The Set Up
Bitcoin and stocks realize .50 correlation, and altcoins are higher bc they have more economic cyclicality (users/VCs) / Trump beta. Gold is also realizing heavy stock correlation whereas bonds are back to negative correlation, as the Fed put kicked in with the Yen Unwind.
Gold vs Long Term Treasuries are realizing heavy beta to S&P because we know there are cuts coming but what we don’t know is whether or not the US government will keep spending like a drunken sailor. We have a suspicion they will which is why stocks are up, and also why gold is trouncing long term bonds. But on the margin - if there is a big unexpected push towards cutting spending or raising taxes from either political candidate stocks drop, gold drops bonds rip.
But in practice nobody ever really stops the fiscal/monetary bonanza. Elizabeth Warren and Trump both pressure the Fed to cut rates - bipartisan absurdity. So political gridlock is the only thing that can slow down the fiscal/monetary impulse. Per Polymarket odds a democrat sweep (President, Senate, and Congress) is 21% while a republican sweep is 29%. Either Sweep situation means the government is empowered to do things which means it spends a lot of money it isn’t collecting in taxes. 50% odds.
But while both benefit from brr/ spend , Bitcoin and stocks are not the same. Bitcoin has no earnings. Stocks do. Bitcoin’s relationship with interest rates is indirect. It has a fixed supply so increasing money supply increases its perceived value as well as the opportunity cost of holding it. Stocks have a much more direct relationship. If corporate bond yields drop, it becomes more appealing to issue debt and buy equity.
Stocks and bonds are always direct alternatives. But. Microstrategy’s tendency to issue debt to buy BTC, and Coinbase’s questionable segregation of assets are 2 reasons why Bitcoin also likes lower corporate bond yields. If there are questions as to whether MSTR won’t be able to roll its debt, or what happens to customer assets in the event of coibase insolvency - BTC drops. So that’s a second element linking rates and market valuations somewhat causally to Bitcoin in a way that isn’t innately related to BTC.
For these reasons I don’t expect the BTC stock correlation to break down in the near term. But I think this correlation is a gift over the long term to own crypto assets and short US equities over longer periods
this ‘correlation cointegration’ paradox isn’t particularly novel. in the QE era, stocks went up intraday while bonds went down for nearly 10 years straight while both constantly making new highs. Useful nonetheless for risk management. Part of my premise is that Crypto/equity correlation will hold and improve the volatility adjusted returns of a mixed strategy.
The criteria
I expect the real long term driver of the delta of stocks and Bitcoin are 1] US judicial system / property right legitimacy 2] whether or not AI will flow through to corporate earnings in aggregate in a meaningful way
If you think the US is going to become a left wing banana republic, equities which currently trade at 24x forward earnings go to MSCI World Ex US (15.2x) - down 30%+. In this case, demand for Bitcoin and Gold probably increase substantially – as you value asset protection and not having to deal with US banking authorities storing value.
If you think the US is going to solve its political differences / normalize over time and avoid things like stacking the Supreme Court, or prosecuting / killing political enemies, and keep its egregiously corporate friendly policies - then why own Bitcoin when you could own global growth engines like META, MSFT, and Nvidia?
It’s worth noting that the US is about as corporate friendly as you could possibly imagine. Corporations essentially write legislation, and realized tax rates for publicly traded stocks are at 40 year lows with the deficit near 40 year highs despite the fact 1% of Americans own 44% of stocks. And this lasted even through an inflation crisis. So the bar is fairly high for the US to reprice in a corporate friendly way unless Trump gets a Republican house and senate and goes wild.
Political repricing happens all the time in other markets - most recently with Argentina (ARGT) outperforming Brazil (EWZ) by over 220% in dollar terms over 5 years. My basic contention is that the US is far more likely to become like Brazil than Argentina. But I’ll get into that.
The second question is whether or not AI is going to drive real economic benefit. The market is pricing that real earnings growth is going to happen to some extent. From the 2022 trough of despair, forward earnings estimates are 14% higher. The top 10 names in the software sector are pricing 17% growth and margin expansion from 46% to 50% - which is very aggressive given that it tech is the largest sectors of the market. And because Nvidia is such a high weight in the S&P 500 now, it drags up index growth expectations (82% growth multiplied by a 6% weight ramps the average index growth expectations by a full 5% depending of course on how you calculate)
Despite the narrative that it’s all fed easing, a lot of movement in the market year to date is based on earnings optimism that’s been priced in from the AI boom. The bond yield for example on the high grade corporate credit index is basically unchanged since the start of the year.
But the market certainly isn’t expecting anything resembling what some of the AGI hype lords are peddling -like GDP going up 40%, mass automation etc. The conference board recently said that 2025 GDP growth actually decelerates from 2.5% to 1.7%. And this is with a 6.1% of GDP anticipated deficit. If there was some kind of AGI utopia coming those estimates wouldn’t be coming down. And interest rates wouldn’t be coming down aggressively if the economy were doing very well because of worker robots, self driving, AI agents or various other memes.
So essentially the apriori market pricing is that AI benefits corporate earnings a lot more than the overall economy. It’s a bit ironic because that’s what the market has told us since this all started in winter of 2022 - AI expands margins, and grows revenue without flow through to GDP. Which is basically the same story as the permanent tech bull started in 2011.
The Thesis
So the contrarian but right thing in current markets is that AI is no different from the cloud, or mobile devices, or any number of distraction economy developments that drive great results for big tech and little to no result for society.
AI makes the distraction economy more efficient, by a large amount. And probably doesn’t have flow through to other sectors. One of the big “tells” of this, is the performance of the genomics sector. Which Jensen, Altman and others hype as the primary beneficiary of AI “accelerating the rate of innovation”. The sector is down over 5 years, and the “AI drug discovery” win - at Moderna, has been a political disaster such that Biden/ Harris don’t even mention the vaccine on the campaign trail.
Bryan Johnson, longevity influencer and previous biotech fund manager - insists that AI is changing everything, even making immortality possible. He was a large public investor in Ginko Bioworks. He wrote in a 2018 promotion of the company that the company would likely rearchitect the gut microbiome, and make genetically engineered trees that could “grow” whatever we want. The stock is down 95% from its IPO.
Johnson is now an influencer with an e-commerce product suite, which AI will undoubtedly substantially help promote. Ginko makes for a great story and a terrible stock. This is a microcosm of a macrocosm - that AI is going to make really good ads, addictive porn, and tell compelling e-commerce stories. Even for longevity! Fantastic for selling olive oil and getting clicks. Not fantastic for making vaccines without heart complications
When you zoom out - what’s really happened in US markets since the GFC is a massive unfunded fiscal transfer to the distraction economy. If you took corporate tax rates all the way back to their 58% highs, you’d wipe out the entire net income expansion in every publicly traded stock since 2006 and you wouldn’t even erase the deficit. The unrealized tax Harris proposed if applied on the top 1% wouldn’t nearly close the gap either - just on the debt side. When you include the $70t of unfunded liabilities (primarily from social security and medicare) you’d need totally unprecedented cuts in both government and personal spending to right size budgets.
We all know this isn’t going to happen. People don’t want to save. They want the government to forgive their student debts. Send helicopter money from thin air to give them starter homes. Give them free ozempic bc they’re too undisciplined to exercise.
And - if you happen to be a hardworking immigrant who finds your way into the US. Or a nefarious one. You’re not coming in with a strong relationship to 1990s America or whatever imagined golden age there was. There is no real reason why you would feel the need to pay off the accumulated deficits spent for other people, and their parents. Social Security will be underfunded as of 2033 so it begs the question… why would I have to cut spending to pay for all these old people who I don’t know or culturally relate to?
So what I think is going to happen - is pretty simple. Trump stated during his debate with Kamala and on the All in Pod that he’d consider naturalizing anyone who went to community college. So even rhetorically - it’s not like he’s the radical immigration hardliner he claimed to be in 2016. Throw on top of that the fact he deported 1/3 of the people Obama did and nowhere close to the millions he said he’d deport, along with not building the wall - and you start to see a clearer picture.
It’s the exact same picture you saw in the UK. Brexit riled people up. Got them to vote. But then the supposed far right hardliners were more beholden to corporate interests who wanted heavy immigration for housing prices and to lower the ability of workers to negotiate effectively. And - because the far right didn’t follow through - you ended up with leftists (Starmer) winning in a landslide.
Populists need population, and population in societies with low birth rates come from immigration. GPT5 won’t be allowed to vote even if it’s as purportedly intelligent as a grad student. There’s no reason for immigrants to feel responsible for debt incurred before they arrived. So it follows that confiscatory policies are going to be politically popular and won’t be influenced by AI. When I say confiscatory I mean wealth taxes, CBDCs which put different sales taxes on different income brackets, financial transaction taxes, property and inheritance taxes.
So the question is – does AI transform the economy first, or does it become clear that left wing populism is where society is headed first?
I think the answer is obvious. GDP decelerating, things like the biotech sector tanking, and rate cuts mean that AI hasn’t delivered growth. AGI probably isn’t coming in 2025 or 2026. Even Altman said it’s 1000+ days out in his hype blog post. Stocks expanding margins with high revenue growth won’t really flow through to common prosperity or GDP - because a lot of this revenue growth will just be generating net extractive economic activity. Like watching porn. Scrolling brain rot endlessly. Or losing yourself in immersively generated AI worlds.
These are great businesses with high repeat rates, but not revolutions.
Not to mention stocks are already very expensive and this margin expansion and growth hasn’t happened yet. So we’re also banking on new models being more useful - which is TBD
You throw on real world problems like dock strikes, wars, etc - and you will get a picture of disconnected elites talking about how AI is going to change everything while the average person is stockpiling frozen chicken legs.
This situation - namely people saying “eat the rich” - is going to be the wedge that sends Bitcoin up and equities down. I think it will happen before AI progress accelerates to the point to where that narrative gets discarded.
Unlike other industrial revolutions where millions were employed, the global AI industry optimistically employs 200,000 people (Nvidia is 30k - of which about half at most are relevant to core IP). So it can easily move. Wealth taxes would uniquely affect these people, so it’s reasonable that it will, in fact, move.
This creates a situation where you have a lot of brain drain, crime and currencies weakening across G10 countries, along with equity derating and deleveraging that happens in burst - at random across the developed world. In the end you’ll have something that looks like Snow Crash - with standard governments sort of presiding over luddite empires that reluctantly buy tech from decentralized gateways.
I think what’s interesting is that AI progress is virtually certain to continue regardless of this - so at some point, you’ll end up with what I call “The Post Fiat” reality – where AI becomes so powerful that it can actually move the needle on societal problems generated by the government and their abuse of their currencies.
Earlier - I was far too optimistic about when this would occur. I thought by now I’d be talking to an AI agent to write this post. when the reality is closer to me doing some light brainstorming / pressure testing.
So - what probably happens first is that things get real, AI moves offshore, and eventually all the hype becomes true. Quite similar to the “AI is 1999” argument proposed by Thiel in recent interviews
Where it all ends up
At a high level this configuration - namely that capital flight / asset protection is the dominant use case of crypto, not tech - is largely already priced in to Bitcoin dominance and the alt market. It’s also priced into Gold. Things are real right now.
But I don’t think this is by any means consensus - given current market narratives around AI, stocks, media coverage, things like OpenAI’s fund raise, or the types of discussions happening at crypto events focused on tech/web3 applications. Overall crypto market caps are below 2021 highs while the S&P 500 is +20% higher than 2021, despite Bitcoin’s hash rate being over 3x higher and ETH inflation dropping aggressively due to a seemingly successful merge.
You also fundamentally can’t predict AI development. And AI development can cause entire equity indices to spike upwards, adding gap risk to equity shorts.
But what is more predictable in my opinion is that: a] fiat currency crosses will continue to be correlated to risk assets as a function of retail traders, capital flows, and geopolitical considerations b] risk assets will not skyrocket over time because of politically driven de-rating - at best they’ll earn their way out of high multiples and stay flat c] heavily indebted governments won’t pay enough carry because they’re incentivized to depress their rates to make their debt nominally more affordable d] there will be political pressure to devalue currencies to win manufacturing votes across G10 economies adding an additional layer of intervention gap risk to already high beta currencies
So at a portfolio level this basically looks like swapping between Bitcoin, Gold, Silver, ETH and other networks that you could call “stores of value” (tbd) and shorting FX crosses that have “risk on” characteristics to manufacture a negative beta over time.
using AI to help manage the process/ create content about it
As I make money, deploying capital into moon-shot type bets specifically betting on AI applications developing offshore and that being a primary use case of crypto (right now, it’s mostly money and gambling)
so before you compulsively ask “what’s the trade” if you wanted to make a tracker index it’d basically look like X = [Long BTC, Gold, Silver, ETH] [short all G10 FX crosses with correlation to S&P >.2] X gets vol targeted to avoid ruin risk (15% max peak to trough dd). this is fairly capital light, or can be bc of liquid futures/swaps - so let’s say typically takes up 20-40% of capital. then the cash balance is deployed in market neutral crypto staking, quant strategies targeting treasury like returns, or other yield trades.
As X goes up in value, PNL gets deployed into coins in 2-3 month uptrends vs BTC/ETH that correspond with overall thesis on crypto facilitating a move of AI offshore, or - to some extent the rise of financial nihilism downstream of the lack of social mobility that will come from the situation I’m describing.
This whole configuration is wholly inappropriate and impossible to implement for most people who don’t live in a tax free jurisdiction and don’t have good custody solutions for crypto assets. But I enjoy writing about my process / hope I can do that some more. Will leave you with this quote from Rust Cohle
“Transference of fear and self-loathing to an authoritarian vessel. It’s catharsis. He absorbs their dread with his narrative. Because of this, he’s effective at proportion to the amount of certainty he can project. Certain linguistic anthropologists think that religion is a language virus that rewrites pathways in the brain. Dulls critical thinking.”