Raoul Pal: four years left before AI "breaks" the economy, keep buying crypto dips

Pal says his entire 2025 strategy was doing almost nothing

Raoul Pal tells the host that he made "two trades the entire year" — buying when the market fell in January/February and again in July — plus adding Zcash. His framework is to buy only when an asset is "two standard deviations oversold versus a log trend," then hold. He argues this compounding approach let his P&L return to all-time highs well before the broader market did, both in 2022 (buying Solana and ETH) and this cycle.

He contrasts this with traders who sell near tops and never manage to re-enter at the same size:

"it's impossible psychologically to do it because you think there's always going to be the other leg lower."

He cites Chris Burniske and Ben Coin as examples of people who were "pretty good on the top" but "kind of missed the bottom," ending up underperforming simply holding.

Pal's SUI bet remains underwater despite years of conviction

Pressed on whether his long-running SUI thesis has actually paid off, Pal admits that excluding tokens granted by the Sui Foundation, his average entry price — buying at 70 cents, around $2.50, and other levels — is roughly $1.70, meaning the position is currently underwater. He says his job is to keep questioning the thesis rather than assume he's right.

His case rests on "economic density per user" — the amount of TVL per active user — which he says sits around $4,000 for SUI, higher than most competitors. He notes that after a roughly 85% drawdown, only Ethereum, Solana and Sui held onto their economic density while other networks like Avalanche "collapsed," which he treats as a signal. Still, he concedes the thesis has not yet proven out and that adoption from agentic finance, gaming (citing Eve Online) and real-world-asset projects needs to materialize.

Pal frames 2030 as an "economic singularity" he wishes weren't coming

Asked what he wishes he were wrong about, Pal says it's his view that an "economic singularity" hits around 2030-2032, as the rise of AI agents — which he says went from zero at the start of the year to "more agents than there are humans on the internet" within ten months — scrambles wages and economic relationships faster than he expected. He says wage growth lagging productivity is the first wave of AI-driven deflation, eventually spreading into goods once AI is pervasive across corporations.

He warns of political backlash as people grow angry about stagnant wages and currency debasement used to fund the AI buildout, alongside a wave of scams enabled by AI-generated voice and video. His conclusion:

"we now have basically four years to unfuck our future because beyond that, I can't see it."

Liquidity has been "looser" than Pal expected, delaying the broad melt-up

Pal says he expected tighter correlation between global liquidity and crypto prices than has actually occurred, blaming liquidity flowing into AI stocks instead, plus a government shutdown and the "Iran war" pulling liquidity out of the system via higher rates, a stronger dollar and oil prices. He says Bitcoin is currently tracking "US narrow liquidity" but remains at a discount to broader liquidity measures, and is "two standard deviations oversold" versus both gold and the Nasdaq on his long-term log trend.

He argues this points to crypto outperforming tech and other assets from here, with smart-contract platforms set to outperform Bitcoin specifically because of stablecoins, tokenized real-world assets, and DeFi all requiring smart contract rails. He floats a rough estimate — unverified in the conversation — that Ethereum's roughly 76% share of the stablecoin ecosystem and large share of DeFi make it "wildly undervalued," suggesting at one point it "should be at least 5x higher."

Digital art as the "most convex trade on Earth"

Pal reveals he is launching a digital art fund called Epoch One, expected to launch in November, built around the idea that blockchain creates genuine digital scarcity. He says roughly 90% of his ETH is now parked in art rather than held as a liquid token, drawing a parallel to buying illiquid real estate in the Cayman Islands years ago as a dollar hedge. He argues that as crypto wealth grows — he floats the crypto economy reaching "a hundred trillion" by 2035 — a share of that wealth will flow into art the way it historically has for the wealthy, calling art "upstream of all wealth."

He is notably bearish on PFP collectibles specifically, saying when asked if favorite JPEGs will come back:

"No."

He argues that market was a cultural, fun-driven moment unlikely to repeat at the same scale, distinguishing it from "serious" fine art collecting.

Zcash, Bitcoin, and the rest of Pal's book

Pal says he bought Zcash around $200 and added more during the flash crash to roughly $250-300 after discussions with people he describes as knowledgeable about the technology, including Barry Silbert, concluding there was no undisclosed "infinite supply" issue. His thesis is simply that some share of global store-of-value demand will want privacy. He says he's unsure how long he'll hold through potential regulatory fights, saying he might take some profit off if it runs significantly while keeping the rest.

On Bitcoin, he says he still holds it as a "collateral asset" akin to digital gold and notes the Iranian Bitcoin seizure coincided with Zcash's price rise as a signal for privacy demand. He also discloses holding "some doge" and "a bit of XRP" as minor leftover positions outside his main thesis.

Living with volatility and public criticism

Pal discusses the psychological difficulty of holding through deep drawdowns, citing Ethereum's post-2021 wedge and gold's long stagnation as examples where holding through ridicule eventually paid off. He acknowledges that when the Zcash flash crash from 650 to 250 happened, he was mocked online, and says public criticism "doesn't get any easier" even after years of experience, though he says he continues sharing his views because he believes he has helped bring people into crypto since 2013.

He closes by promoting his upcoming book, "The Everything Code," out in November, and an AI chatbot trained on his writing and conversations, called rpal.ai, which he estimates is currently "65%, 70%" as useful as he wants it to eventually be.