Original Source:《When Shift Happens》
Original Compilation:Felix, PANews
Macro investor and Real Vision co-founder Raoul Pal once again appeared on the 《When Shift Happens》 podcast, delving into why the AI race is the largest capital event in human history, and why cryptocurrency holders are in an advantageous position. Raoul Pal explained the economic singularity, why traders always lose to long-term holders, and why he continued to buy during the recent pullback.
PANews has compiled the highlights of the interview.
Host: You shared and commented on an interesting video a few weeks ago. It satirized the US stock market's constant rise, saying you just "buy the dip" to make money, borrow money to buy if you don't have any. It's not a pyramid scheme; it's the dip-buying plan. What's really going on with the stock market?
Raoul Pal: There are two main reasons. First, obviously, is liquidity—we are witnessing an expansion of liquidity. The other thing is, we are in the most extraordinary period in human history, and nothing else matters. All capital is flooding into artificial intelligence. This is the largest race ever. It's a race among nations, among companies. So, of course, it will suck every penny of capital because you can't slow it down.
Host: Tell me about this race.
Raoul Pal: In this world, no one will allow a single superpower to monopolize AGI (Artificial General Intelligence), so there must be two, and only the US and China can afford this race globally. Game theory suggests no nation will stop now because stopping means the other side gains an advantage. Even in a scenario like OpenAI going bankrupt and running out of funds, the US government would immediately auction its assets to companies like Microsoft, Google, etc., never allowing a single company to have a monopoly. This game is too big; no one will stop. It's a game of converting "units of energy into units of intelligence."
Host: Too big to fail. So, just "buy the dip" forever? Will it ever end?
Raoul Pal: I wrote in 《Global Macro Investor》 that it won't end unless it reaches the economic singularity. The economic singularity is when the system can no longer cope with the pace of technological development. You know the magic formula: population growth + productivity growth + debt growth. When you count AI and robots as population, our maximum population is currently 9 billion, but we could reach 18 billion, 100 billion, even a trillion intelligent agents. With 10 billion or 50 billion agents, the economic system simply cannot function as originally intended; it runs too fast.
Almost all past technology adoption followed Metcalfe's Law (PANews note: network value is proportional to the square of the number of connected users), showing logarithmic growth. But AI is the first observable case of "Reed's Law" in human history (PANews note: the value of a network scales not just with user count but exponentially due to group formation), which is exponential of exponential. It's estimated that by 2028, the annual text output by AI will exceed the total text output by humans from the Gutenberg printing press to the present.
Host: Anthropic's interview today also said their Q1 expected 10x growth but achieved 80x. That's incredible.
Raoul Pal: Yes, the economic singularity happens when you have economic agents (AI agents) capable of instant capital formation, which is what meme coins signify: instant capital creation and destruction. They can instantly build digital businesses, quickly capture markets, and exit when opportunities fade. In this economy, where is the role for large traditional corporations? Who are the workers? The system already isn't functioning normally. Because carbon-based life (human neurons) operates at 1 millisecond, but now we have electricity passing through sand (silicon, the second most common element on Earth) creating intelligence that is six orders of magnitude (a million times) faster than human neurons. It's insane.
Host: Given AI's exponential explosion, many in crypto want to switch to AI, thinking crypto is boring now, even fearing getting stuck. How do you view the investment choice between AI and crypto?
Raoul Pal: Even though the AI industry is extremely hot, I still believe that over the long cycle, cryptocurrency remains one of the investments with the optimal risk-reward. But it is indeed hard to compete with chip companies like Nvidia because they are the core part converting energy into intelligence. However, cryptocurrency has an "infinite TAM (Total Addressable Market)." Around October last year, we witnessed the birth of the agent economy (AI Agents). When these agents begin scaling massively, they will have their own wallets and conduct business on-chain. Previously, our estimate of the crypto market reaching $100 trillion was based on human users; now with unlimited AI agents, it completely changes the game. Moreover, the Fed will run the economy like in the Greenspan era, relying on productivity miracles to reduce debt-to-GDP, fiat currency depreciation won't stop, and the entire financial system is moving towards blockchain infrastructure, so you just need to front-run the institutions. The worst is over because global liquidity is accelerating.
Host: So for you, Bitcoin falling from recent peaks back to $60k isn't a bear market?
Raoul Pal: It's just an uncomfortable correction in a bull market. I've been in crypto since 2013. A 50% Bitcoin correction is normal, and other altcoins fall even harder, like Solana last cycle dropped 80% before skyrocketing. The difference is, the 2021 correction was very fast, a sharp drop followed by a quick recovery; this correction is more volatile, taking months, so people feel very painful. But look at it another way: the longer the consolidation, the longer and larger the potential bull market after the breakout.
Host: The problem is in 2021, the fall was fast, the rise was fast, but this market is choppy, time-consuming. Also, some well-performing companies (like stablecoins, RWA) don't have tokens; retail can't invest, making people feel the early fortune-making promise is broken.
Raoul Pal: I don't think that's true; product-market fit is king. Just because your altcoin hasn't risen doesn't mean the promise is broken; the market owes you nothing. People got used to the easy money days of previous years, but liquidity in 2024 is still somewhat suppressed; we haven't entered the real "banana zone" (referring to crazy upside). While ordinary people may not be able to buy equity in stablecoin companies, that's not the problem at all; you just need to hold the underlying Layer1 tokens. This is our "Universal Basic Equity." If a large part of the future economy is dominated by AI and agents, and they use crypto networks, we just need to hold Layer1 tokens to share in their success. We didn't have this opportunity in the internet era; now there's no excuse to miss it.
Host: What did you add to your portfolio during the recent dip?
Raoul Pal: I bought some Sui and a bit of Zcash. I didn't chase Zcash when it surged last year; I started buying during the pullback. In the store of value space, privacy has value. It's a very simple "left-curve" trade (intuitive): it's Bitcoin with privacy. The "right-curve" trade (thoughtful) considers it has quantum-resistant properties. While this may attract government bans, in the future, it provides a very important hedge attribute.
Host: Can you elaborate on why smart contract Layer1s will capture most of crypto's value over time?
Raoul Pal: Layer1s are investment-grade infrastructure layers. Just as the OS market eventually converged to three or four major players, Layer1s will also converge to 3-5 core chains. How to understand Layer1 value? If you unplug Ethereum today, the economic value destroyed is immense: all Layer2s, DeFi, NFTs, RWA would go to zero. ETH's current valuation might even be undervalued. Bitcoin has a single function; its goal is to capture a share of global savings. But the scalability of smart contract infrastructure is infinite.
Host: So which Layer1s will win?
Raoul Pal: ETH has the densest economic value and developer intellectual resources (security, Lindy Effect, etc.), like Microsoft—buying it won't be a big mistake. Solana has proven successful; it's more efficient, faster, cheaper. Sui is very early, but when the market fell 80%, ETH, Solana, and Sui were the only three tokens that maintained economic density. Sui's programmability within a single block, transaction throughput of thousands per second, and finality speed are on a completely different level. You can't evaluate blockchains with traditional "Discounted Cash Flow (DCF)" models because the network's purpose is to provide the cheapest, fastest service; valuing it by how much fee it generates is nonsense. The cheapest, fastest, most programmable chain will ultimately outperform.
Host: Some say DeFi has "died" after many hacks in the past few months. How could traditional finance put money on DeFi that's easily hacked?
Raoul Pal: But this only forces people to develop better products. Just like installing antivirus on computers, hacking is everywhere. Every bank internally has teams dealing with hacks and stolen funds ratios; they just don't publicize it. I predicted back in 2014 that the entire financial system's infrastructure would move to blockchain. Why? Because it's the most efficient way to output energy. The financial system will always migrate to the most profitable, most efficient track. And, DeFi is actually more suitable for machines (AI agents) than humans. Machines don't even need front-end websites; they can rebalance assets and trade instantly across multiple chains, using various stablecoins, with extremely low friction, in milliseconds. They will be the largest user group of DeFi, and we won't even notice these transactions.
Host: Do you think NFTs will gain huge value from the wealth effect mentioned above? My CryptoPunks and XCOPY aren't moving at all; I don't even want to look at them.
Raoul Pal: That's because NFT activity is a function of the crypto economy's prosperity. You have to wait until the overall crypto market reaches trillions of dollars. When ETH breaks above current prices to, say, $5000 or higher, you'll see a massive recovery in NFT activity. Think about it: we're experiencing the biggest turning point in human history: we will no longer be the top intelligence on the planet, and art is the carrier recording the culture of our era. When people make big money in this vast machine economy, they naturally buy "trophy assets" (PANews note: refers to those top-tier assets that, due to extreme scarcity, prime location, or major historical/cultural value, showcase the buyer's social status, bring high psychological satisfaction, and are often "priceless" in the market), just like tech moguls, real estate tycoons, and hedge fund billionaires buy art after getting rich.
Host: How do you plan to configure an NFT portfolio? Is only the top "holy grail" worth buying?
Raoul Pal: I'm actually preparing to launch an NFT fund. Many high-net-worth individuals, family offices, even OGs who made money in crypto but never bought digital art don't know how to buy. Our fund will have two parts: one invests in "holy grail" assets (like Alien Punks, XCOPY, Beeple, valued from hundreds of thousands to tens of millions), which already have proven social consensus. The other part invests in mid-tier but highly convex artists' works. For example, "Die with the most likes," who humorously, even crudely, documents the decline of the American middle class; or German artist Kim Asendorf, at the forefront of AI art. If these artists' works repriced from 20 ETH to 200 ETH (5-10x), and ETH itself may rise 10x in the future, you get an astonishing 100x dual return.
Don't worry about common NFTs because the whole industry is still small; everything will be repriced. Even if you buy a common Punk from the same series, it's a good trade. Additionally, our fund will engage in NFT collateralized lending, earning over 15% yield, reinvesting to support liquidity for the entire art ecosystem.
Host: Is Bitcoin a proxy for investing in AI?
Raoul Pal: You could say so, because AI promotes economic growth, and fiat depreciation from massive debt benefits Bitcoin's function as a digital store of value. However, Layer1 smart contract platforms are a better, more direct bet.
Host: You said everyone is too focused on cycles, while the big picture is so clear: unless you urgently need to, you should never sell.
Raoul Pal: Exactly. In an era with Agents, depreciating fiat, everything going on-chain, why would you sell? If we know the long-term direction of market cap, why sell? This is the pension plan for all humanity. The economic singularity arrives in about 4 years. You have 4 years now to hold as much of these assets as possible; they will carry you through the greatest uncertainty ahead.
Host: Can you prove with data that "buy and hold" beats those trying to time the market?
Raoul Pal: Absolutely. I've built models. If it reaches oversold territory 1-2 standard deviations below the lower bound of the logarithmic trend channel, you buy, then do nothing—the compounding is astonishing. If you try to sell at the top and buy back at the bottom, 99% of people can't do it; it's too difficult. People often chase the highs during rallies. I've been in this industry for 35 years; I don't know anyone who consistently makes big money through short-term trading. Those traders who make big money actually earn asset management fees.
It's proven that those who make the most money in crypto are the ones who "do nothing." Why are the most profitable accounts at major brokerages often "dead accounts (forgotten accounts)"? Retail tries to time highs and lows, not only fails but also expends enormous emotional and mental energy, getting angry or euphoric over daily price fluctuations—this is absolutely the most inefficient use of personal energy. If you have spare productive energy, go study AI, then hold onto your Bitcoin for dear life. If prices become overbought by two standard deviations, you can sell a bit to enjoy life; otherwise, shut up, buy the dip, and patiently hold.
Host: How do people keep faith when portfolios drop 60-80% and stay down for months?
Raoul Pal: I simply don't care. I live on my salary. If I have spare cash and the market is severely oversold, I keep buying. Because my core logic hasn't changed: tomorrow will be more digital than today.
Host: Now AI stocks are soaring; many charts are straight up. Won't this attract people away from boring crypto to buy AI?
Raoul Pal: Your duty is to be a mercenary for your own capital; go where the money is. But I believe crypto offers higher compound returns. Compared to Nasdaq, Bitcoin is currently at a severely oversold position relative to its long-term trend. This means, relative to Nasdaq, you should allocate more to crypto now.
Host: Finally, give us some optimistic hope for 2026-2027, to boost morale.
Raoul Pal: There's a lot of good news. First, banks are entering; stablecoins will explode in growth over the next two years. Regulatory-wise, the 《Clarity Act》 will be signed, allowing almost everyone to start building on blockchain. Macro-wise, the US government has trillions in interest to roll over and pay; they must keep printing, and global liquidity will increase. The business cycle remains strong; more people's income will be recycled into speculative assets. Most importantly, current crypto assets, relative to assets like Nasdaq, are at their cheapest point on the long-term logarithmic uptrend. We've also experienced the longest-lasting, lowest-reading "extreme fear" period historically (Fear & Greed Index below 10), and there's a high probability the Middle East war will be permanently resolved. This is a perfect storm of bullish combinations. I think the probability of this positive bullish outcome is 70%. The remaining 30% downside risk lies mainly in the Middle East war not being resolved, triggering inflation and liquidity tightening, but I haven't seen signs of that yet.
