GPUs is not as junk as some current IG gems may turn out in 2026. Especially if they do have merely 3 year lives and are already pricing in declining prices and utilization. Seemingly pristine collateral is often modeled so and precisely what's entirely wrong...
btw, "democratizing access" to deep OTM calls with days to expiration is perhaps the biggest travesty out there by a factor of 1,000,000. Financial WMDs "Gone Wild" (despite SECs foundational mandate re 50% retail margin). So we'll see how much anyone understood anything since the 2019 ignition of that short-dated deeply-out-of-money synchronized tail wagging the TSLA/NVDA dogs.
But w/ stablecoins Ethena is the literal hedge fund and one dependent on non-zero funding rates to exist. WORSE than Junk. it was a temporary garbage assumption to begin/pretend with... But if that wasn't enough, their ironically "strategic" growth lever was to position THAT wrapped strat as "ideal" COLLATERAL on Binance and other exchanges. Worse than garbage - borderline FTX type recklessness.
Protocolization of user-owned assets with or without financial engineering is something I'd bet on over any other DeFi/RWA... tokenizing equities/bonds that are already hyper-wrapped in MF/ETFs/Managed strats seems at most 0.1-1%/yr back-end efficiency gains. Global access to US assets is literally the LEAST increased access as everyone is already clamoring for it or has access to it if they have any money to invest. Americans subsequently don't invest enough in Small/value internationally, but I guess they/we will learn that timeless lesson of diversification soon enough. And that's just existing RWAs being tokenized. Hardly anyone can model or understand all the emissions, frictions, and hidden risks in DeFi...
just my 2 cents. from soviet style meltdowns to GFC-style. been around, too. not holding my breath on USSA exceptionality or hot-takes re US dollars/treasuries being spread via "Genius Act". ;)
This analysis is absolutley brilliant in how it reframes fintech as part of a larger capital-labor shift. The comparison of crypto's $3T market cap to the UK's entire equity market really drives home how quickly synthetic value creation is outpacing traditional economies. What strikes me though is that the "robot money" thesis might underestimate how regulatory capture will shape ownership structures, since history shows capital concentration rarely happens without state blessing or atleast acquiescence. The real question becomes whether DAOs and decentralized ownership models can actually compete against the Mag7's entrenched positionbefore regulation crystalizes the current power dynamics.
Yep, this is also my fear. Full centralization backed by state power is several orders of magnitude more effective at this stage than our Web3 community
Want change … one word - ‘Yiannis’
GREAT ANALYSIS LEX!! I love the main thesis also super funny that I shared the same initial first principles thinking re Cobb Douglas function when working through where the gains of AI would accrue too. Love it.
Thanks! I ended up taking a loot at the academic research and it's like "yep robot owners win"
Incredible analysis and read 👏
I am a fan of Lex’s thought provoking articles here. Adding my own comment here as it is germain to the argument.
As A.I. and robots replace labour then the value of money ceases to be the traditional ‘store of labour’ that it was previously… simply because people will increasingly no longer be employed.
In the current economic cycle (with high global debt, inflation and debased fiat currencies), for our ‘money’ to retain value, it MUST move to become a store of the title of real world assets (RWA) for it to continue as a medium of exchange. The key point here is that A.I. and robots cannot ‘make’ real estate.
Real estate is arguably the ONLY common thing that AI and robots can never ‘make’ and which is both finite and in increasing demand.
TPX Property Exchanges has very carefully over many hears created the end-to-end mechanism of ‘real estate money’ under a global SaaS model to meet this requirement.
Check it out here: https://www.tpx-global.com
Hi Lex, thank you for another great writeup. I'd love to get your next take on DePIN (seeing it as a foundational layer, quite literally) around specific investment opportunities.
ie. https://app.usd.ai/loans
15%, 3yr notes, at 65-70% LTV? risk tranche of 26%? Is the principal the current GPU value (deflated from purchase price)? How to assess risks (declining GPU rental prices?) and jump risk (TPUs adoption and efficiency vs GPUs)?
If you hover over APY, it shows different GPUs with different APYs... What are the tradeoffs?
obviously, this is just one stablecoin's approach. Albeit, doesn't seem like stablecoin's the most attractive/interesting aspect, imho.
Robot money itself is interesting of course, too. Rooting for Virtuals continued thought-leadership, not web2 Goog/OpenAI monopoly expansion... :)
I don't love packaging junk bonds as stablecoins, that's my main issue with it as ppl don't really understand what they are buying. It's not a bad risk for a large PE shop. But other than that ...
Haha. We're all sensitive to different types of risk, aren't we.
50% LTV on high utilization GPUs is not as junk as some current IG gems may turn out in 2026. Especially if they do have merely 3 year lives and are already pricing in declining prices and utilization. Seemingly pristine collateral is often modeled so and precisely what's entirely wrong...
btw, "democratizing access" to deep OTM calls with days to expiration is perhaps the biggest travesty out there by a factor of 1,000,000. Financial WMDs "Gone Wild" (despite SECs foundational mandate re 50% retail margin). So we'll see how much anyone understood anything since the 2019 ignition of that short-dated deeply-out-of-money synchronized tail wagging the TSLA/NVDA dogs.
But w/ stablecoins Ethena is the literal hedge fund and one dependent on non-zero funding rates to exist. WORSE than Junk. it was a temporary garbage assumption to begin/pretend with... But if that wasn't enough, their ironically "strategic" growth lever was to position THAT wrapped strat as "ideal" COLLATERAL on Binance and other exchanges. Worse than garbage - borderline FTX type recklessness.
Protocolization of user-owned assets with or without financial engineering is something I'd bet on over any other DeFi/RWA... tokenizing equities/bonds that are already hyper-wrapped in MF/ETFs/Managed strats seems at most 0.1-1%/yr back-end efficiency gains. Global access to US assets is literally the LEAST increased access as everyone is already clamoring for it or has access to it if they have any money to invest. Americans subsequently don't invest enough in Small/value internationally, but I guess they/we will learn that timeless lesson of diversification soon enough. And that's just existing RWAs being tokenized. Hardly anyone can model or understand all the emissions, frictions, and hidden risks in DeFi...
just my 2 cents. from soviet style meltdowns to GFC-style. been around, too. not holding my breath on USSA exceptionality or hot-takes re US dollars/treasuries being spread via "Genius Act". ;)