Delphi Digital researcher Marcus says the key question in crypto valuation is rarely whether a protocol looks cheap on revenue. The real question is whether that revenue can hold. If a project can cover its current market cap with less than two years of revenue and buyers still hesitate, the issue is usually not the number itself. It is the market’s belief that the number will last.
That is what xRev, or market capitalization divided by annualized revenue, is actually measuring in his framework. Marcus argues that xRev is not a pure cheapness metric. It is a durability metric. He uses PUMP and AERO to make that case. Based on recent revenue, both traded at low single-digit multiples, with PUMP at 2.3x and AERO at 3.5x. Yet over the past 30 days, PUMP rose 87% while AERO fell 14.5%.
The screen looks similar. The price action does not.
Marcus writes that in June this year, the market gave PUMP only a 1.3x valuation, which implied that traders did not believe the protocol could sustain then-current revenue for more than 16 months. By July, that skepticism started to break. Most of the move since then came from rerating. AERO moved the other way. Since its price peak in December 2024, its multiple has nearly doubled, not because confidence improved, but because revenue fell faster than the market repriced the token.
Under that framework, the trade is not simply “buy the lowest multiple.” It is “buy the multiple that is about to stop being doubted.” Once a revenue stream that had been dismissed as fragile proves it can persist, a rerating can drive most of the upside even if market cap rises faster than revenue.
What xRev is actually telling the market
Marcus defines xRev in simple terms: market cap divided by annualized revenue. If xRev is 1.0, one year of protocol revenue equals the entire market value. If it is below 1.0, the payback period is under a year.
Numbers like that often trigger the same reaction: the market must be wrong. Marcus says that reading misses the point. A very low xRev often means the market is applying a steep durability discount to the revenue stream. In effect, the market is saying that current revenue is temporary, that it will fall, and that it will not recover.
That is why a very low xRev is not, on its own, a buy signal. It is better read as a market expression of doubt. The alpha lies in figuring out whether that doubt is justified.
Before getting into the examples, Marcus separates tokens into two buckets based on where they begin on the xRev spectrum, because the starting point shapes the kind of opportunity that can emerge later.
Type A tokens: born cheap
These are new protocols that find product-market fit in high-fee categories and suddenly generate strong revenue, but the market does not trust that revenue to persist. Their xRev starts near 1x or even below 1x. High revenue. Small market cap. Deep skepticism. In Marcus’ framing, the market cap is low for one reason: the market has not bought into the revenue story yet. That is what makes these names candidates for a belief-driven rerating.
Type B tokens: born expensive
These are the opposite. The market prices them from day one as future revenue giants, so xRev starts high and optimism is already embedded in price. There is no obvious “skepticism discount” to exploit. The trade is mainly about whether the project can deliver on expectations that were paid for upfront.
Marcus says AERO clearly launched as a Type A token. PUMP clearly launched as Type B. From there, the paths split.
PUMP: a Type B project starting to earn back its old premium
PUMP is Marcus’ clearest Type B case. The token launched after a $1 billion raise, implying a fully diluted valuation of $4 billion. That represented more than nine years of protocol revenue at the time. On circulating market cap, PUMP opened at 4.5x xRev.
In his reading, the market paid a premium for belief before the business had fully proven itself. Over the next year, that premium was gradually taken back. That process was not hard to understand in context. Meme coin trading volume is cyclical, rivals were competing for order flow, and there was no certainty that the platform could defend market share.
The xRev chart shows that early confidence bleeding out of the name over roughly 11 months, even as the protocol was still producing more than $200 million in quarterly gross revenue.
Marcus highlights two points on the chart.
First, the April 2026 restructuring caused a mechanical drop in xRev rather than a simple mood shift. Pump burned all tokens it had previously bought back, worth about $370 million, equal to 36% of circulating supply at the time. It then committed 50% of net revenue to an irrevocable buyback-and-burn contract that will run for one year.
Because the burn removed about one-third of circulating market cap overnight, xRev had to fall immediately by construction. What stood out to Marcus was what happened next. The market kept applying a discount anyway. After the burn, xRev slid from around 1.7x to 1.29x by June 6. At that level, PUMP’s market cap was worth less than 16 months of revenue, and half of that revenue had already been contractually set aside to buy and burn PUMP. Marcus describes that point as peak skepticism.
Second, the rerating only began once the revenue story itself started to change. During the week of Aug. 3 to Aug. 9, Pump posted more than $10 million in weekly fees for the first time. Of that total, $5.02 million was used to buy and burn 2.15 billion PUMP.
Revenue over the past 30 days reached $37.9 million, which annualizes to roughly $460 million. Revenue over the past 12 months was $457 million. In other words, the current run rate is almost identical to the last year’s level, while cumulative protocol revenue has already passed $1.2 billion. The market spent a year treating this business as a temporary spike. Marcus says the latest data show that the revenue base is still there and has started to bend upward again.
The result is a rerating. PUMP’s xRev expanded about 80% from the June low of 1.29x to 2.3x. Over the same 30-day period, the token rose 87%, and circulating market cap reached $1.07 billion. Marcus’ point is that this move did not require explosive revenue growth. Most of the gain came from the market changing its mind about the durability of revenue. If investors believe roughly 2.3 years of revenue can equal today’s market cap, they will pay more for the same revenue stream.
He adds an important caveat: his preferred lens is circulating market cap. On a fully diluted basis, PUMP’s FDV stands at $2.3 billion, which puts xRev closer to 5x. August unlocks are also narrowing the gap between circulating and fully diluted figures. He also notes that, regardless of the buyback structure, a meme coin issuance platform remains a cyclical revenue business. If weekly fees fall back below the roughly $5 million to $6 million range that held through the summer consolidation, this rerating could reverse as quickly as it formed.
Aerodrome: when the low multiple is telling the truth
AERO is the opposite case in Marcus’ framework. On paper, it looks attractive. The token carries a $404 million market cap against $116 million in trailing 12-month annualized revenue, which puts xRev at only 3.5x. Unlike most protocols, 100% of AERO revenue flows to veAERO lockers. There is no value leakage in that setup, and 54% of supply is locked.
That sounds like a reasonable long. Marcus says the xRev chart argues otherwise, at least for now.
In early 2024, AERO was the cleanest Type A setup and one of the strongest bullish examples for this entire framework. It was a new protocol that had just found product-market fit, and Base’s rapid growth helped push revenue up much faster than market confidence. At one point, the market valued it at less than 1x revenue, or less than one year of revenue. In that phase, the market’s skepticism was wrong.
Revenue kept growing, confidence followed, and by the time AERO peaked in December 2024, xRev had expanded to 3.4x and market cap had reached $1.5 billion. The real opportunity, Marcus argues, was to buy when AERO traded below 1x xRev and hold until the market changed its mind.
What followed was different. Since that peak, the token is down 73%, but xRev has nearly doubled to 8.8x. Marcus says there is only one way to get that combination: revenue is falling faster than price.
AERO’s annualized current run-rate revenue has dropped from about $443 million at the top to roughly $46 million now. Quarterly total protocol revenue peaked at $106 million in the fourth quarter of 2024 and fell to $29.3 million by the second quarter of 2026, a 72% decline. This is the mirror image of the PUMP case.
In his view, AERO’s price weakness does not reflect irrational underpricing of a highly profitable protocol. The market is repricing a deteriorating revenue curve, and it is not even keeping up with how quickly fundamentals are worsening. Over the last 30 days, AERO fell 14.5%, while PUMP climbed 87%. Marcus points to that spread as a direct example of the durability discount at work.
He does note several possible catalysts that could change the picture for AERO, including a merger with Velodrome, the predictive emissions upgrade launched in July, and a Binance listing. Any of those could alter the revenue trend. If the run rate turns back up, the same rerating mechanism that pushed PUMP higher — and that lifted AERO itself in 2024 — could return.
But that is exactly the point, Marcus says. Buying AERO today is not really a bet that the trailing multiple is cheap. It is a bet that the revenue trend will change.
What counts as a fair xRev multiple
xRev is just a ratio, so a rising or falling multiple means little by itself. Marcus breaks every move into two parts: in log terms, the change in xRev equals the change in market cap minus the change in revenue.
That means multiple compression can come from price falling faster than revenue, or from revenue growing faster than price. Multiple expansion can come from price rising faster than revenue, or from revenue shrinking while price declines more slowly. The same chart shape can reflect opposite trade setups.
That is why he says the first question should always be: which leg is driving the move? Once the ratio is decomposed that way, “fair value” becomes more practical.
When both market cap and revenue move around but xRev stays broadly stable over a long period, Marcus treats that as the token trading at its fair multiple. In that regime, the market is repricing the token one-for-one with the changes in fundamentals. The multiple that holds through that stretch is effectively the market’s estimate of revenue durability.
He divides AERO’s history into three stages.
Stage one was the belief rerating in early 2024. Market cap rose 58x. Revenue rose 26x. Both exploded higher, but price moved faster because the market began to believe.
Stage two was the fair-value band, which Marcus describes as the most easily overlooked part of the chart. From May 2024 through March 2025, xRev stayed between 1.7x and 3.0x for 11 straight months, with a median of 2.2x. During that span, market cap ranged from $223 million to $1.57 billion and annualized revenue moved from $141 million to $491 million. Those are huge swings — roughly 4x to 7x changes in the underlying figures — yet xRev barely moved. Marcus says that showed the market had settled on a consensus that a dominant Base DEX was worth about two to three years of revenue.
Stage three broke that range. Since April 2025, AERO’s market cap has barely changed, moving from $350 million to $404 million, while annualized revenue has fallen 67%. So xRev nearly doubled, with almost no help from price on the numerator side.
PUMP looks very different under the same lens. Its chart is almost a mirror image of AERO’s, but with one critical distinction. PUMP started from the classic Type B position, opening at 4.5x xRev because confidence had already been priced in.
Its first stage was therefore the steady removal of that prepaid confidence. In that stretch, market cap fell 42% while revenue declined only 13%. Marcus says the business itself did not suffer a structural break. The market was simply pulling back the premium it had paid at launch.
The shaded region on his chart marks the next roughly six months, when xRev traded between 2.6x and 4.4x. It looked stable on the surface, but it was not. By February, that range broke again due entirely to price-side selling, even as protocol revenue posted its best quarter on record. Then the April restructuring mechanically reset the numerator. Pump burned all previously repurchased tokens, worth about $370 million, equal to 36% of circulating supply.
Because market cap is price times circulating supply, removing one-third of supply without any token sales cuts circulating market cap by one-third overnight. That pushed xRev from around 3x to around 1.7x immediately. Marcus says that part of the move contained no new information about market belief. The two months after that were more like capitulation. Both market cap and revenue moved only slowly, and xRev eventually touched a low of 1.29x.
He describes that moment as the point where a Type B token fully transformed into a Type A structure: high and already validated revenue, a compressed market cap, and maximum skepticism. The third stage shows what happens after that skepticism breaks. Market cap rose 87%, and revenue rose 43%. Marcus says that is the same quadrant AERO occupied in early 2024.
Is revenue holding up or rolling over?
To separate stable revenue from decaying revenue more clearly, Marcus suggests calculating xRev in two ways.
The first is trailing 12-month xRev, which divides market cap by revenue accumulated over the last 12 months. That number blends all months from the prior year, strong and weak alike.
The second is current run-rate xRev, which uses the same market cap but divides it by the annualized revenue from the most recent 30 days. That asks a different question: at the protocol’s actual earning pace today, how many years of revenue would equal the current market cap?
The numerator is identical in both. Any gap between the two numbers therefore has to come from the revenue trend. If revenue is stable, the latest 30-day pace should look similar to the past year’s average, and the two multiples should be close. If revenue is falling, the trailing 12-month figure still includes older high-revenue periods that no longer exist, making trailing xRev look misleadingly cheap.
AERO is Marcus’ main trap example. On a trailing basis, it trades at 3.5x xRev, implying that about 3.5 years of revenue would equal its current market cap. That only holds if the protocol can continue earning at its average pace from the last year. The problem is that the last year included quarters with revenue two to three times current levels. On today’s real revenue pace, the payback period is closer to 8.8 years.
Marcus uses a restaurant analogy. A restaurant may have posted $1.2 million in sales last year, but if it now does only $30,000 a month, saying it trades at 1x sales using last year’s figure misses the current reality. Based on actual present operations, a buyer may really be paying 3.3x current annual sales. The roughly 2.5x gap between AERO’s two xRev measures tells the same story.
PUMP is different. Whether measured on the trailing 12-month basis or the current run rate, its xRev is roughly 2.3x. Today’s revenue pace is close to the average from the last year, which suggests the revenue base has not broken. That means future moves in xRev are likely to be driven mainly by price — in other words, by changes in market belief.
Marcus’ bottom line
Marcus concludes that PUMP began as a Type B asset. It spent a year digesting the rich premium attached to its launch, and only recently moved out of the phase where the market doubted it most and into a phase where rerating can generate returns. Its revenue base, in his view, has stayed largely intact.
AERO began as a Type A asset. It already experienced one major rerating in 2024, but its multiple has since risen for the wrong reason: revenue has deteriorated faster than price has adjusted.
The setups he wants to find are the tokens currently sitting in a Type A structure: validated high revenue, compressed market cap, and a market that still does not believe the story. That, he argues, is where rerating can do most of the work.
Marcus adds that a number of newer revenue-generating projects now fit that profile. They are producing six-figure daily revenue, yet their xRev remains below 1x. The market’s current view is that those revenues are temporary — just as it once thought when AERO traded at 0.7x xRev, and just as it thought when PUMP fell to 1.29x.