The Fat App Thesis

Why the upside is in apps

This is the third—hopefully last!—part in an ongoing series on the value of blockchain apps after part 1 (the proto-app thesis) and part 2 (the social app thesis). They do not need to be read consecutively, so don’t worry: you can go ahead and read this thing.

I.

No essay in crypto history has served as a self-fulfilling prophecy quite like Joel Monegro’s 2016 “Fat Protocol Thesis.” A future map of the market so good that it seems suspiciously like a catalyst for it, Monegro’s piece remains the definitive lightning rod for why crypto markets should be valued differently from any of the 20th century markets to which crypto’s motley mix of estimable economists and degenerate gamblers alike were accustomed. The broader market is still catching up to Monegro’s kind of groundbreaking implication. In the 20th century, we could only invest in businesses based on cash flows. But what if, in the 21st century, we could invest in something else completely: the digital states, free of land-and-body-based colonialism, that underlie these businesses?

More simply, Monegro argues that protocols (ie chains) will capture more value than the apps built on them. Why? Well partly crypto apps have weak moats (they can be forked). But mainly, the success of apps will drive users to accumulate protocol tokens to use them, generating network effects for chains as every app drives up the token price of whatever chain it’s built on. What that means—though Monegro never says so explicitly—is that protocol tokens operate like sort of national currencies for a digital nation, underlying transactions within an ecosystem both as a medium of exchange as well as a representation of a legal order (smart contracts) that will guarantee the transaction’s validity, all while collecting tax for that ecosystem. Apps, meanwhile, are plain-old businesses generating revenue.

You can do your own math, but it should be clear: the market cap of a nation-state’s currency, which largely correlates to the GDP of everything built on top of it, tends to be much greater than the market cap of a company. Slice that however you want, and you’ll land at an argument the market has continually vindicated over the past 8 years. Everyone knows this: the market cap of chains has crushed the market cap of apps. Protocols consistently raise at valuations of hundreds of millions of dollars pre-product, while apps with dedicated users struggle to raise at all.

Isn’t it clear? Protocols are immensely more valuable than apps. The Fat Protocol Thesis was right all along.

And yet. And yet.

As someone who loves the “Fat Protocol Thesis,” I would argue that the damage it has visited on this space has been immense and set us back years.

And what was true for the past decade is not necessarily true for the next.

II.

The first issue is that the private market has thrown itself on the altar of protocols without any of the original reasons to do so. At this point we are basically just worshipping protocols as icons, for the mimetic value of being a protocol, with the assumption that others will do so as well—that the emperor’s clothes will shine resplendently if we all believe he’s wearing them.

More concretely, to judge how much the market has bought into the Fat Protocol Thesis to the point of complete irrationality, you just have to look at the private valuations of many of the latest interchangeable, random general-purpose L2s. These L2s don’t fulfill any of the requirements of the Fat Protocol Thesis since their tokens don’t need to be used in transactions at all—in fact, these L2s often don’t even need to have a token in the first place. But narrative is often stronger than sense in crypto, and many of these L2s have comfortably reached 9-figure valuations while apps have struggled to get any kind of valuation in the first place.

Public markets, however, are not always reciprocating these valuations.

After all, there is a very simple explanation for why all this Chain Supremacy ballyhoo is problematic: a chain is only valuable if it has valuable apps. You’ll hear this ad nauseam from the chains themselves as they tout their massive performance improvements. “Of course we need to scale blockspace,” they say, “because the next top apps will need it.” But unfortunately, the logic that we need to fund apps in order for chains to be successful is never going to be compelling enough to get VCs to fund a whole category they still think will lose. For VCs, it doesn’t matter if apps will help chains to become valuable, if they don’t think that apps themselves have value of their own.

The term of “financial nihilism” is usually used in the context of retail traders reflexively pumping and dumping useless tokens under the assumption that others will pump and dump them as well. But in the framework of 2024, I think it probably better applies to the countless VCs who all thought they could pump and dump app-less protocols without funding the apps on them as well. It’s arguably “protocol,” not “memecoin” that had the strongest mimetics and weakest fundamentals this year.

And there’s mounting evidence that game is done.

III.

Because here’s the second issue:

If there’s one lesson to take from 2024, it’s that for the first time in crypto history, apps were the winners. Polymarket. Pumpdotfun. Hyperliquid. Virtuals. ai16z.

Even the most-hyped protocols have been hyped because of the novel ways they support apps. MegaEth incubating 18 “MegaMafia” projects that require high performance. Celestia building the foundations for anyone to launch a customized appchain with whatever permissions or privacy they want. And Berachain and EigenLayer both supporting apps at a protocol level by letting users put up tokens to give them liquidity and decentralized trust, respectively, while earning rewards from protocol and app alike.

The protocols themselves have started to acknowledge that if they’re digital nation-states valued by GDP, then they need to have GDP in the first place. They need to be the best place to build apps.

The most important flippening in crypto history is underway, one that can trace its roots to Ethereum launching in 2015 as a smart contract environment—a place, unlike Bitcoin, where devs could launch apps. It’s the flippening from protocols to apps, and it requires a whole new thesis.

Or rather, it requires a whole old thesis. Because what I want to claim here has been so patently true for the duration of the internet that it should be boring even to state. Actually, most of the value to be found in crypto today is to be found in apps.

Let’s call this the Fat App Thesis.

IV.

Why will the next 5-10 years be defined by apps, rather than protocols, seeing the most upside?

I’ll outline three reasons in ascending order of importance.

1.

The first, most speculative reason, is simply historical cycles. Apps are massively undervalued and protocols are massively overvalued. The internet tends to shift between decade-long infra and app cycles, and we’re reaching the tailend of a massive infra boom in which we’ve created extraordinary tech that finally works (which was not the case even two years ago). Apps have never been as undervalued as they are now relative to their actual ability to succeed.

Simply put: 2024 was the year where we started to shift from the protocol cycle to the app cycle.

2.

The second reason, more compelling reason, is that apps and protocols have switched places since the “Fat Protocol Thesis” was coined in 2016. At that point, apps were all largely interchangeable forks of each other’s trading tools, while chains were walled gardens with massive liquidity moats. But things have changed. Today, apps have all failed to fully fork each other (see: Sushiswap) because their real moat is users.

Chains, meanwhile, have become almost entirely interchangeable—the difference between one general-purpose L2 and another is brand. As new interop and chain abstraction solutions launch every week, users are increasingly able to bridge between ecosystems without knowing what chain they’re on.

That doesn’t mean chains aren’t working desperately under a polite veneer of “public goods” and “ethereum alignment” to build a wall so they can privatize users within their ecosystem. Of course they are. For some, this is the familiar wall of liquidity, though the point there is that in the case of protocols like Berachain and EigenLayer, this liquidity is essential not as a vanity metric but to support actual apps. For others, it is a relatively new wall of supporting native apps that builders actually need to get users: smart wallets and accounting software and local DEXes and popular bridges, all of which are required just for users to transact and liquidate currency.

The point in either case is the same. The only moat that chains have today is in the apps they support. It is apps who kingmake the chains, not chains who kingmake the apps.

The value at the bottom level is no longer the protocol. It’s the apps that make the protocol usable in the first place.

3.

When technical features aren’t the moat, the user is. Users spend their time where other users spend their time, and the way they find and interact with other users is through apps.

This is also, paradoxically, why I suspect everyone (in and out of crypto) has been so bearish on apps. Only a few apps ever win—because users will ultimately all draw each other to a few distinct cities on the internet. Indeed, a few web2 social apps won a decade ago, and it’s been impossible to compete with their users’ attention ever since. It’s also frankly been *hard* for anyone to come up with new ideas for apps within the limits of web2 constraints—namely app store fees, closed APIs, and the inability to spend money in a tap. The full limits of the technology were exploited by apps a decade ago.

But onchain rails enable entire new types of app experiences with financial and reputational upside that was never possible before: they get rid of the app store fees, open the API on public blockchains, and let users spend and store money easily. New technology means new types of apps, and the lesson of the internet is that a few of these, too, will win. If past internet cycles teach us anything, these apps will also become mega-apps that consume the majority of blockspace.

Of course, this time could be different. We could see the flourishing of millions of mini-apps, like all the apps in telegram or on World App, sustaining the kind of ephemeral 2-month apps that have thrived in web2 over the past few years. Ephemeral apps in particular have defined crypto, as short speculation cycles lead to gold-rush boom-and-busts (hello NFTs), and as it’s been impossible to build for any other use case than speculation among a few whales willing to brave crypto UX.

But we also have every sign now that that era is ending. Look at the current state of apps: prediction markets, contests, nfc chips, depin, and even vapes. All of these integrate speculation with social consensus to build a user base. But there’s a key difference from previous crypto apps that doesn’t get enough attention:

None of these new categories of apps depend on token price going up as the use case. For the first time, crypto is the means, not the end.

What I mean is that apps can actually win longterm and start claiming all that blockspace we’ve been generating for years. And what happens then? Well, these apps can do novel things. They can return money to users rather than to the Apple app store to incentivize their growth. They can collect revenue from every tap. And they can, ultimately, generate vast revenue, of which only a tiny fraction will go to the chain.

Oh and—as in the case of Hyperliquid or Pudgy Penguins—they can become chains. The protocol is born out of the app that has a built-in user base that solves protocols’ top issue today: the cold start problem.

Of course, I said before that chain’s don’t need vast revenue to receive vast valuations since they should be valued on something like GDP. But when most of that GDP is being generated by a few apps, it is worth asking: who really is the fattest of the stack? Is it still the chain? Or is it, far more likely, the app?

V.

Let me end by saying that I’m not pessimistic on chains—at all. Many chains are *not* interchangeable, alternately because of unique VMs or opcodes (like Solana, Irys, Movement, Eclipse), native incentives (Berachain), high-level performance in familiar VMs (MegaEth, Monad), or because they’re actual appchains customized to a specific use case (Hyperliquid, or rollups deployed on Celestia or Delta).

Apps built on these chains could only be possible on these chains. And ultimately, even if just a few apps win market share, investing in a chain is by far the best way to index on all the apps that will be built on top.

We like to think that infra and apps are at war with each other as they fight for funding from private markets. But there’s no real war for value between the two—each builds the value for the other and could not survive without each other. Besides, most apps will, I suspect, operate like protocols themselves for others to build on.

Still, though, we’ve behaved not only as though there’s a war, but as though infra has won. We’re realizing that’s deadly to infra. But what we need to realize is that it’s also a huge missed opportunity for apps.

It’s startup canon that you’re supposed to be solving a problem, and protocols all have so many glaring technical issues that there is easy funding for any founder who wants to fix them. But this mindset has also begotten a fundamental lack of imagination. For once we stop asking ourselves what problems we need to fix, and instead ask what new experiences can be enabled by new technology, we can start to focus on something this space hasn’t deigned to touch in years—the big boogeyman question of how this tech can be used. And the only projects that can answer that are apps.

For the past decade, we’ve assumed that protocols would capture value because they control the rails, and we’ve ignored the corollary: that apps control the users. There is value in both, of course. But I suspect it will be much easier to guide users to rails than to guide rails to users.

It’s apps, as they always have, that will provide value for the ecosystem as a whole.

— David Phelps
September 27-December 31, 2024
Modified from an original post here.