원문 보기
What if the most important indicator in the world is not the price of Bitcoin, but the rate of cheapening of intelligence? One investor says it is now over 40,000 times cheaper than it was in 2020. And he believes that by 2030, the economy will outgrow its own measurement tools . In this video: the network law, the new curve, one remaining term, and the supercycle.
The man behind this idea is Raul Pal. He is a macro investor, founder of Real Vision and author of Global Macro Investor, an analytical newsletter he has been writing since 2005. He calls his main idea "The Exponential Era, Chapter One." Network law. Most people believe that new technologies develop in a straight line .
This is not true. For years it seems like nothing is happening. Then the graph becomes vertical and everyone asks where it came from. Raoul has a simple answer, and it starts with how the network acquires its value. Listen to this. And then comes implementation. Years of apparent irrelevance, and then a phase transition so rapid it seems as if it came out of nowhere.
This is not true. All this time, it has been growing exponentially. In November 2013, when Bitcoin was worth around $200, I wrote in GMI: "It's either worth nothing or it's worth a really big amount of money." From that point to its peak, Bitcoin rose 456%. So why is adoption growing exponentially? Every new user makes the network more valuable to everyone else.
This is Metcalfe's law. The value of a network grows in proportion to the square of the number of its participants. A telephone network with 10 subscribers supports 45 possible connections. A network with a million subscribers supports approximately 500 billion. As I wrote in 2021, the entire crypto market is being evaluated according to Metcalfe's Law, and therefore it all depends on implementation.
My two predictions from January 2020 spoke precisely about this. When measured against their best performance, Bitcoin is up 720% and Ethereum is up 3,455%. Let's check these numbers. A network with 10 users has 45 possible connections. A network with a million users has almost half a trillion. This is why adoption, not profits, drives cryptocurrency prices.
But this path is never smooth. Bitcoin is currently trading around $77,935. This is about 38% below its record of $126,080, set in October 2025 . The network continues to grow, but the price moves in waves with deep dips in between. And Raul doesn't stop at one network. He believes that the real story is the merging of many networks into one giant network.
He also says that most people misunderstood blockchains from the beginning . See what he says next. Almost my entire thesis about the Exponential Era combines numerous network effects into one giant network effect. In June 2022, it was Exponential Era technology through the Nasdaq, which rose 151% from its entry point to its peak.
Then in 2024, Sui came along, and that's how I look at blockchains. Blockchains are not companies selling a product. These are coordination substrates similar to the Internet itself. From January 2024 to its peak, Sui grew by 555%. This is a loud statement. So, here's the other side. Today, Sui is trading around $0.90.
This is about 83% below its peak of $ 5.35 in January 2025. The sweet spot is not where the price is today. According to Raul, this is the price of owning young networks. The fluctuations are crazy, but the implementation continues to grow internally. Chapter 2, new curve. For half a century, one rule has told us how quickly computers get better.
The power of chips doubled approximately every 2 years. Raul says that this rule is no longer decisive. Something much faster replaced it around 2020. Here is the new curve. For 50 years, the history of computing has been Moore's Law. Around 2020, the limitation stopped being the chip, and became the intelligence that that chip could create.
From 2020, the line turns into computational efficiency, intelligent output per joule, and look at the change in slope. Since 1971, it has risen seven orders of magnitude. Since 2020, it has risen another five. Moore's Law doubled every 2 years, and what replaced it doubles about every 4-5 months. This is not a faster version of the old curve, it is a completely different curve.
So we created a way to measure it—the Exponential Era Index. Intelligence multiplied by joules per dollar is intelligence per dollar, the number around which the entire economy is now organized. It is now more than 40,000 times higher than its level on January 20, 2020. The energy engine, half of the "joules per dollar", has not yet started.
The cost of electricity supply in the US is higher than it was in 2020. More than 40,000 times in about 6 years. This is the intellectual half of his index. The energy half is a weak point, and the data confirms it. In January 2020, the average American paid 13.4 cents per kilowatt-hour of electricity.
By September 2026, it was already 19.6 cents. This is an increase of almost 50%. So, chips are ahead of everyone, but electricity bills are not falling yet. Remember this because we'll come back to it at the end of the video. Now the second law. Metcalfe's Law is about people connecting with people.
And what happens when groups form within a network, and these groups connect with other groups? Mathematics becomes much more extreme. Listen to this. And when networks connect, and their members can form groups, you get Reed's law. The value of the network grows as two to the nth power. Not like n squared.
Two to the nth power. This is an exponent of an exponent. In 2023, I wrote that artificial intelligence is the first example of Reed's law in history. One way to monitor the development of AI is the time it takes for a model to independently complete a task half the time, measured in terms of the time it would take a human.
In 2019, it was 3 seconds, and as of April 2026, it is about 17 hours. Since 2024, this figure has doubled approximately every 3–4 months. This AI measurement comes from a research group called METR. It tracks how long an AI agent can perform tasks independently. Their first study in March 2025 showed that this time horizon doubled approximately every 7 months for 6 years.
In some programming tasks, doubling took less than 3 months. Raoul's point is simple: the pace is accelerating, not slowing down. Chapter 3, one term left: why should an investor care? Because Raul believes that the old economy has run out of engines. Previously, growth came at the expense of more workers and more debt.
He says both of these factors are now broken. Here is his formula. Phase two is the era of exponential growth. This is what is happening around us right now. All key technologies are advancing exponentially simultaneously for the first time in history, and each one fuels the others. This is one of my four models, and as I noted in GMI, the era of exponential growth is driving the adoption of networks.
" Code of Everything" stimulates liquidity. Here is the logic. GDP growth is population growth plus productivity growth plus debt growth. Demographics as a driving force is dead, and debt is a Band-Aid for a gunshot wound, so only one term remains. We can't change demographics, but we can change productivity through technology.
This is the essence of the era of exponential growth. In April 2021, after 16 years of writing GMI, I wrote, "This is the first time I'm completely changing my macro model." I left the old one next to her just in case . It's a rare case of a macro investor abandoning their own model after 16 years, but look at the logic.
In most parts of the world, the population is aging. The debt is already huge. This leaves productivity, and the only thing that can raise it on such a scale is technology. That's the whole point. So what feeds it? Raoul borrows an idea from futurologist David Mattin: a cycle where energy, computing, and intelligence feed off each other.
And he says the cost of almost everything is falling along with it. See what he says next. The exponential era runs on a flywheel , as David Mattin says. Energy paves the way for computation. Computation opens the way to intelligence. Intelligence opens up new sources of energy. The cycle is contracting.
The best AI designs the best chips. Better chips enable better AI. And these networks don't just grow separately, they connect with each other. So each subsequent revolution is faster than the previous one. Let's start with energy. Solar energy obeys Wright's law. Every doubling of total installed capacity reduces the cost by about a fifth, and this relationship has not been broken for 50 years .
Governments are pushing this transition through a carbon price, and emissions quotas were one of my predictions in 2021. They grew 111% from the time of the forecast to the peak , and every forecast here is measured in exactly this way . And this applies not only to energy. As I wrote in 2022, technology makes everything cheaper.
Even the technology itself is becoming cheaper, creating an endless cycle. In 1975, a megabyte of memory cost about a quarter of a million dollars , and by 2023, it was about a tenth of a cent. Think about this number regarding memory. From a quarter of a million dollars to a tenth of a cent. This is a price that is 250 million times lower.
Solar energy shows the same pattern. Every time the world doubles the number of panels installed, the next panel becomes cheaper. This is a flywheel in real life. Cheaper energy means cheaper computing. Cheaper computing means cheaper intelligence. Chapter four, supercycle. This is where things get really big.
Technology companies are investing record amounts in data centers, chips, and power plants. Some call it a bubble. Raul calls this the beginning of a supercycle, and he knows a way to determine exactly what it is. Listen to this. In May, I said that the computing cycle and infrastructure buildout we are in through 2030 should have pauses for rethinking and periodic shocks, but without a systemic break in the core engine.
Five-year productivity figures have not yet caught up. As I explained in June, it's like three dominoes. The first one has already fallen. The second one is falling now. The third one hasn't fallen yet. Intelligence is about 6 months ahead of capital spending. Capital expenditures are four to five years ahead of productivity.
The last domino is still standing because the cascade did not reach it, not because it stopped. And you'll know it's fallen when five-year productivity growth exceeds 2% for the first time since the early 2000s. We believe this could create a supercycle, the greatest development in computing power, energy, and credit that we will ever see.
So, we are assessing his condition. "Coal Mine" tracks eight warning signs of this development and rates them on a 100-point scale. The indicator is in the 80s, none of the eight signals worked, and the "bottleneck" is energy. So, again it's energy, the same weak link in his index. The chips are ready, the models are ready, the bottleneck is power supply.
That's why much of this development now concerns power plants and transmission lines, not just chips. And if productivity growth does indeed exceed 2%, it will be the first time in over 20 years. There is something else that Raul wants you to understand. Experiencing exponential growth doesn't feel like progress at all.
See what he says next. But that's the thing about having to go through it . Exponents don't feel like growth. They are felt as instability. And if intelligence continues to develop at this rate, at some point the tools will stop working—I think around 2030, and that will be the economic singularity.
So, the main conclusion. Raoul Pal sees three forces at work. Networks that grow in proportion to the square of the number of users, intelligence that gets cheaper every few months, and the latest economic engine—productivity— now powered by technology. He may be wrong about the terms. The falls along the way can be brutal, as Bitcoin and Sui are currently showing.
But if he's right, the next few years will be chaotic, and those who understand this curve won't panic when it changes dramatically. If this helped you see the bigger picture, please subscribe. We break down the most important ideas in crypto and macroeconomics one video at a time.
Comments 0