원문 보기
may have already begun , and according to Raoul Pal and Tom Lee, artificial intelligence and crypto could be at the very center of it. Raul Pal makes a shocking prediction for the next four years, claiming that AI intelligence is accelerating so rapidly that by 2030 the economy will be almost impossible to predict.
He believes that AI work and exponential productivity growth could push economic growth far beyond the 3% or 4% we are used to, potentially reaching 10%, 20%, or even 30%. But Pal's prediction becomes even more interesting when he links AI to crypto. He says autonomous AI agents are starting to make transactions at the speed of machines, creating a whole new economy where blockchains and crypto can become the financial infrastructure for intelligent machines.
Tom Lee develops this thesis even further. He argues that money itself is becoming software, paving the way for stablecoins, crypto, tokenized assets, and micropayments, which will become part of an entirely new financial system built for AI agents. In this video, we break down explosive predictions from Raoul Pal and Tom Lee, including the economic singularity, the economics of AI agents, exponential productivity, and why crypto could become the money standard for the machines of the future.
If their predictions are correct, the next few years could completely change our understanding of AI, money, and crypto. oday I want to return to a topic that is incredibly important. This is what everyone is worried about: the economic singularity. You see, we all spend time working in a world where we have a position, a pension, we slowly accumulate life experience, trying to do our best to succeed.
But suddenly we all start to see what I've been saying for a long time: everything is going to change . And the things you value may not be the things that will carry you into the future. And it's quite difficult to realize this all at once. At first, it causes a lot of fear. But the reality is that it also opens up opportunities.
And I'm not going to say it will be easy. So, what is the singularity? When I talk about the singularity, I mean the moment when the foundations of physics, mathematics, fall apart, and our understanding of everything begins to collapse. And this is driven by one single force—the exponential curve of intelligence.
As we cease to be the highest intelligence on planet Earth or anywhere else, everything must change. And everything has to be reconfigured. And this is a very important thing. This is becoming a black hole for the economy. It becomes a black hole for everything. And I often say this: I don't see what will happen after 2030.
Earlier I talked about 2030–2032. I don't see anything beyond that. I think 2030 is just the thing. I think there's enough evidence that intelligence is becoming exponential, and exponential curves are everywhere, to understand: we only have 4 years left. And it becomes impossible to make predictions after tha unlike anything modern economics has seen before.
His thesis about the economic singularity begins with one main force. Intelligence itself is accelerating. Pal argues that around 2030, AI could develop so rapidly that traditional economic forecasting will become extremely difficult. This may sound radical, but the events of 2026 give this idea much more weight.
The International Monetary Fund now describes artificial intelligence as a major structural shift with the potential to boost productivity and economic growth. Its April 2026 forecast estimated that faster adoption of AI could boost global growth by 0.3 percentage points in the near term and by 0.1–0.8 percentage points in the medium term.
The important thing is that this is still just the beginning. Companies are pouring capital into computing, data centers, software, and energy infrastructure, while AI continues to permeate real economic activity. The IMF reported that AI-related investments are already becoming a major source of growth in the United States.
There are obvious risks. Implementation may be slower than expected, infrastructure may become a “ bottleneck,” and markets may overvalue the technology. But even conservative institutions now recognize that AI can significantly change productivity. And if intelligence continues to accelerate, Pal's 2030 deadline may prove to be much more important than it seems.
Now, in the next part, Pal explains why he believes this acceleration becomes exponential. Here's, first, a graph I made: intelligence per unit of energy, to me, is how everything works, part of my thesis about the universal code. And you can see from the graph that it was actually driven by Moore's Law before.
Moore's Law was essentially the accumulation of intelligence through chips. But suddenly we created intelligence beyond chips. And the curve now turns into a double exponential. That's why everything moves so fast. This is because this is the first time we have observed Reed's law in any visible mechanism.
This doesn't even exist in nature. Reed's Law is the double exponent, the square of Metcalfe's Law, and that's what it all comes down to. Raoul Pal says that the key metric is intelligence produced per unit of energy. Essentially, how much intelligence can AI produce from the same amount of energy. According to him, AI is now moving beyond the traditional Moore's Law model, where progress has been largely about creating faster and more powerful chips.
With AI, technology itself can help improve the way intelligence is created. That's why Pal describes the current trend as a double exponential. His idea is that better chips, smarter AI models, faster networks, and greater energy efficiency all reinforce each other. Instead of AI evolving steadily, these improvements could reinforce each other, forcing the pace of progress to accelerate ever faster.
There is already clear evidence that AI is becoming much more energy efficient. A 2026 study found that current AI models can respond to a typical query using only about 0.31 watt-hours of electricity. And as AI models, software, and computer chips improve, this energy consumption could eventually drop another eight to twenty times.
The hardware race is moving just as fast. Nvidia claims that its newer Blackwell Ultra systems can handle up to 50 times more AI tasks with the same power consumption, while reducing the cost of each AI token by up to 35 times compared to previous Hopper systems in certain workflows. This is extremely important for cryptocurrencies because cheaper intelligence means more autonomous software can run continuously.
And as intelligence becomes cheaper, faster, and more autonomous, the demand for digital infrastructure could grow much more than expected today. If you want to stay ahead of the market insights to chang. And that brings us to Pal's next question. What happens when this intelligence starts to become part of the economy itself?
So, what do I mean? What I mean is that if GDP growth—that's population growth plus productivity growth plus debt growth, we're going to—that was the equation that the "code of everything " worked on. And now we are moving to a new version, where we are adding artificial intelligence and robots. And then we added energy density, efficiency, and intelligence management efficiency .
This is a new economy. The economics of the economic singularity. And what happens is that the economy can theoretically move faster than we've ever seen because the growth in population and intelligence, which is productivity, is exponential. So unlike a world with 3% or 4% GDP growth, we could end up in a world of 10%, 20%, 30%.
But most of GDP is driven by machines in the invisible economy. We are currently witnessing the rapid development of the agent economy. So, this is a new population. They are still in the early stages, but we are starting to see agents becoming more autonomous , performing more complex , “human” tasks at scale, but invisibly.
They start making transactions. We see this in the X402 protocol and many other large blockchains where agents are now the main focus. Because agents need to move money at the speed of a machine, not a person. Pell argues that the old economic formula of population growth, productivity and debt is evolving.
AI and robotics could become a new source of productivity, while improved energy efficiency will allow machines to produce more at much lower prices. His most daring scenario suggests that economic growth could go far beyond what investors considered normal. Pel mentions the possibility of growth of 10%, 20%, or even 30% if machine intelligence begins to exponentially increase productivity.
This is not a prediction that can be guaranteed, but there is already evidence that AI agents are moving from simple chatbots to systems capable of performing multi-step tasks. In 2026, large technology companies are increasingly creating agent-based systems designed to interact with software, execute workflows, and operate with less human intervention.
This is especially important for cryptocurrencies, as autonomous agents need digital money that can be moved programmatically. Visa, Mastercard, and major stablecoin companies are already developing infrastructure for digital payments at machine speed, while the use of stablecoins continues to expand in global markets.
Ho points to blockchains as a new financial layer for these agents, as machines don't need banking systems designed for human schedules. They can transact continuously, globally, and in extremely small amounts. If the agent economy continues to develop, cryptocurrencies could gain a whole new source of demand that has little to do with traditional investors.
And this leads us straight to Tom Lee's thesis that money itself becomes software. So, I think there are two things that will happen that agents need and people don't, you know? And the first is that money becomes software, now that it' bits. And that means that other things can start to act as money.
And in fact, we could already be transacting in this. We can think of it like, “Hey, there’s fiat, like dollars or euros, but there’s other things like gold, stocks, cryptocurrency, or reputation,” but over time, they’ll all become units of money. And banking systems are not ready for this , because literally the entire accounting book of the banking system has to use a currency, like the dollar or the euro.
If someone created a new currency, say, “Wisconsin state mone essentially completely rebuild the entire payment system for banks in the US and the world. And the second thing is micropayments, because in the future everything will happen a thousand or a million times faster than current transactions.
That is, the smallest unit will not be a cent . There will be micropayments everywhere. Even payment shares, and cryptocurrency is actually a good fit cent transactions. And this is great es can be collected very easily. Because you have the appropriate units available for this. Tom Lee's argument starts with a simple idea that could have huge implications for cryptocurrencies.
As money becomes fully digital, it increasingly behaves like software. This means that financial value can be programmed, automatically transferred, divided into smaller units, and ultimately used by machines without human intervention in each transaction. Lee believes this is important because AI agents will have economic needs that humans do not have.
An autonomous agent may need to pay for computing power, buy data, access an application, or instantly complete a transaction. Traditional banking infrastructure was built for people and institutions, not for millions of autonomous software agents making payments 24/7. This is where stablecoins and cryptocurrencies become particularly interesting.
Stablecoins already allow digital dollars to move across blockchain networks without relying entirely on traditional payment processes. In 2026, stablecoin legislation and their institutional implementation brought these assets to the center of financial discussions. As payment companies continue to experiment with blockchain-based payments, Lee is also turning his attention to micropayments.
If machines eventually complete transactions thousands or even millions of times faster than humans , the typical 1 cent payment will become too large for many cases. Blockchain systems can support extremely small denominations, potentially allowing machines to pay for resources almost continuously.
This creates a much bigger opportunity than simply replacing credit cards. If money becomes a programmable infrastructure for a software-dominated economy , cryptoassets could become part of a financial system that powers autonomous agents. There is one final element to this thesis that could make this possibility even more significant.
What will happen when billions of transactions occur between machines every day? The main idea that unites Raoul Pal and Tom Lee is not simply that artificial intelligence will become more powerful. And that AI can create a whole new economic system, and cryptoassets can provide some of the infrastructure necessary for this system to work.
Pal sees an economy where AI agents and robots become productive participants, not just passive software. Lee sees a financial system where money becomes programmable and able to move at the speed of machines. When you combine these ideas, the potential becomes enormous. This opportunity is already attracting serious capital.
According to the International Data Corporation, global spending on AI infrastructure, software, and services will continue to grow rapidly through the end of the decade. At the same time, stablecoins have become a large category of digital payments, with billions of dollars passing through public blockchains.
The bullish scenario for crypto assets becomes even stronger if AI agents start controlling their own wallets and autonomously paying for resources. Theoretically, an agent could purchase computing time, pay for data, access applications, enter into contracts, and transfer value without human intervention.
This does not mean that every blockchain or cryptocurrency will benefit equally. Infrastructure must become faster, cheaper, safer, and more machine-friendly. Regulation also remains an important variable, but it is the vector of development that makes this thesis so compelling. AI can significantly increase the number of economic actors in the digital space, while cryptoassets will provide them with the tools to transfer value.
If Pal and Lee are at least partially right, the next wave of crypto market growth may not be driven by people buying assets. And their need for machines . This is why the convergence of AI and crypto assets could become one of the most important investment themes of the next decad.
Comments 0