A Simple Framework for Modeling Network Value Capture
Who collects the value created by different kinds of networks, be they centralized or decentralized, Lawyer 1 or Layer 2, etc?
The Cambrian explosion of shared corporate databases pretending to be blockchains, combined with lethargic price action in layer-1 blockchain coins like ETH & Sol, has resurrected the debate around network value capture.
In crypto, as in life, price action dictates opinions, not the other way around. This isn’t the first bear market that’s hijacked people’s beliefs around what works, and it won’t be the last. So as the Suits look ascendant, and podcasters crash out, I thought it helpful to share my basic framework for what ends up capturing the value in different kinds of networks.
In a decentralized settlement system, the native coin captures substantial value
This is the closest thing we have to a tautology in our industry. The relationship between network effects and coin value has to be true for a decentralized settlement system to remain secure. The mechanics of Proof of Work and Proof of Stake break down otherwise. If you don’t believe in this relationship, you don’t really believe in crypto.
Note that I said substantial value, not all value, or even most. History will reveal whether some version of the Fat Protocol thesis is accurate. But it’s reasonable to believe that you can’t have trillions of dollars secured by a Layer-1 blockchain with only billions in economic security, there’d be too many attack vectors. At the same time, you don’t need a trillion dollars in economic security to protect a trillion dollars in assets, that would be overkill. Economic incentives are only one part of what makes a decentralized system secure.
For the first decade of crypto, there was a popular belief that blockchain was valuable, but Bitcoin wasn’t. That turned out backwards, as countless enterprise networks failed to get anywhere while Bitcoin became a trillion-dollar asset. In this second decade, it’s popular to argue that a smart contract network like Ethereum will be useful, but ETH the asset will not be. This too will be proven farcical.
It’s not that the asset issuers, dApps, or even Layer-2s that rely on Ethereum won’t capture value, they will. It’s just that Ethereum, as the settlement system, has a stronger moat than any of those. Bootstrapping a decentralized settlement system is hard, building on top of such a system is easier.
That which has the strongest moat is most likely to capture substantial value. Anyone who argues otherwise is overthinking it, and likely overinterpreting current price action.
In a centralized settlement system, the equity holders capture substantial value.
This category includes TradFi exchanges, card networks, clearinghouses, and so on. There is no debate as to who captures the value here, it is always the equity holders because they are the only game in town.
If the centralized system is run by the government, then value flows to the equity holders of the firms given exclusive access; governments don’t run open systems. Think: RTGS systems, ACH, RTP, etc.
If the centralized system is operated by the private sector, value can flow either to the equity holders of the system itself, or the equity holders of the firms given exclusive access to it.
The first category includes card networks and TradFi exchanges. The shareholders of Visa and the NASDAQ capture substantial value.
The second category includes clearinghouses and ancillary enablers of settlement activity where the network operator is structured more as a governance layer. The shareholders of the DTCC and SWIFT do not directly capture much value, but the shareholders of the brokers who can access DTCC’s services, and the shareholders of the banks that use SWIFT to coordinate payments, do. And guess what? The largest shareholders of both systems are said banks and brokers.
The “private, but not profit-making” nature of these systems is part accident of history, part political cover. DTCC’s dominance is enshrined in law, it’s a government-protected monopoly. It would be problematic if it also made a lot of money. Passing value to the next layer of shareholders is a clever way of protecting the existence of a highly profitable cabal.
Given their entrenched position and their protected profits, the biggest threat such cabals face is from decentralized systems that perform similar (or superior) services without the exclusive access. In TradFi, incumbent banks and brokers constantly machinate behind the scenes to keep startups out of the systems they control. They can’t do that in DeFi.
That’s why they’ve lobbied so hard to have the government deem decentralized systems out of bounds for regulated activity. That strategy is no longer working, so they’ve resorted to launching fake alternatives that they fully control.
In any permissioned database, the equity holders capture all the value.
So-called “enterprise or permissioned blockchains” are not a thing. They are just corporate-controlled databases that incorporate mostly useless cryptography to pretend to be of the crypto world. In reality they are no different from legacy TradFi systems; they offer none of the features of an actual blockchain.
Like legacy centralized settlement systems, the value capture from these networks goes to the equity holders. It remains to be seen whether this happens directly or indirectly. Some are structured as non-profits and run by a foundation, indicating indirect value capture. Others have raised hundreds of millions of dollars from VCs, indicating a more direct form of monetization.
Regardless, every permissioned database has a single button that controls who gets access to it, that’s why the cryptography is performative. And whoever controls that button dictates value capture. To argue otherwise is to defy the basic properties of power. It’s also ignorant of history: most of the world’s card networks and exchanges began as non-profits and utilities. Today they are massive profit centers. Why? Because there was a button.
I am skeptical such networks will ever get enough traction for value capture to even be possible. Their fake decentralization makes them chronically unsecure; they are neither regulated by a government nor by cryptoeconomic security. TradFi executives know better than anyone how dangerous it is to put your business on a network controlled by an unregulated button. The most likely outcome for the permissioned approach is a bunch of fragmented systems, controlled by competitors, with little adoption.
In a permissioned database with a coin, the equity holders capture even more value, as the coin is purely extractive.
Permissionless networks like Bitcoin, Ethereum and Solana need a coin. They can’t be secure or censorship resistant without one. A permissioned database doesn’t have this problem. Security is provided by the controller of the button (for better or worse) and censorship is the point.
It makes no sense for centralized databases to have a coin; that’s why the first generation of enterprise networks never had one. They even advertised the lack of such as a perk; “all the benefits of Ethereum, none of the hassles or regulatory uncertainty of ETH.”
Alas, crypto’s volatility proved to be a feature as quality coins soared in value, and regulatory clarity is coming, so the newest crop of corporate networks masquerading as chains also feature a coin.
You could argue that’s their main innovation over their predecessors. It’s almost like the purveyors of these products studied the various grifts the crypto industry ran on Main Street and decided to run the biggest one—that of an unnecessary coin—on Wall Street. It would be highly clever if it wasn’t immoral.
Not only does a permissioned network not need a coin for security, it also doesn’t need it to charge for usage. Permissionless systems have to charge for blockspace because it is scarce—scaling decentralized networks is hard. They charge the required fees in their own coin (as opposed to an off-chain payment or a stablecoin) to preserve censorship resistance.
Corporate databases can scale endlessly. And since they are not public, they don’t even need to charge a fee. They can still charge one to extract value, but it would make more sense to charge in fiat money, like the TradFi systems they emulate always have.
Some permissioned systems claim to need a coin to incentivise adoption, but this makes no sense. If the network solves a problem for its participants—the ostensible reason this network was created—why do you have to also pay them to adopt it? Amazon doesn’t pay companies to use AWS, it charges them for it, AWS is useful. Actual blockchains pay miners and validators because they are mercenaries, they do work that is valuable for others (users, dApps, etc) and wouldn’t show up to work if they weren’t paid. Corporate databases where the validators are the same firms who rely on its service don’t have this problem.
(It’s also worth noting that even on permissionless systems, paying for adoption, as opposed to security, has never worked).
A pointless coin that is paid to entities who would be doing an activity anyway has no value capture mechanism. Like the users of a social media platform, it is the value capture mechanism. In time, it will get dumped on unsuspecting buyers who mistake it for a real cryptocurrency. This extraction is compounded by the fact that in any permissioned system, a coin can only be launched unfairly—the controller of the button also decides who gets to earn the coin on day 1 (usually itself).
I am obviously skeptical such networks will ever get real adoption (though they’ll advertise plenty of fake adoption to promote their coin). But even if one did, the coin would also have a major supply problem. Not only is the demand for the coin of a database contrived, but the real supply is infinite.
In any permissionless system, the decentralization of the network makes it hard to change the inflation schedule. That’s why people trust Bitcoin’s 21 million supply cap. But in a permissioned setting, the controller of the button can change the inflation schedule whenever it wants. If the network advertises privacy, they can do so privately.
Endless shades of grey
There are many different designs that this mental model doesn’t fully capture, such as the utility and governance coins of decentralized applications, L1 networks with permissionless but highly concentrated validators, and L2s. I am still learning myself how to best model token value capture for all of these. But as a general rule of thumb, I always ask two questions:
Does the project even need a coin to function? If it doesn’t, then the coin won’t capture substantial value.
Is there a proximate equity holder who also holds a claim on the project’s success? If there is, coin holders won’t capture substantial value.
Let’s run through a few examples of what I mean:
Bitcoin can’t exist without a coin, and there’s no equity-enabled entity that has any major say over how the network operates. Thus BTC is the primary way to capture value from the underlying network’s success.
Ethereum can’t exist without a coin, and there’s a weak Foundation that goes out of its way to not be captured. ETH is the primary way to capture the value of the network’s success.
Hyperliquid launched without a coin so it can work without one. It has since issued one to decentralize the network, but the validator set is highly concentrated. Also noteworthy: there is a powerful founder, a for-profit, and a Foundation. The project is highly profitable, but I believe these factors nevertheless limit its upside. I am a fan, so hope to see them decentralize further. Doing so would add value to their coin.
Base doesn’t have a token and does well without one. All the value capture goes to Coinbase shareholders. If they do issue a token, they’d have to change the design drastically for it to capture value.
The value capture of any L2 token, even ones tied to a more decentralized design, will always be throttled by the fact that some of the security is derived from an L1.
Polymarket has been a massive success without a coin. That in and of itself raises serious questions about the value-capture properties of any future coin.
DeFi lending protocols benefit from having a governance token more than decentralized exchanges; credit systems require an underwriter and equity layer to scale, trading systems do not. Ceteris paribus, this dynamic means the coins of lending protocols have a higher ceiling than trading ones.
There is a long tail of supposedly permissionless Layer-1 systems that launched with a for-profit arm, one that simultaneously raised money against its own equity and the coin. Almost all of them were dead on arrival.
Databases work fine without a coin.
In conclusion
It’s still early, and much remains to be learned as to what kinds of coins capture value in the long run. There have been long stretches where the above rubric would have cost traders money, and crypto bull cycles often feature the most useless coins pumping the hardest before collapsing. Today, XRP and BNB remain among the more valuable assets in crypto, despite failing my litmus test.
So I’d understand if anyone wants to dismiss everything I just said. But I stand by it. If applied over the past decade, it would have mostly picked sustainable winners, and perhaps more importantly, not blown up.