When you really boil down the entirety of what’s going on - the current state of the market boils down to one simple observation. The people buying the highs of the 1999 tech bubble were right. But they were too early.
Sama’s most recent interview highlighting how “humanoid robotics are 100%” on OpenAI’s roadmap means we are now in the end game move of this observation coming to roost. This end game move pits the economy of bits directly against the economy of atoms and its political backers - a new regime with historic implications.
After the tech bubble, people decided to invest in everything BUT tech. And for the largest part, these investments have been disastrous - not only in terms of returns, but also in terms of their negative impacts on society. Private equity companies gutted the American middle class and shipped American manufacturing and supply chains to a country that we now count as a geopolitical enemy. And ironically the equities of said enemy haven’t even worked, because they were - as it turns out, communist. The military industrial complex grew unchecked, into an uneasy world policeman that ultimately failed to install democracies in its wave of unpopular invasions. Easy credit and debt driven expansion led the United States to turn its entire housing market into a global Ponzi scheme and caused a global financial crisis that got the entire world hooked on money printing, and fiscal and monetary excess.
At the core of all these “bets” have been the idea that you can engineer growth in the “real world”. “Real” being housing, emerging economies and so forth. “Real” being constrained by bodies, not compute or IP.
But every real world bet doesn’t have a high enough ROI - because it’s fundamentally not scale invariant the way that technology is. The tech giants of today grew via network effects, not debt. But physical infrastructure projects and private equity require low rates to work - borrow at 4% and lever a 12% return 2-3x and you’re compounding capital very effectively. The government has subsidized these types of projects to a ridiculous extent via like kind taxation exemptions, as well as so called “opportunity zone” tax credits
The “real world” types always would smugly observe, “sure you have your fancy Youtube or Algorithms but that’s not gonna build roads”. In the meanwhile $1 of debt fueled road investment drives $.3-.4 of GDP growth whereas the big infrastructure pushes they cited post WW2 were in the wake of a baby boom and a much lower base. Physical investments used to pay themselves off. But they haven’t for 30+ years. Now - a consortium of corrupt unions proximate to politicians controls an endless deluge of grift ironically termed “The Inflation Reduction Act”.
But everyone loves the ‘real economy story’ - especially politicians who get kick backs for financing all the projects. That’s why the politicians are so rich despite having low salaries. When Joe Biden says he’s a car guy, what he really means is he and his family members / various related entities are getting paid off by auto manufacturers.
Auto manufacturing is interesting because the market has priced that Tesla is by far the most valuable car maker for years. People now take it for granted that a monthly move in Tesla market capitalization is +/- a General Motors or Ford. This is because there’s a fundamental understanding that cars are tech products - especially electric vehicles. And the tech arms of most other car companies such as Ford and General Motor are utterly inept, and only exist to hope to “close the valuation gap” with Tesla. Which of course they won’t.
Elon is a contrarian and strange beast because he’s been investing in the real economy, and doing “hard tech” projects such as Spacex. Which essentially nobody wanted to do - because it’s so risky.
But there’s a new chapter. Elon has fallen from grace. And he is the last vestige of the ‘real economy’ interesting with the world of technology.
The rise of Artificial Intelligence is the retort to the “Real Economy Resource Extraction” folks. How is your tech company going to flip a hamburger? Is no longer a smug or rhetorical question. Now there’s an answer. Humanoid robotics with bolted on vision technology, powered by AI systems. The world of bits and the world of atoms will collide - it just requires a greater amount of compute to make it happen.
Fittingly, Elon is suing OpenAI and casting Sam Altman as a megalomaniac. This is because - at its core - Tesla is an ‘atoms’ story - and if more powerful models can obsolete Tesla’s AI engines, which rely on vast amounts of data - he faces an existential threat. If you listen to Elon’s argument for why Tesla is an AI winner - it’s because he has vast amounts of training data. Whereas when Altman discusses why the New York Times IP lawsuit is irrelevant, he asserts that in the medium term training data is not going to matter because models are getting better at training themselves.
While targeted at the NYT lawsuit, this is key to understanding the existential conflict between Altman and Elon. If Altman is right, Tesla is Ford or GM. Another real world economy bet that is doomed to the scythe of real world AI application. And if you look at the price of Tesla versus the rest of the index, it is certainly beginning to trade that way.
The fundamental vision of an AI enabled economy is at complete odds with the previous investment cycles of the past 20 years. The investment cycles that gained traction due to the crash of the 1999 tech bubble.
In an AI based economy, new resource extraction methods are invented - such as fusion power. That instantly obsolete oil production. This is - perhaps, why Altman is even deigning to raise $7 trillion and is being invited to talk so frequently in the Middle East. There is existential fear that the entire atoms based economy is on the way out, and it’s time to invest in AI now or get left behind.
New methods of warfare and weaponry will be designed by AI systems that make tanks analogous to mid-evil swords.
Workers will be fully autonomous.
But the key thing to understand and why these trillion dollar investments are taking place at Sovereign Levels - is because NONE of the value capture will be by existing companies. Elon is the first of many conflicts and beefs. If Sam Altman’s nuclear fusion start up takes off, the share prices of oil companies will collapse. If new AI warfare comes online, Lockheed Martin will implode. And it’s not just the traditional “atoms” companies that are at risk.
If we all start getting all our information from local models run from Nvidia chips in our homes, even Google Search will go into free-fall. What happens if there’s a dynamic video game personalized to your preference instead of GTA?
There are no traditional “beneficiaries” of a complete paradigm shift, where the entire value complex migrates to whoever has the most correct weights.
As this becomes increasingly clear - with the rise of humanoid robotics, we are going to get the first real wave of backlash. I think the big markets catalyst will be Robotic McDonalds. The restaurant industry has about 12.5 million employees. Imagine if most of that got automated, and a couple of other low skilled industries. You get to 10-15% unemployment very quickly.
This is no longer a hypothetical or speculative. Bezos is investing heavily in humanoid robotics. Elon is doing the same. Softbank is all in. And Altman has explicitly stated that robots are coming.
The uncomfortable reality is that the robotic wave is going to cause huge amounts of unemployment, and suddenly make life very dangerous for technologists and AI enthusiasts. This is something that Altman highlighted explicitly in his most recent Fridman interview - noting that assassination risk was something he’s considering with the launch of GPT5.
I keep tweeting about the “3 million” key employees. There are about 3 million people propping up the entire AI based stock market. And this is probably on the high side because it includes support people not core backend devs or researchers.
Singapore’s population is 5.5 million. The population of the UAE is 9.4 million. The population of Switzerland is about 9 million. Together, these 3 small countries and a few others could absorb the vast majority of high performance global AI talent.
The idea of IP - not population- being the driver of value capture, ironically should not be at all foreign to the current breed of equity investors. EM economies continue to drive global GDP, but their stock markets are at multi decade lows versus Western markets precisely because it’s the owners of the tech who get the margin, not the low skilled employees. The paradigm shift of AI is that basically every company involved in any “bits” industry becomes an Emerging Market, and therefore will trade at a similar multiples (currently enjoy about a 30% premium).
This takes us to the real near term use case of crypto.
Governments are not going to want to let the 3 million leave. They’re going to try and lock them in. Or try and deploy hundreds of billions of crony capital as Biden is now doing with Intel to bribe them to stay. But the rockstars won’t stay or work at Intel. Imagine the smell. These are the world’s most tech proficient people. Many of them grew up in the backdrop of the tech crash and 2-3 decades of disparaging “nerds” and do not have strong loyalty to “their roots”. In fact, many of the best engineers are skilled immigrants from India or China who only vaguely have ties to America. And they know just as well as anybody that there are huge debts that need to get paid off - accumulating funding projects that did not for the most part benefit them.
So - simple roadmap. The robots get deployed. It starts becoming obvious that it’s not just coders who are going to lose their jobs. Unemployment spikes. The Fed eases. There’s another push in risk assets. But — whether Trump or Biden wins, after the unemployment hits. There’s going to be a spike in violence in San Francisco.
The politicians won’t respond. In fact, politicians like AOC or Trump will gain more traction as the problems accelerate. Trump has already called Meta - the largest developer of Open Source AI models - an “enemy of the people”. The Donald borrowed this phrase from the history’s most infamous populist who believed in the “real economy” and big infrastructure projects. And he’s using this term for someone - who for all intents and purposes, helped him get elected the first time. Imagine his sentiment towards other tech companies.
CBDCs are going to be weaponized by right or left wing populists to financially censor people who try and get out. Geopolitical excuses are going to be given. Note that governments keep relentlessly pressing on CBDCs despite the fact they’re deeply unpopular with ordinary citizens - and are not being “requested” by any large tech companies.
Thus - because fiat currencies are going to be used to try and contain AI talent, the US dollar, Euro or Yuan are not going to be the base layers for the AI economy. Anyone seeing a configuration of 2 old men - Trump and Biden - as presidential candidates - who barely understand email and naturally hate tech workers, will intuitively grasp this. Tech workers cannot fundamentally trust majoritarian systems to ensure their property rights or their physical safety which will be increasingly endangered within the next 2 years.
There will be a full scale flight of capital out of the system. A giant Whoosh as the key talent moves offshore and ships core IP back on-shore. No, not all the salespeople milling about downtown San Francisco. Only the small town of people that actually matter. The new wave of IP won’t just be chatbots, or video games. It will include a robotic workforce, energy generation and a wholescale replacement for the so-called economies of Atoms. Which perhaps is better dubbed henceforth The Economy of Smithereens.
There’s another possibility is that the tech elites have already fully captured the government, and that this is all “part of the plan”. In this event - the base case is closer to global government and mass MMT. While it’s an interesting bet to make via things like Worldcoin ,that is a lower likelihood in my opinion, than there simply being no plan. No great conspiracy. Just a Great Whoosh of the economy of atoms giving way formally to technological superiority.
So - going back to the original statement. The 1999 tech investors were right. But they were too early. And being early is indistinct from being wrong. As now - we live in the world of consequences of investing in the wrong things. Foolish infra projects. Foolish wars. Endless accumulation of debt that now has higher and higher rates associated with it. Lost generations of productivity. Cratering birth rates.
You throw robotic replacement of the work force onto this powder keg, and there will be consequences. These consequences are not really being considered by the vast majority of investors across many markets.
And that is the essence of The Big Whoosh. The sound of the entire IP value capture of the system flowing outside the grasp of the future authoritarians before it is too late. And it will be too late by the time the CBDCs are implemented and deployed by 2026-2027.
What is the use case of crypto? It honestly hasn’t arrived yet. But we are on the cusp. Crypto is how AI gets deployed and licensed back to the decaying carcass of the Empires of Atoms while retaining an underlying economic model and serving as currency for The 3 Million High Value Citizens.
We are still relatively early to this thesis across many asset markets - especially regarding stocks that have run too far on a Utopian view of the world. There’s time to get appropriately positioned as market consensus for the most part - is that there will be no Big Whoosh. That the talent will stay onshore. That the Old Boys will win again. That it will happen naturally and continuously and somehow benefit incumbent players.
This natural preference towards equilibrium conditions might have made sense if there weren’t robots on the verge of flipping burgers. It’s not the world of Bits vs the world of Atoms. Soon it will be the world of Bits vs the decaying empire of smithereens.
Some will view this as a threat. Others will ignore it as delusional rambling of a bull market. A small number will view the Big Whoosh as the opportunity of our lifetime.
“See, the world is full of things more powerful than us. But if you know how to catch a ride, you can go places.”
- Neal Stephenson, Snow Crash