Raoul Pal said a number out loud on a stage in Singapore, and the number was one hundred trillion dollars. By 2035. He said it twice, I think, because rooms need to hear it twice. By then the second number had already landed — $10 trillion in total crypto market cap inside 18 to 24 months — and underneath both sat the softer line doing the actual work. Ethereum is wildly undervalued relative to what is coming.
I was on a surveillance desk in Lisbon when the clip crossed my feed. 03:40 local time. A burn-rate chart was already open on my second monitor, and the two things did not want to sit next to each other. Pal described a chain carrying 70% to 76% of the world's stablecoins, most of DeFi, the tokenized real-world asset market, the digital art market, and every Layer-2 built on top of it — while the same chain's fee market was paying validators less than it did before its own scaling roadmap shipped. Pulse on the chain, breath in the market. One of those two pulses is faint.
That is the whole story, and almost nobody is telling it.
Context first, because context is what gets stripped out when a clip goes viral.
Raoul Pal runs Real Vision. Before that he ran macro money at Goldman Sachs. He is a public figure, he is a good communicator, and he is not pretending to be neutral. He holds Bitcoin, Solana, Ethereum, Sui and Zcash. He said so. The piece that carried his comments even flagged that the numbers behind his argument deserve a second look — which, credit where it is due, is more editorial spine than most KOL coverage gets.
But here is the thing about a bull market. Bull markets do not reward the person who checks the arithmetic. They reward the person who says the number first, loudest, on a stage. I learned that at 23, in 2017, when I filed 1,200 words on the OmiseGO token sale forty-five minutes after the announcement — before I had read the whitepaper's technical section properly, before I had verified a single claim inside it. My quality score dropped 15% that year. My byline recognition went up. I am not proud of the trade-off. I am just telling you which one the market paid for.
At 27 I tracked whale wallets in real time through the NFT mania and published fifteen threads in a single week. Speed built the audience. Depth would have built the accuracy. I know which one compounds and which one evaporates.
So let me do the boring part that the stage skipped.
Pal's central argument is a systemic-importance argument dressed as a valuation argument, and those are not the same claim. He says that if Ethereum went offline tomorrow, 70% to 76% of stablecoin supply would freeze, most DeFi would halt, tokenized real-world assets would stall, digital art markets would go dark, and every Layer-2 would be affected — a wipeout he sizes at $1 trillion to $1.5 trillion. Fine. That is probably roughly the right order of magnitude for the blast radius.
Now notice what he did not do. He did not show how that blast radius converts into ETH price. TCP/IP is arguably the most systemically important protocol in human commerce. There is no TCP token. System importance is a statement about dependency. It is not a statement about capture. Pal slid from one to the other in a single breath, and the room applauded the second half.
The stress test itself has a methodology problem too. "Assume Ethereum goes offline tomorrow" is a rhetorical device, not an economic model. Ethereum mainnet has run since 2015 without a prolonged halt. Building an investment case on a scenario you yourself believe will never happen is not analysis. It is a mood, wearing analysis as a costume.
Let me define the claim properly, because "wildly undervalued" is doing a lot of unexamined work. To justify a re-rating you need either a rising claim on future cash flows or a rising scarcity premium. Ethereum has neither in a clean form right now. Fee revenue is thin. The burn is thin. The scarcity premium depends on activity that Layer-2s were explicitly designed to absorb off-chain. What you actually have is a rising dependency premium — a growing set of systems that would break without it. Dependencies are real. They are just not priced by the same mechanism as cash flows, and pretending otherwise is how people end up holding a load-bearing asset through a cycle that pays somebody else.
Here is the part that actually keeps me up.
Every Layer-2 that Pal lists as evidence of Ethereum's strength is also a machine for routing fee revenue away from Ethereum. Since the blob upgrade, L2s post their data to L1 for a rounding error. The burn that made ETH deflationary in the 2021-2022 window has thinned to a whisper. The settlement layer got more important and less paid at the same time. That is not a contradiction in the data. That is the design. And it is the single largest blind spot in the bull case as it was delivered on that stage.
I know this shape. I watched it in 2020 during DeFi Summer, on a Lisbon trading desk, when bZx got exploited and I filed my alert late because I had stepped away for a spontaneous after-work thing. Adrenaline does not scale. I rebuilt my workflow around automated alerts after that — not because I got smarter, but because I stopped trusting my own attention. My writing changed the same way. Shorter. Data-forward. Actionable lines over narrative arcs. Sensing the tremor before the earthquake hits only works if the sensor is on, and sensors do not get distracted.
In 2024 I spent most of a quarter modeling ETF creation and redemption flows, mapping BlackRock's entry strategy against on-chain settlement data hour by hour. That work taught me something that has nothing to do with Bitcoin. Institutional capital does not follow narratives. It follows plumbing — custody, settlement finality, regulatory clarity, borrow availability. When the plumbing is ready, the flow arrives whether the story is exciting or not. When the plumbing is not ready, no stage number moves a basis point of it. Pal's $1 trillion stablecoin float is a plumbing claim wearing a narrative costume, and it deserves to be stress-tested like one.
So let me put the sensors on the numbers he put on the stage.
Ten trillion dollars of total crypto market cap in 18 to 24 months. The current aggregate sits in the low-to-mid single-digit trillions depending on how you count. Call it a three-to-four-times move. Not impossible in a liquidity flood. But Pal attached his own precondition: stablecoin supply reaching $1 trillion. It sits around $200 to $300 billion today, depending on which float you count and whether you net out treasury collateral. So the prediction requires stablecoin float to roughly triple or quadruple, inside a window where US and EU stablecoin rules are still being written, and where the largest issuers are fighting over reserve yield rather than chasing float.
Then the $100 trillion by 2035. From a $2.5 to $3.5 trillion base, that is a 30-to-40x. I have an MS in applied mathematics. I have modeled capital flows for ETF desks. I can build you a spreadsheet that produces any terminal value you want, as long as I am allowed to choose the inputs. That is not a forecast. That is a wish with a discount rate attached.
The AI thread is where the argument gets genuinely interesting and genuinely shaky. Pal's read is that crypto and tech underperformed money supply growth because new capital went into AI infrastructure instead, and that rate cuts plus a capital expenditure cycle will redirect that flow back. Running where the liquidity flows fastest — that is the thesis. It is a macro liquidity bet. It has almost nothing to do with Ethereum's technology. It lives or dies on the Fed and on hyperscaler capex, and if AI capex keeps beating expectations, the crypto rotation gets pushed out again. And again. And again. I have watched that "it is coming in six months" clock reset four times since 2022.
Worth naming, too: the framing treats AI and crypto as competitors for the same dollar. That is too clean. DePIN, agent-to-agent payments, verifiable compute — those are real intersections, and they are not zero-sum with AI capex. Pal picked the simpler story because the simpler story travels.
Now the position he did not spend much time defending.
He holds Zcash, on the argument that privacy demand is rising. He also likes stablecoins and tokenized real-world assets, and he wants stablecoin float at a trillion. Those three buckets — privacy coins, stablecoins, tokenized securities — are the three most regulatorily exposed corners of this entire industry, and the stage version of the argument contained zero regulatory risk. Privacy coins keep getting delisted from major venues under AML pressure. Stablecoin issuance is being redrawn by legislation in Washington and Brussels. RWA touches custody, securities law and cross-border settlement. The privacy demand Pal cites and the delisting pressure Zcash faces are the same coin, two faces. Caught in the flash, framed in fact — and the fact here is that the upside he names is bounded by a rulebook he never opened.
There is also the "economic density" line on Sui. I went looking for a definition. There is not one. No metric, no formula, no comparison set. It is a phrase that sounds like a metric. In a surveillance job you learn fast that the words without numbers are the words doing the marketing.
And the alignment question. Pal holds every asset he is bullish on. Public holdings are a form of disclosure, and I will take them over the alternative. But "this asset is wildly undervalued" said by someone holding the asset is a position, not a finding. He was speaking at a Sui event. He runs a media business whose reach scales with the boldness of its calls. None of that makes him wrong. All of it makes him interested.
Here is my contrarian read, and it is the one I would put on a desk note.
The line that should worry bulls most is the one Pal offered as a compliment — that the market has not priced an Ethereum outage. An unpriced tail risk is not a discount. It is an unpriced tail risk. You can read it as cheap insurance that will re-rate upward. You can also read it as the market quietly deciding the dependency is real but the capture is not. Both readings fit the same sentence. Pal picked the first. The tape has been picking the second, which is why ETH keeps trading at a multiple that makes holders angry.
And the real beneficiary of the settlement-layer thesis may not be ETH holders at all. It is the stablecoin issuers collecting reserve yield on a trillion-dollar float. It is the sequencer operators — and I will say plainly, most of those sequencers are single nodes with a company logo on them, whatever the roadmap slides say about decentralized sequencing. It is the custody and compliance vendors. The chain can be load-bearing while the token remains the least-paid participant in its own economy. That is not a bug in the thesis. That is the thesis, misread.
I have seen this exact divergence before, in a different corner of the market. After the fourth halving, miner revenue compressed hard, and hash power kept concentrating into fewer pools. The decentralization story got louder as the underlying distribution got narrower. Nobody was lying. The narrative and the plumbing had simply stopped describing the same object. Ethereum's settlement story and Ethereum's fee capture are drifting the same way.
So what do I actually watch from here? Three numbers, no narrative.
The ETH burn rate, because it is the cleanest available proxy for whether L1 activity is being paid for or merely routed through.
L2 blob fee routing, because if the blob market stays cheap forever, the settlement premium never shows up in the token.
Stablecoin float, because $1 trillion is the load-bearing assumption under the $10 trillion call. If float stalls below $400 billion by mid-2027, the whole clock slips, and the rotation story becomes a permanent six-months-away.
Pal is not wrong that Ethereum sits at the center of something enormous. He is not wrong that the dependency is real. He is unproven — at minimum — that centrality converts into price on the schedule he named.
Seventy-two hours without sleep, zero doubts. I have been that person. It is a great way to break news and a terrible way to size a position. The stage number was $100 trillion. The desk number is the burn rate. One of them is auditable, and it is not the one that got the applause.
