RV Pro Deep Dive: Ethereum, Solana, Sui

First, the market

Ethereum hasn't had a stand-alone deep dive from us since July 2022, before the Merge, which is far too long for a core position. Solana was last treated properly in The Solana Trilogy in July 2024.

And Sui, the one I know most of you are anxious about, gets the most careful read of the three. The news you can get anywhere, but this piece is the analysis. Let's get into it.

Before the chains, the context that governs them all: Crypto doesn't trade on headline liquidity; it trades on excess liquidity - what is left over after the real economy absorbs its share.

Headline liquidity made new highs through 2025, but excess went negative in August 2025, and that was the top. The whole complex then fell in proportion to its beta... BTC 36% from its high, ETH 50%, SOL 66%, SUI 86%.

As liquidity slowed, so did crypto. That's the story of the past year, and it needs no more explanation than that...

Excess broad liquidity leads crypto by roughly 90 days: US excess broad liquidity (total liquidity money flow minus nominal GDP) against total crypto market cap, log scale. Excess went negative into the 2025 top and has been fighting its way back toward the zero line since.

The same dial now says the regime is ending. Excess sits at the zero line, with the June official print at -0.2 points, and the chart is in this report so you can watch it with us. The ignition is two consecutive monthly prints above zero, most plausibly in the October-November window.

Generally speaking, even in midterm election years, the market tends to have a very strong period after the election, though expect some volatility in the middle. And, as we've explained in our research, we expect this cycle to run longer than most imagine... the recent Bessent put has only strengthened that case.

The mechanics beneath it: liquidity leads crypto by roughly 50 to 90 days, the accumulation zone runs August to October, and the June lows are the line the whole map stands on: BTC $58,000, ETH $1,560, SOL $61, SUI $0.66. Alt rotation typically follows ignition by around 60 days, confirmed when Bitcoin dominance closes a week below 55%...

Bitcoin dominance, weekly: the multi-year converging wedge, currently ~60%. A weekly close below 55% is the rotation confirmation.

With that on the table once, we can look at each chain for what it actually is, rather than telling the same liquidity story three times.

Ethereum: the settlement layer grew up

The last time Ethereum had a report of its own was July 2022, written about a chain that still ran on proof-of-work. The position went on the book the month before, at $1,426 in June 2022, almost exactly the bear market low.

Everything about the network has changed since. The thesis, oddly, has barely moved.

Let's begin with what happened while we weren't writing about it. The Merge took Ethereum to proof-of-stake in September 2022.

Shanghai unlocked staking withdrawals in April 2023, removing the last structural fear discount. Dencun landed in March 2024 and collapsed L2 transaction costs, which is what turned the rollup thesis from a whitepaper into Base, Arbitrum and Optimism carrying roughly $28bn of value between them today. And the spot ETFs arrived in mid-2024, giving TradFi its first clean pipe into the asset.

But the number that defines modern Ethereum is none of those. It's $148bn... the stablecoin supply now sitting on the L1. Ethereum's own market cap is around $296bn, which means the dollar economy living on Ethereum is now half the size of Ethereum itself. Add $49bn of L1 TVL and the L2 stack on top, and you get the cleanest description of what this asset became while everyone argued about narratives: the settlement and collateral layer for the on-chain dollar system.

And that role is compounding, not static. The tokenization wave... funds, treasuries, private credit... issues on many chains but settles back against the same dollar collateral, and the first generation of agent-to-agent payment rails is being built on top of those exact-same dollars.

ETH did $9.7bn of DEX volume in the past week alone, up 86% week on week as the complex turned. Trading activity across the stack is turning back up.

The valuation matrix we built for Sui in the March GMI reads ETH very distinctly. TVL per daily active address on DeFiLlama's counts and measured at the base layer deliberately (the L2s carry the activity, the L1 keeps the settlement... that division of labor is the thesis) is roughly $82,000, against about $2,100 on Solana and $3,200 on Sui. One health warning that applies to every cross-chain number in this piece: address counts differ widely between trackers, so we use one source consistently and trust the trends more than the absolute levels.

Transactions per active address run at about four a day. In plain English: fewer users, vastly deeper capital, doing deliberate things.

Ethereum is where the money lives; the other chains are where the activity lives. Both are viable economies. They're just different ones, and pricing them off the same metric is how people get this sector wrong.

The bear case deserves its paragraph too, because it's the one you'll hear most often. Dencun pushed activity out to the L2s and cut L1 fee revenue hard, and the market read that as value leaking out of the asset... ETH the token earning less while Ethereum the ecosystem grows. It's a real argument, and it explains a great deal of the two years of underperformance.

Our answer is the same one we gave on Sui in March: fees are friction, and what you own here is the settlement and collateral layer, not a fee take. $148bn of dollars and the deepest capital base in the sector chose this chain to live on, and they pay for that security in ways fee dashboards don't capture. If the market keeps pricing ETH as a fee business, it will keep being wrong at the turns... in both directions.

In its simplest form: ETH's value is the throughput on ETH plus the businesses built on top of it. Stablecoins, $148bn of them. DeFi, $49bn of TVL on the L1 alone.

The L2s... Base, Arbitrum and Optimism carry roughly $28bn between them, whole chains with real economic value of their own, added on top of Ethereum because Ethereum is what they settle to. NFTs. And now real-world assets arriving through the tokenization wave.

Not all of that value accrues to the ETH token, and I won't pretend it does; the L2s keep most of their own economics. But every one of those businesses requires ETH underneath, and no other network has an economy remotely this size built on top of it.

Think of Ethereum the way you think of the US economy. America's legal structure and culture are what allow vast business value to be built on top of it, and you don't value the US economy by its tax intake. You value it on GDP... the total value transacted... and even more on the value of all the businesses built on top.

Ethereum's fees are its tax take. The stack above is its GDP and its corporate sector. This is the part that is being incorrectly priced.

So, how is it moving? ETH is around $2,500 as I write this, down 50% from the August 2025 high, having peaked at $4,954... a long way short of what this cycle was supposed to deliver for ETH, ours included, and the liquidity dial up top is the whole reason why.

The June low was $1,560, and the Stink Bid we set filled at $1,650 in June, near the low. Price has reclaimed the 200-day moving average near $2,000 in a sharp August recovery that has lifted the whole complex 20-30% in a matter of weeks. The alts get their proper turn when dominance breaks, not before.

The chart that matters most here isn't ETH/USD, it's ETH/BTC: roughly 0.031 now against 0.064 at the August 2025 peak. Halved in a year, sitting on decade-long support, with weekly exhaustion buy signals printed on the ratio itself.

A move through 0.035 is the confirmation that rotation has begun. Until then, ETH strength is BTC strength with lag...

ETH/BTC weekly: the decade-long downtrend from 2017, the falling wedge at long-term support, currently 0.0317. Through 0.035 is the rotation confirmation.

ETH/USD weekly: the multi-year ascending triangle... rising support from the 2020 lows, the $4,700-4,800 ceiling tested three times.

ETH/USD daily: the June retest of rising support near $1,500-1,560, the recovery through $2,500, the big level overhead at $4,800.

$2500 has always been an attractor, and we need to see ETH go through it convincingly for bigger confirmation.

Positioning. Core hold, high conviction, and we add on weakness. The add zone is $2,000 to $2,100, the 200-day retest.

$1,875 to $1,900 is major support beneath it and would be a gift. I'd be surprised to see ETH trade back below the June low at $1,560, and if it does, we'd think hard about reassessing where we are... starting with whether the liquidity picture has changed or something else has gone wrong.

We're not putting cycle targets on this. The long-term view is unchanged: we still maintain that ETH will be worth over $20,000 in the coming several years, as the entire financial system comes on-chain.

Solana: the full round trip

Solana has received the most coverage of the three: the October 2023 In Focus piece, written when it was the most hated major asset in the world (at a fraction of today's price), and The Solana Trilogy in July 2024. Since then, it has done the entire round trip. $293 at the January 2025 top, $61 at the June 2026 low, about $101 now.

The original position went on in June 2022 at $26.74, so even after a 66% drawdown from the peak, the line is still up roughly 3.8x. We added again via the Stink Bid at $65 on the 5th of June, four dollars off the bottom. What did the round trip actually teach us?

That the thing we bought is real and the way the market prices it is manic.

The network kept building straight through the drawdown: stablecoins on Solana now stand at $15.9bn and are still growing, with USDT supply up 9% in the past 30 days; daily active addresses run around 2.7 million, more than four times Ethereum's L1. Solana processed a record 4.2 billion transactions in July, up 91% from last December... roughly 135 million a day, a count that includes validator vote messages (the network's own consensus chatter), so treat the growth rate as the signal rather than the level. Strip the votes out and real user activity still set its own record in August: 1.3 billion non-vote transactions in the busiest week in the chain's history. Either way it tells you two things at once: throughput is genuinely the product, and a meaningful share of that count is machines trading with machines. Both are true, which is why we read the activity through stablecoins and users rather than raw transaction counts.

The newest driver of that machine activity isn't meme coins; it's agents. Solana and Base are currently trading the lead in x402 transactions, the emerging standard for AI agents paying each other for services. The absolute volumes are still tiny and the lead changes hands week to week, which is exactly what an embryonic standard looks like.

The agent-to-agent commerce we've been writing about as the endgame for these networks is no longer a forecast; it's showing up in the transaction mix, and agents go where settlement is fastest and cheapest. Watch that series as an early, noisy read on the biggest demand source of the next cycle.

On the matrix, Solana is the mirror image of Ethereum: capital density of about $2,100 per active address against ETH's $82,000, stablecoin depth of $15.9bn against ETH's $148bn, but activity and user counts ETH can't touch. At roughly $56bn of market cap against $5.6bn of TVL, Solana trades at around 10x TVL... the richest multiple of our three majors, against roughly 6x for ETH and 7x for Sui. A caveat on these multiples: TVL itself moves with token prices, so the ratio breathes with the market; the level is the information, not the wiggle.

And turn the same numbers into a turnover ratio, and you see what the premium is actually buying: Solana did over 3x its entire TVL in DEX volume in a single week, against about 20% for Ethereum. Capital doesn't sit still on Solana; it spins.

That premium is the market paying up for throughput leadership and for the ETF flows now arriving. Premiums are fine when you know you own one and size for it, because they re-rate hardest in both directions... 2026 demonstrated the downside of that arithmetic in full. Solana's economy is also levered to trading activity in a way Ethereum's isn't: a large share of its fees, its transaction count and its stablecoin velocity is markets trading markets, meme coins included.

In a risk-on regime, that's a feature, because activity compounds and the fee take compounds with it. In a risk-off regime, it drains faster than user counts suggest, which is exactly what the 2026 drawdown showed. So, when we say SOL is a high-beta major, we don't just mean the token... the underlying economy is cyclical too, and the 10x multiple is being paid on cyclical revenue.

None of that changes the thesis. It changes how much of the book it should be, which is why it's sized as a core line and not the whole crypto allocation.

Over the past twelve months, SOL is still down more than 50%, and SOL/BTC gave back roughly a third over the same stretch. High beta does what high beta does, and the dial-up top explains all of it. The same mechanism runs in reverse: when dominance breaks, SOL has historically been the first big beneficiary of the rotation flow.

Solana stablecoin supply, weekly since 2024 (DeFiLlama): from under $2bn to $16bn, still grinding higher through the entire 2026 drawdown.

SOL has bounced hardest of our three, back to around $101 and already pressing into the first resistance zone, with DEX volumes running at $17.5bn, up 59% week on week, so it's obviously a bit stretched short term. Resistance comes in at $100 to $110 and then $120, but we already own it, so there's nothing to chase here... the plan is for weakness, not strength...

SOL/USD weekly: the giant ascending triangle... rising support from the 2023 lows, the $250-265 ceiling, price back at $101 after holding the trendline in June.

SOL/USD daily: the June trendline low, the recovery to the $101 breakdown retest, the $265 ceiling above.

Like $2500 in ETH, $100 has always been a price attractor, and we need to also see it clear this level convincingly.

Positioning. Core hold, high conviction, add on weakness. The add zone is $85 to $90.

Support below sits at $80, then the $61 to $65 area where the June low and our fill both live. It's the same as ETH: I'd be surprised to see SOL back below the June low at $61, and if that happened we'd be reassessing where we are and whether something has gone wrong with the liquidity read. Above the June low, pullbacks into the add zone are for accumulating, not agonising.

Sui: the honest read

Let's do Sui properly, because I know exactly where most of you are; looking at a token that went from $5.35 at the peak to about $0.75 now, an 86% drawdown, the deepest of any major we hold, and wondering whether we're seeing something real or refusing to admit a mistake. It's an important question, so let's answer it properly.

First, the record, with no dressing up. We entered at $1.76 in September 2024, added at $2.20 in April 2025, added again at $0.97 in February 2026, and the Stink Bid filled at $0.75 in June.

At about $0.75, the two earlier lines are down 57% and 66%, the February add is down 22%, and the June fill is flat. The record also includes this: the March matrix was written from a $0.91 trough, and the market then took it to $0.66 in June, 27% lower.

The price of the token has been absolute dogshit: The question that matters is whether the network has been dogshit, too.

The March GMI built a four-metric valuation matrix for exactly this question: TVL per active address, transactions per active address, stablecoin depth, and applications per developer. Four numbers chosen to reduce the price effect and resist easy gaming. Here is the refresh, five months on, same sources...

横向滑动查看完整图表

So, what does the table tell us? Two of the four core metrics have deteriorated since March. The worst of them is the one we told you to watch most closely.

We set the stablecoin depth trigger at roughly $1.1bn, about 2x TVL, and instead of climbing toward it, depth sits between roughly $420m and $480m depending on whose count you use... down 15-25% from March, at about 1x TVL against 3x on Ethereum and nearly 3x on Solana (and yes, deposited stablecoins are themselves part of TVL, so the two overlap rather than being independent checks). It has ticked up in recent weeks alongside the rally, but I won't reframe five months of the wrong direction on two good weeks: the leading indicator we picked went backwards.

Full stop.

Sui stablecoin depth, weekly since 2024 (DeFiLlama): the $1.2bn peak in July 2025, the collapse through the drawdown, latest $431m against the $600m checkpoint and $1.1bn trigger.

What held is capital density: TVL per active address was about $2,120 at the February trough and, on our own dashboard, sits essentially flat at ~$2,070 now, having held through the entire drawdown (DeFiLlama's counting shows it higher, near $3,230; we use our own basis for continuity with March). TVL fell while active addresses fell faster: each remaining user is at least as deep in, and there are fewer of them.

The bullish read is commitment concentrating; the bearish read is that what remains is yield capital that would also leave in a worse regime. The staking picture leans toward the first... the bulk of supply remains staked despite the easiest exit door in the majors... but it's a lean, not proof.

The developer layer needs a caveat the size of a house. Electric Capital has sector-wide monthly active developers down 56% over the year as AI absorbs engineering talent, so every chain's denominator collapsed at once. Sui sits at 633 monthly actives with 147 live DeFi protocols against the 300+ trigger we set in March... treat that as directionally below target rather than a precise miss, since the counting basis may differ from March's.

In March we cited 14.3 protocols per 100 developers, double Solana and triple Ethereum. On today's like-for-like basis, all three chains sit between 23 and 26 per 100, with Sui marginally the lowest, though both sides of that ratio changed measurement basis since March, so it's directional at best. Some of it may still be real erosion, and we'll treat it as real until the data proves otherwise.

What the price doesn't show...

The price chart doesn't show you what has changed most since March: the institutional build-out on Sui has accelerated straight through the drawdown. Securitize has launched its first tokenized fund on the chain. Tether's Hadron platform, built to put stocks, bonds and commodities on-chain, is live on Sui. tZERO is bringing seven years of regulated digital-securities infrastructure across.

And the stablecoin strategy behind the depth trigger is deliberate. Stablecoin transfer fees went to zero, and $65bn moved across the chain in a single five-day stretch in June... more than $1 trillion of stablecoin volume has now crossed Sui in the past year, against roughly half a billion that stays deposited. The rails are clearly being used; the dollars don't yet live there, and closing that gap is exactly what the depth trigger measures. That's what the native yield-bearing dollar launched in March is for. None of this shows up in TVL yet, and all of it aims at the one metric we care most about... deposited dollars.

Let's step above the metrics.

Platform markets don't produce twenty winners; they produce four or five. In the smart-contract world we have Solana and Ethereum, and outside of Sui I can't see the third. There will be specialist cases... NEAR and others... but the general-purpose winners' circle is small, and most of it is already visible.

So, what is Sui actually built for? The machine age. This is the agentic economy argument we laid out in The Intelligence Flywheel back in February, and it is the biggest bull factor here, bigger than the RWA wave.

A single Sui transaction can carry up to 1,024 composed operations, executed atomically, with finality at roughly 300 milliseconds and heading for 200. There's nothing like it in the majors for an agent that needs to transact, verify, and transact again at machine speed.

And in July, Sui's programmable tunnels... Lightning-style channels with programmability added... peaked at over six million transactions per second in a live agent-to-agent experiment. That's not direct on-chain throughput, to be clear, but that isn't the point.

As Kostas Chalkias, the Mysten co-founder who ran it, put it, "A company's trading agent could play chess or poker against another company's agent millions of times without touching the base chain... Right now, there are only four or five proven product-market fits in crypto: stablecoins, DeFi, payments, prediction markets. I think programmable tunnels just opened the door to a fifth."

Privacy matters just as much, and not only for agents; the entire financial system cannot operate in public, and no company can run payroll, treasury, credit, or a trading book on a fully transparent ledger, which is why so little of finance has moved on-chain so far. Privacy is of utmost importance, and Sui is building it in at the protocol level rather than bolting it on; identity privacy is live today, confidential transactions are on the current roadmap, and NIST-approved quantum-resistant signatures are rolling out at the protocol level alongside them. Sui is specified for a world that doesn't fully exist yet, which is precisely the point of owning it early.

RWAs and stablecoins are the here-and-now use case. Agents are the one that re-rates it.

And yes, Sui hasn't picked up massively yet. It's still early in the game, and the environment could hardly have been worse for proving it: token prices down 86%, the AI boom pulling generalized activity and attention away from crypto altogether, and a very lacklustre liquidity backdrop that has to fund both AI and crypto at once.

Our job is to be early. That is the whole job.

Regardless of all of that, I don't see a world in which Sui is not entirely re-rated for its capabilities in the new world we're moving into. I think it's a matter of time. Not because of my board seat, and not because it's one of my biggest bags. It's because it just seems so fucking obvious to me.

And look, I know that attitude may be hard for some of you to sit with while you stare at this drawdown. I get it. Alas, I remain convinced of Sui’s integral part to play in our emerging world.

That doesn't mean Solana and Ethereum won't do stunningly well... they should... and Sui may not even outperform them. But if given our own pricing metrics, over time, it really should. The table above is exactly how we'll know whether we're wrong or right.

The cycle mechanics explain the path: Sui is the most beta thing we own, and the dial-up top predicts an 86% drawdown in exactly this regime... the drawdown is not, by itself, information about the chain. The real information is in the table above and, as we've seen, it's genuinely mixed: parts have held up impressively, one important part went the wrong way, and none of it looks like a chain that's failing.

Part of the stablecoin decline is the regime too... deposited dollars leave the riskiest venues first when excess is negative. But we won't lean on that: the test is whether depth returns when excess turns positive, and if it doesn't, it was never the regime.

And beta cuts both ways. If the ignition fires, the same mechanism that took Sui down 86% is the one that moves it furthest on the way back. So here is the structure, with dates and dials, because any forecast you can't falsify is just a story.

The base case: the June low at $0.66 printed on 2023 support with weekly exhaustion signals, part of the same buy alignment across the majors. The accumulation zone and ignition dial up top govern here, with alt rotation following ignition by roughly 60 days if the sequence works.

First price confirmation is a reclaim of $1.00. First fundamental confirmation is stablecoin depth turning back up through $600m... not the target, just proof of direction.

The March trigger of roughly $1.1bn stands, and it is further away today than when we set it. But it isn't uncharted territory... depth peaked above $1.2bn in July 2025, so what we're asking for is a reclaim of where the dollars have already been, not somewhere they've never gone. Price can lie for months; deposited dollars don't.

So, how is it trading? Sui lagged the August bounce and has already handed part of it back, down 11% this week to about $0.75 and back inside the accumulation zone... exactly what you'd expect from the highest-beta major while Bitcoin dominance is unbroken. DEX volume still more than doubled week on week to $377m from a low base... a rounding error next to the big two, but on the turnover measure that's about 80% of Sui's entire TVL traded in a week, so what capital remains is being used.

SUI/USD weekly with DeMark 9-13 counts: the falling wedge into rising support from the 2023 lows, weekly buy 13s printed at the June low, price back in the accumulation zone at $0.75.

Sui bounced exactly where it should, exactly on the trend line with the monthly and weekly DeMark signals. Although it's an early-stage chart, I think the pattern is much bigger than people can yet perceive, but let's see...

I'd be surprised to see Sui back below the June low at $0.66, and a weekly close down there would have us reassessing the whole setup, starting with the liquidity read. The other thing we're watching is stablecoin depth. If it's still shrinking by Q1 2027, after the regime has had time to turn, then the fundamental thesis has a real problem regardless of price, and we'd rather reduce than re-tell the story.

Why Q1? Because the regime turn itself is dated October-November, and two quarters is a fair window for deposits to respond... and if the ignition never prints, the timetable resets with it.

Positioning. Hold on conviction. For those underweight, the accumulation zone is $0.70 to $0.75, where the Stink Bid filled.

Critical support is $0.64 to $0.66. Resistance is $1.00, which is also the first confirmation.

None of this is a promise that Sui rescues the average entry by Christmas. It is the honest shape of the bet: the regime that crushed it is, on our map, ending; the chain's own data is intact in places and wounded in one place that matters; and the levels and dates above tell us when to reassess.

One map, three positions

Step back, and the three sections are one picture. Platform markets produce four or five winners, and we believe we own the core of the smart-contract winners' circle: Ethereum the deep-capital settlement layer, Solana the high-throughput activity layer, Sui the highest-beta bet on the next architecture.

All three sit in the same accumulation zone waiting on the same dial, and the June lows up top are the line. And for those of you who came to this piece for the Sui answer: you now have all of it, both the wound and the case, and the dials that decide between them.

Above the June lows, we accumulate patiently into weakness. Below them, we'd be reassessing where we are, exactly as laid out above.

Next month we rotate into the next block of the book. Bring the hard questions... this format exists for them.

Raoul Pal

Raoul Pal has been publishing Global Macro Investor since January 2005 to provide original, high quality, quantifiable, and easily readable research for the global macro investment community, hedge funds, family offices, pension funds, and sovereign wealth funds. It draws on his considerable thirty-one years of experience in advising hedge funds and managing a global macro hedge fund. Global Macro Investor has one of the very best, proven track records of any newsletter in the industry, producing extremely positive returns since inception: www.globalmacroinvestor.com.

Raoul retired from managing client money at the age of 36 in 2004 and now lives in the tiny Caribbean island of Little Cayman in the Cayman Islands.

He is also the founder and CEO of Real Vision, which is a digital media group: www.realvision.com.

Previously he co-managed the GLG Global Macro Fund in London for GLG Partners, one of the largest hedge fund groups in the world. Raoul moved to GLG from Goldman Sachs where he co-managed the hedge fund sales business in Equities and Equity Derivatives in Europe. In this role, Raoul established strong relationships with many of the world’s pre-eminent hedge funds, learning from their styles and experiences. Other stop-off points on the way were NatWest Markets and HSBC, although he began his career by training traders in technical analysis.