The Great Divergence
Perception vs Market Realities
Over the past few months, we have avoided talking about hypotheticals and, instead, focused on economic realities in our writing. The reason is fairly straightforward. At a time when institutional pools of money flirt with crypto, the economic fundamentals would matter infinitely more. Founders will need to position themselves well to benefit from it.
The article today is the summation of several conversations & observations we’ve had internally over the past few months. Krishna Sriram helped me put together these points in a way that flows well. Think of it as an open memo we’d have sent our own portfolio of ~23 companies. The primary argument I make is that crypto is past its phase of infancy. We are crossing the chasm to a period where firms with revenue will run laps around those who don’t.
I break down the reasoning behind this—first through comparisons with the open-source ecosystem, then with specific instances of products that have done millions in revenue. As always, if you are a founder or an investor we’d love to talk. Reach out go us at venture@decentralised.co.
On to the story…
Hey, hi, hello!
Humans are creations of sentiment. Shaped by and made of intense emotions. Of these, nostalgia stands out. A state of yearning for what used to be the norm—the status quo.
This state of nostalgia makes us vulnerable to the changes brought about by technology. Let’s call it mental inertia; an inability to let go of previous models of thinking. When the fundamentals of an industry shift, insiders or early adopters cling to how things used to work. When electric bulbs came around, some lamented how oil lanterns were better. In 1976, Bill Gates had to write an open letter to hobbyists who were annoyed at him for trying to build paid and privately owned software.
Today, Crypto is experiencing its own version of mental inertia.
Between copious amounts of Kopi and Nasi Lemak, I have been trying to think about how we evolve as an industry from here. In practical terms, we are living the dreams of DeFi summer. Robinhood is issuing stocks on a blockchain!
What do founders and capital allocators do when we finally cross the chasm? How does crypto’s core culture evolve when the marginal person on the internet starts using these tools? I explored these themes in loose terms in past essays on monetary velocity and volatility as a service.
Today’s piece tries to explain how monetary premium is earned as a function of underlying economic activity being distilled into compelling stories.
Let’s dig in.
Power Laws Power Loss
Venture capital finds its roots in the 1800s when whaling expeditions used to occur. Pools of money would be allocated towards gathering ships, men and the necessary tooling required to hunt whales. A successful endeavour would usually yield a tenfold return on capital. This meant most endeavours could fail owing to bad weather, a ship sinking, or even employees revolting, so long as one returned a multiple on the original investment.
The same philosophy applies to venture capital today. Most startups can fail in a venture portfolio so long as there is one outlier.
There is a common denominator linking the whaling expeditions of the past century and the surge of applications developed in the late 2000s: market size. Whale hunting missions worked so long as there was a large enough market for whale oil. Building applications worked so long as there were enough users to build network effects. The density of potential users in both instances helped create large enough markets to warrant a high enough return.
In contrast, much of the ecosystems built on L2s have divided a small, and increasingly desperate market. In the absence of volatility or new wealth effects (like from meme assets on Solana), these users will have little incentive to bridge from one chain to another. It is the equivalent of going all the way from North America to Australia for hunting whales. The lack of economic output is reflected in what we see in the price of these tokens.
One way to understand this phenomenon is through the lens of protocol socialism. It is a point in time, where protocols subsidise open source applications, even when they have no users or economic output through grants. The benchmark for such grants was often social proximity or technological alignment. A popularity contest, instead of an efficient market. One that was subsidised by increasing interest in tokens.
In 2021, when liquidity for large-cap assets was abundant, it did not matter if a token generated enough fee revenue. It did not matter if many of its users were bots. In fact, you could skip having apps, too. The reason was that individuals were betting on the hypothetical probability that a single protocol would attract large numbers of users.
It would be like being able to own equity in Android or Linux when they were just about to take off.
The problem is that for much of open-source innovation’s history, tying capital incentives with forkable code has not worked. Corporations like Amazon, IBM, Lenovo, Google and Microsoft directly incentivise their developers to contribute to open source.
As of 2023, Oracle was a leading contributor to kernel changes for Linux. Why would for-profit enterprises allocate money towards these operating systems? The answer is fairly straightforward
They can use these things to build other things that can make money. AWS generates billions of dollars in revenue partly because much of the server infrastructure leans on Linux. Google’s focus on Android being open source enabled it to loop in Samsung, Huawei and other manufacturers that helped create the mobile ecosystem it powers.
These operating systems have sufficient network effects to warrant allocating money towards their development. Over three decades, there has been a sufficient number of people using them for large enough economic activity to warrant pursuing clout.
Contrast this with the L1 ecosystem today. According to data from DeFillama, there are over 300 L1s and L2s. Of these, only seven have had enough economic activity to warrant north of $200k in fees in the past day. Only ten have more than a billion in TVL in their ecosystem.
If you are a developer, trying to build many of these L2s is like trying to set up shop in the middle of a desert. Liquidity is scarce, foundations are shaky, and users have no reason to drive all the way to you unless you are giving away dollars. Ironically, between grants, incentives & airdrops, that is what most applications are forced to do. As a developer, you are not competing for a portion of protocol fees, but those fees are symbolic of what happens on the protocol.
In such an environment, economic output matters less. Attention & theatrics tend to garner more attention. The thing does not need to make money as much as it seems like the people are building the thing. So long as individual participants are buying tokens, this would work. Living in Dubai, I have often wondered why there are drone shows or taxi ads for tokens.
Do CMOs expect users to come out of the little desert haven I call home? Or why are so many startup founders oriented towards “KOL Rounds”?
The reason is that there’s a bridge between attention and capital infusion in Web3. Garner enough eyeballs, infuse enough FOMO, and you have a shot at a high valuation.
All economic endeavour stems from attention. Without holding a person’s attention long enough, you cannot convince them to talk, date, work or transact with you. But when attention is all we seek, the bills eventually come due in all aspects of life. In the age of AI slop, redundant playbooks used by L2s to be perceived as valuable no longer work.
If everyone is doing the same old-tier 1 VC, large exchange listing, random token grants & non-sticky TVL games, nobody stands out as special. This is the sad reality that much of crypto is waking up to.
In 2017, you could afford to build on Ethereum, even with a lack of users, because the underlying asset (ETH) would rally some 200 times in the space of a year. A similar wealth effect happened in 2023 with Solana, where the base asset rallied ~20 times from the bottom, and had a series of meme asset rallies.
New wealth effects sustain open-source innovation in crypto when there is appetite from investors and founders. Over the past few quarters, that playbook has inverted. Fewer individuals angel invest in startups. Fewer founders have personal runways to last a venture funding freeze. And even fewer funds do large rounds.
One place where the after-effect of lagging apps becomes apparent is in the Price to Sales ratio for prominent networks. The lower the figure here, the healthier it often is. Conventionally (as you’ll soon see for Aethir), the P/S ratio trends lower as revenue increases. But for most networks that is not the case. Valuations stay intact owing to new token emissions, while revenue either declines or stays intact.
The sample of networks below consists of those that have been legitimately trying to build over the past few years. The numbers, however, are reflective of economic realities and not subject to my opinion. For Optimism and Arbitrum, the P/S ratio is closer to a sustainable 40-60x. For some of the other networks, the numbers are as high as a 1000x multiple.
So, where do we go from here?
Revenue Eats Perception
I have been fortunate enough to be early to multiple data products within crypto. In my experience, there are two that have been the most impactful
Nansen - The first to label wallets using AI, and give a bundled view of how capital flows
Kaito-The first to use AI to observe which products on crypto-Twitter have the most mindshare. Or interestingly, track down creators that make the most noise on individual protocols.
I find the timing of the release of these two products rather interesting. Nansen was released in the middle of the NFT and DeFi boom. A time when individuals wanted to keep track of what kind of wallets were accumulating a certain token. Or where the capital from a certain wallet was deployed. Even today, I use Nansen’s stablecoin index as a measure of risk appetite within Web3.
Kaito, on the other hand, released around Q2 of 2024. Right after Bitcoin’s ETF rally began. It was a time when capital flows did not matter as much. Instead, what was being managed was perception. Make the most amount of noise on Twitter, and individuals would presume you are hot. One was a transaction economy product, the other focused on how attention is distributed at a time when on-chain transactions were in decline.
Kaito became the source to benchmark how attention flows. In turn, it disrupted how marketing worked in crypto. You could no longer spin up a bunch of bots or work with fake metrics to be perceived as valuable.
With the benefit of hindsight, I think it is fair to say that while perception fuels discovery, it does not sustain growth. Many of the names that were “hot” in 2024 are now down 90%. In contrast, there are a bundle of applications that have been built slowly and steadily over the last few years.
Broadly, you can categorise them into niche, verticalised ones with tokens. And centralised ones without tokens. Both of them follow the conventional arc of a product finding PMF over time.
To understand my point, consider the chart above exploring how Aave and Maple Finance’s TVL have evolved. According to data from TokenTerminal, in aggregate, Aave has spent $230 million to create a loan book that is worth $16 billion as of writing. Maple Finance has similarly spent $30 million to build a loan book of $1.2 billion. However, if you look at the volatility in earnings over time, you will see Maple’s gradual steadiness.
Instead of focusing on the broad, large-scale DeFi market Aave focused on, Maple’s approach was to go after institutions. Currently, both protocols make roughly similar amounts in earnings. Both Aave and Maple trade at a P/S ratio of around 40.
Aave is more volatile from an earnings perspective. The initial burn in Q4 2021 may have been for building up TVL, which in turn did establish a strong moat. As late as Q1 of 2025, that moat has translated to rising earnings. But compare it with the chart for Maple and you will see that one is more capital efficient than the other. Crypto’s great divergence is visible in these two metrics.
On one end, you have protocols that were early, spent heavily and built moats of capital. On the other, you have products that have built niche, vertical markets.
Consider, for example, the chart below that breaks down Maple’s AUM. Much of it is focused on institutions looking for yield on Bitcoin or USDC-based yield. The yield on USDC here comes from loans made to institutional borrowers. As of writing, the interest rate on the dollars lent here is at 9.9%. Where Aave looks at the broader over-collateralised lending market, Maple looks at the underserved corporate lending market. This is not to imply Aave is unprofitable. Or that Maple is a clear winner. However, it does show the dispersion of how ventures are being built. What we are seeing is applications focusing more vertically on how they scale.
A different place where this has become apparent is with Phantom wallet and Metamask. According to DeFiLLama, Metamask has generated $135 million in cumulative fees since April 2023. In comparison, Phantom has made close to $422 million in fees since April 2024. Granted, the meme coin ecosystem in Solana is much bigger; these numbers point to a larger trend in Web3.
Metamask has been around since 2018, under the ownership of Consensys. It is a brand that is recognised by everyone who has engaged with Ethereum. Phantom, in contrast, was a new, unknown entrant. But they built where the users were. On Solana. And that has rewarded them richly. It also helped that they were one of the best wallets in the Solana ecosystem.
Axiom is an instance of this phenomenon going to the extreme. The product has done a cumulative $140 million in fees since February of this year. Just yesterday, the product did nearly $1.8 million in fees. Much of the revenue in the app layer last year came from products that were trading interfaces. Instead of dancing with decentralisation theatrics, these products build for what the user ultimately wants. Whether it sustains or not is something to be seen.
But if your product has done ±200 million in revenue in six months, and you’re privately owned, the question remains: do you even want to sustain?
To presume all of crypto will be focused on gambling alone or that tokens are not a necessity in the future we are trending is a weak assumption. That is like presuming the GDP of the United States will concentrate on Las Vegas. Or that porn is all that is there on the internet. Blockchains, at their core, are money rails. (I know we love saying that on repeat around here.)
And as long as products can utilise these money rails in niche, unorganised markets to bring together unrelated parties to engage in an economic transaction, there is value to be generated. Few products highlight this the way Aethir protocol does.
When the AI boom happened last year, there was a lack of high-value GPUs that could be rented out. Aethir created a marketplace for such GPUs to run models. Part of their market also comes from the gaming industry. For those running data centres, Aethir offers an alternative, steady source of income.
As of writing, Aethir has done ±78 million in revenue since the end of last year. They have made a little over $9 million in earnings.
Is it “perceived” to be a “viral product” on crypto-Twitter? Probably not. But its economics lean towards being one that is sustainable, all while the token’s price trends lower. This divergence between price & economic fundamentals defines the existing vibecession within crypto. On one end, you have protocols with barely any users. On the other, you have a handful of products whose revenue has surged whilst token prices have not reflected any of it.
The Imitation Game
The Imitation Game is a movie on how Alan Turing cracked the Enigma. There is a scene where they finally figured out how to crack the code on one of the most powerful encryption devices at the time. The Allies were now able to read what Germany was about to do. But they had to hold off on taking any action on the basis of the information they deciphered. Take action too soon, and they would know that the encryption mechanism has been cracked. A lot about markets, work in the same order.
Startups play a perception game. You are almost always selling the probability that the future value of a firm is more than its economic realities. Equity becomes valuable when the probability of a firm’s economic fundamentals surging increases. This is why we see Palantir’s stock price rising whenever there are hints of a war. Or why Musk’s Tesla Stock surged right around the time Trump was elected president.
The problem is, the same perception game can work in the opposite direction, too. A failure to communicate progress can reflect in price. And this lack of “communication” is where a new crop of investment opportunities will emerge.
This is crypto’s great Age of Divergence. A time where assets with revenue and PMF will dwarf ones without. An age where founders can launch applications without ever considering a token launch because the underlying protocols have matured. A time period where liquid funds will scrutinise the economics behind base protocols because exchange listings no longer justify high valuations.
Our great divergence is one where the market matures to pave the way for the next influx of capital. One where traditional equity markets flirt with crypto-native assets.
What we have are assets that have sufficient revenue to warrant economic flywheels. And ones that have enough attention to warrant virality as meme assets. You have a barbell structure—on one end, you have fartcoin, and on the other, you have the Morphos and Maples of the world. Ironically, both appeal to institutions.
Protocols that bootstrapped moats, like Aave did for DeFi, will continue to sustain, but what about founders building new products? The writing on the wall spells a few clear insights.
Issuing a token may no longer be ideal. We are increasingly seeing a subset of founders that barely have VCs on board, doing millions in revenue via trading interfaces.
Tokens that do exist will be scrutinised by traditional capital allocators based on their network revenue. This would mean fewer investible assets becoming more crowded trades.
M&As from corporate BD teams of listed firms will become more commonplace. It brings in a new pool of capital alongside token holders and VC funds into the mix for crypto.
These trends are not new on their own. Investors like Arthur (from DeFiance) and Noah from Theia Capital have been at the forefront of a switch towards revenue-based investing. What is new, however, is that more traditional pools of money are now flirting with crypto. For founders, what that means is focusing on niche verticals to unlock value from smaller customer cohorts, which can be massively profitable, as there are pools of money that are now allocated towards buying them out. This expansion of the capital subset available for founders is perhaps one of the most bullish things to happen to the industry in a while.
The question that remains to be answered is, will we will let go of our mental inertia and wake up to this shift? As with many pertinent questions in life, only time will reveal the choices we make.
Listening to this Punjabi track on loop,
Joel
Disclaimer: DCo and/or its team members may have exposure to assets discussed in the article. No part of the article is either financial or legal advice.