Raoul Pal says Ethereum is wildly undervalued. In the same breath, he says that if the chain went dark tomorrow, somewhere between $1 trillion and $1.5 trillion would evaporate โ 70% to 76% of all stablecoins frozen mid-settlement, most of DeFi dead on arrival, every Layer-2 hollowed out, the RWA market and the digital art market dragged down with them.
Hold that image. A man who publicly holds Bitcoin, Solana, Ethereum, Sui and Zcash is telling a packed room in Singapore that the most important settlement layer on earth is trading at a discount to its own importance. Then he attaches two numbers that should make any honest analyst flinch: a $10 trillion crypto market cap inside 18 to 24 months, and $100 trillion by 2035.
The clips went viral. The replies filled up with rocket emojis. And almost nobody asked the only question that actually pays: what is the mechanism that converts "Ethereum matters" into "ETH goes up"?
I spent the week pulling that argument apart. Parts of it are genuinely strong. Parts of it collapse the moment you follow where the fees go.
Pal is not a protocol researcher. He's the founder of Real Vision, a former Goldman Sachs hedge fund manager, and one of the most-followed macro voices in this industry. That matters for how you read him. He thinks in liquidity cycles, not in opcodes.
His disclosed book โ BTC, SOL, ETH, SUI, ZEC โ is public, which is more than most influencers offer. He bought ETH and SOL at the 2022 lows and says he hasn't changed his method in two years. The interview ran at Sui Base Camp in Singapore, and the outlet that carried it even slipped in a warning that "the numbers behind his argument deserve further scrutiny." That single line is the most valuable sentence in the entire piece, and I'll tell you why in a minute.
The backdrop is a market that has done nothing for months. Chop. Range. Crypto and tech both underperforming money supply growth, because the marginal dollar is going into AI infrastructure instead of on-chain rails. Pal's entire thesis rests on that flow reversing.
I remember the night this clicked for me. During the Merge watch parties I hosted in Mexico City in 2022, fifty-odd exhausted traders packed into a room to watch epoch transitions projected on a wall. The merge wasn't a technical event for most of them. It was a feeling. That's when I learned the thing Pal is now monetizing: narrative moves price weeks before fundamentals show up, and sometimes fundamentals never show up at all.
So let's do the work. Here's the strongest version of his argument, and here's where the numbers go soft.
The offline stress test is directionally real, but it's a rhetorical device, not a valuation model. His claim that 70%โ76% of stablecoins live on Ethereum is broadly correct โ most issuance still settles on the L1 and its rollups. The $1Tโ$1.5T impairment figure is a defensible order-of-magnitude guess. But Ethereum mainnet has not suffered a prolonged outage since 2015. Using a near-zero-probability scenario to derive a buy recommendation is a confidence trick dressed as risk analysis. Systemic importance and token value capture are two different variables, and Pal slides between them without ever showing the bridge.
That bridge is where the argument actually lives. After EIP-1559, ETH's supply curve became a direct function of L1 blockspace demand. Then EIP-4844 arrived, handed rollups cheap blobs, and the rollups stopped paying meaningful L1 fees. The result is a structural irony nobody on that stage wanted to touch: the more successful Layer-2s become, the less ETH gets burned. Ecosystem growth and the scarcity thesis are pulling in opposite directions. And the DA layer that was supposed to fix the economics is solving a problem most rollups don't have โ I've been saying this for two years and I'll say it again: the overwhelming majority of rollups do not generate enough data to justify dedicated availability infrastructure. They buy blobs because blobs are cheap, not because they need them. Hackers don't need a zero-day when the incentive layer is already broken.
I tested the other half of this thesis the hard way. When Solana stalled in early 2024, I collected 200+ user testimonials โ failed swaps, stuck withdrawals, people watching liquidation prices tick toward them with nothing they could do about it. Uptime matters enormously to users. But here's what the uptime bulls keep missing: user pain does not automatically accrue to token holders. Solana came back, users came back, and SOL's price had far more to do with memecoin volume and narrative than with the outage itself. The same discipline applies to Pal's Ethereum catastrophe. Even if the scenario is real, the trade doesn't follow from it.
Then there's Sui. Pal praises its "economic density" as the reason it holds up in drawdowns. I went looking for a definition. There isn't one. No formula, no metric, no historical series, no benchmark. It's a phrase, not a fundamental. When a thesis rests on a word nobody can measure, you're not analyzing โ you're listening to a pitch.
Verdict first, then the why: the $10T and $100T numbers are not forecasts, they are mood boards. Crypto's total market cap sits in the low-to-mid single-digit trillions depending on the week. Reaching $10 trillion inside 18โ24 months implies a 3โ4x on the entire asset class. Reaching $100 trillion by 2035 implies 30โ50x. Both can only happen if several independent things break right simultaneously: Fed cuts that actually reach risk assets, AI capex rotating out of data centers and into on-chain rails, stablecoin supply hitting $1 trillion, and an AI-agent "economic singularity" arriving on schedule around 2030. Stack four optimistic assumptions and you don't have a forecast โ you have a lottery ticket with a chart attached.
And one more thing that never made the highlight reel. Pal holds five assets and singles out exactly one as "wildly undervalued." The one he names is the one he's most positioned in. That's not a crime โ public disclosure is genuinely better than the alternative โ but it is a conflict of interest wearing a research hat. The outlet's own caveat about "numbers deserving scrutiny" was doing more work than the entire interview.
Here's the angle nobody in that room raised. Pal's stress test, read honestly, is an argument against his own conclusion.
If the market truly has not priced the risk of Ethereum going offline โ his words โ then the market also hasn't priced the risk of value leaking out of the L1 and into the rollups, the stablecoin issuers, and the RWA custodians. Those are the entities that actually capture the fees generated by a settlement layer being systemically important. Circle and Tether earn float on reserves. L2 sequencers earn the spread. ETH holders earn the burn, and the burn is shrinking. The bull case for "Ethereum the network" is quietly a bear case for "ETH the asset."
Then there's the silence around regulation. Pal is bullish on stablecoins, tokenized RWAs and Zcash โ the three most regulator-sensitive corners of this market. Stablecoins live under GENIUS and MiCA. RWA touches securities law, custody and cross-border compliance. Zcash faces delisting pressure across Japanese, Korean and several European venues on AML grounds. He mentioned none of it. "Privacy demand is rising" and "privacy coins are getting delisted" are the same sentence read from opposite ends. A thesis that only works if regulators stay friendly is not a thesis โ it's a bet on someone else's patience.
Watch three numbers over the next two quarters: ETH's net issuance, L1 fee revenue net of blob subsidies, and total stablecoin supply. If issuance stays positive while stablecoin supply climbs, the ecosystem thesis is right and the token thesis is wrong โ and that gap is where the next eighteen months get decided. The chain doesn't have to go offline to hurt you. It just has to keep working exactly as designed.