a16z Crypto Startup School: How Does Blockchain Capture Value?
In early 2020, the well-known venture capital firm a16z held an online video course called "Crypto Startup School," covering the basics of cryptocurrency technology and important considerations for building a crypto company.
To help domestic crypto market entrepreneurs gain more methodologies, Chain Catcher translated the series of courses and made edits that do not affect the original meaning. The speaker of this article is Allie Yehia, an investment partner at a16z, who has conducted interdisciplinary research in computer security, computer networks, and distributed systems at Stanford University. She mainly introduces the core business models of cryptocurrencies and how to sustainably capture value in a balanced state.
Blockchain, in a sense, is a type of computer where participants come together through consensus algorithms, forming something that resembles a computer for developers. The applications we will discuss run on this type of computer, and they differ significantly from applications typically written using traditional computing paradigms.
We start our analysis from the ground up. The blockchain first requires computing hardware, which is referred to as the "Layer 0" hardware layer. This includes miners, validators, and peer-to-peer network protocols that enable participants at this layer to communicate with each other and ultimately reach consensus on the network state of the blockchain, such as how much cryptocurrency a participant owns. Once all participants can agree on the state, they must be able to compute on top of that state in a verifiable manner and ensure the network is correct through game theory, which is where the "Layer 1.5" computing layer comes in.
It is worth noting that the Layer 1 consensus layer and the Layer 1.5 computing layer are almost always bundled together in every blockchain system, so these two layers are often collectively referred to as Layer 1. The combination of Layer 0, Layer 1, and Layer 1.5 constitutes a complete blockchain computer, where developers can actually deploy the programs they write and let the computer run those programs. Each layer faces various challenges, but several startups have begun to tackle the various challenges at each layer.
For example, at the hardware layer, the main challenges often involve providing computing resources to ensure bandwidth and connectivity among the various nodes in the network, as well as managing and operating large data centers. The layers above the consensus layer need to ensure that there is only one true version of the computer state at any given time, and that no one can overturn the fact that everyone agrees on the key challenges of Layer 1.5 to ensure correct execution of computations, and that no one can secretly perform computations to steal cryptocurrency.
The aforementioned three layers provide users with a blockchain computer, and above this is the Layer 2 smart contracts, which serve as the fundamental infrastructure for everything on the blockchain.
Next is the Layer 3 user interface, which is actually composed of code running on mobile phones or web browsers, including wallets, exchanges, and other products used to interact with smart contracts. It ultimately becomes the bridge connecting the protocol world and the human world.
Layer 0 and Layer 3 are akin to peripherals in the crypto world, and their business models resemble more traditional business models, maximizing revenue in a way that does not confuse users.
Layers 1 and 2 require more focus, while the Layer 1.5 computing layer does not need to be discussed in detail, as it is often bundled with the consensus layer. The reason to focus on these two layers is that they ultimately become the aggregation points, and most of the value in the blockchain stack is ultimately captured by them.
Now, what is truly remarkable is that the business models of these two layers ultimately follow the same pattern, with their basic value capture mechanisms being instances of multi-sided platforms. What is a multi-sided platform? My favorite definition is that a multi-sided platform is a common foundation that creates value by enabling multiple participants in a network to interact directly.
For a concrete example, a marketplace is a multi-sided platform because it allows merchants to interact directly with villagers who want to buy goods.
Ride-hailing app Lyft is also an example, where drivers interact with the Lyft app, while the Lyft app interacts with writers on the other side of the operating system. Windows, iOS, and similar platforms are typical examples of multi-sided platforms that connect developers with users of those applications.
Therefore, I believe multi-sided platforms are the core template for providing value creation and value capture in the crypto world, and the key is that it applies to Layers 1 and 2. The core business model of Layer 1 is the most important business model in the crypto world because it enables all crypto applications. Bitcoin pioneered the incentive structure in 2008, and then Ethereum made it fully programmable and widely adopted.
Next, delving deeper into how this model works, I tend to refer to it as the Layer 1 "flywheel." It starts with a founding team that may propose a vision for a protocol, and with the help of a core development team and external investors, they can build the protocol and give initial tokens value, incentivizing validators to provide computing resources and hardware, thus securing the functionality of the platform.
Once functionality exists, there is motivation for third-party developers to build useful applications that provide utility to end users. The end result is that a community begins to form, reinforcing the protocol and its original vision.
Around this "flywheel," the stronger our vision for the protocol, the more the token is valued in a broader market, leading to greater motivation for validators to provide security and functionality, which in turn encourages developers to create more applications, ultimately creating more utility for end users.
The protocol itself is a common foundation that allows direct interaction among all participants. The protocol is a multi-sided platform that happens to have five different aspects: the founding team, investors, miners/validators, developers, and end users. Because everyone can rely on the protocol to create enforceable rules, they can interact directly and trust each other.
There are many nuances to how it ultimately captures value, which is a puzzle we must face, as you can even ask why it captures any value at all. This inevitably becomes a competitive landscape and starts the open-source process of this "flywheel."
Thus, it is a question of how to defend against competition, while also being an apparent paradox that I call the value capture paradox. At its core, it is because all the code in the stack is open-source; if this model is effective, anyone can directly copy it and commoditize any value captured by the intrinsic mechanisms over time.
Here we need to define defensibility, which refers to a company's ability to maintain a positive return over time, or a network's ability not to disappear due to competition from other participants in the ecosystem, or a project creating a moat around its core value.
Historically, almost all value capture has relied on three things. First, it is a company that has some secret that allows it to build something that others cannot; second, the company has some intellectual property that helps it gain licenses, charging rights, or similar things, enforced by some legal jurisdiction; third, the company has ownership control over certain resources that others do not. Through these things, a company ultimately has defensibility, but the blockchain industry lacks these things; all code is open-source, every technological insight is open, and available to the entire ecosystem.
This is the issue of Layer 1 defensibility. Let's do another thought experiment: if you are a loyal user of Lyft or Uber, and these companies announce that they will make all their technology code open-source, imagine the multiple significant impacts this would have, and whether you would continue to use it. I bet Lyft would be less affected compared to other platforms because if someone were to copy that technology and create a clone product, there would not be drivers to provide service on the clone product's server, so you would not easily switch.
Another important factor is the network effect, defined as the phenomenon where the value of a network to new users often increases with the number of existing users. This was already evident when the telecommunications industry was just starting out. If a telephone network has only one user, it has no value because there is no one to call. If there are already two users in the network, then the new user can call two people, increasing the network's value.
The more people in the network, the greater the value of the network to the next marginal user. This principle also applies to social networks and even monetary systems, as the more people accept any form of currency, the greater the value that currency provides, making new users more willing to trade or hold that currency.
The previous cases belong to single-sided network effects, while another type of network effect is called cross-sided network effects. Taking Lyft as an example again, the number of drivers also affects the platform's availability to passengers; the more drivers join the platform, the stronger the availability for passengers, rather than for drivers. The same situation applies to the iOS operating system; the more developers on one side of the platform, the more useful it is to users on the other side.
Returning to the topic, each connection in the above diagram is an example of cross-sided network effects. For instance, the more financial capital that enters the ecosystem through external investors, the greater the network's attractiveness to miners and validators, as there are greater token value incentives for them to provide the same production capital, thus making the network more practical for users.
However, the functionalities provided by blockchain networks differ from other examples (like iOS and other operating systems) because applications running on the blockchain can interact with each other, which is the concept of smart contracts being able to compose with one another.
For a smart contract, it can seamlessly and directly connect to other smart contracts already deployed on the platform, thereby creating the same single-sided network effects for third-party developers. Developers will want to build projects on blockchains that already have the most important projects in the ecosystem, such as Maker, Compound, Uniswap, etc., most of which are currently on Ethereum.
Similarly, on the user side, there are single-sided network effects; users want all their friends to be on this platform, and all interesting events to happen on this platform.
In addition to network effects, economies of scale can also create defensibility. When miners and validators build tools and specialize in providing services for a particular blockchain network, that network can benefit from economies of scale, as each component in the network is integrated into the ecosystem, practical applications are integrated into the user interface, and token value is embedded in a non-obvious way, making it difficult for external developers to replicate the entire ecosystem built around token value by copying the technology.
A blockchain network can gain defensibility from protocol centralization, which often leads to a winner-takes-all situation. The winning network ultimately creates an irresistible pull, making developers and users have to join this winning network rather than enter a new network, simply because the network effects are so strong.
Ultimately, tokens become the form of currency that coordinates all economic activities within the network, as well as the medium of value storage and exchange relied upon by all participants in the network.
Now, to briefly summarize the business model and value capture mechanism of Layer 1, this layer is a multi-sided platform whose defensibility comes from serving as a common foundation that supports five different types of participants to interact and trust each other.
The uniqueness of Layer 1 lies in the fact that it is owned and controlled by a broad and representative community, and the programs running at this layer also possess a form of sovereignty, as once deployed by the creators, they run autonomously without interference from anyone, including the original code writers or those controlling the physical machines executing the logic.
Next, let's talk about Layer 2. Applications based on smart contracts are also a multi-sided platform, because smart contracts enable people to trust each other, allowing them to interact in ways that were previously impossible due to sovereignty issues, thus establishing consensus through enforceable rules.
Looking at a few specific examples, Maker is one of the most important projects on the Ethereum network and is a multi-sided platform with five types of participants. Maker is a lending platform with lenders on the supply side and borrowers on the demand side, while Maker also produces a cryptocurrency called DAI, which has its own set of use cases.
The platform also has so-called keepers, who are the guardians of the network and ensure its financial health. Finally, there are holders of the Maker governance token MKR, which serves both as a mechanism to ensure all system parameters are sound and to manage risk, thus providing governance and serving as a medium for financial capital to enter the network.
Another case is Uniswap, which also runs on Ethereum as a smart contract and is a three-sided market, with buyers, sellers, and liquidity providers who need to deposit and lock assets so that buyers and sellers can match with each other, and they have a different set of incentives, namely collecting transaction fees.
Compound is a decentralized lending market where any lender can lend crypto assets, and then smart contracts can lend a portion of those assets to borrowers in exchange for interest paid on the loan.
Like Maker, the Compound market also has keepers that maintain network security, ensuring that loans with undesirable risk profiles are liquidated in a timely manner, thus keeping the entire system in good shape. At the same time, governance token holders ultimately provide governance, thus providing management rights over the system. This is also a way for financial capital to enter the system.
Further analyzing Compound, on one hand, lenders can earn a claim token after depositing into the network's smart contract, which can be redeemed for their deposit if they want to withdraw. On the other hand, borrowers can borrow from the capital pool managed by the contract by first providing some collateral, and then withdrawing the loan while paying interest on it. Most of the interest is paid to the initial depositors who placed assets in the contract, with a portion of the interest returned to governance token holders.
This is the template for the Layer 2 business model. We can see this in various examples, where there are demand and supply sides, the supply side provides services, the demand side pays for those services, and a portion of the fees is automatically received by governance token holders.
This also involves the question of how to establish defensibility. Smart contracts are completely open-source, and anyone can easily copy them. Given that governance token holders will charge users fees, it is likely that someone will copy the project code and remove or reduce the fees.
But what prevents this from happening brings us to the Layer 2 defensibility issue and the answer of network effects. Different projects at Layer 2 often exhibit different types of network effects because there is a large design space, and only projects that consider network effects can benefit from them.
Taking Compound as an example, there is a positive feedback loop between lenders and borrowers, as trading activity accumulates on the Compound smart contract, increasing the total assets within the protocol, thus providing both sides with better and more stable interest rates. Therefore, all participants in the market can generally have a better experience, and this positive feedback loop makes Compound a more useful platform for participation.
Looking at the chart above, we can see that as the assets within the Compound protocol increase, costs and user risks decrease, thus increasing the project's defensibility, making the protocol more valuable to the next new user.
If we accept the premise that the utility of smart contracts increases with the increase of capital within the protocol, then we need to do some thought exercises on how the protocol's defensibility is affected by the following three factors: the distribution of protocol assets among different smart contracts, the distribution of protocol assets among different users, and the asset attrition rate within the protocol.
In fact, Layer 2 smart contracts themselves are also multi-sided platforms, so they can gain defensibility from the network effects built into their mechanisms, as well as through integration with the surrounding ecosystem, just like Layer 1.
The final conclusion is that multi-sided platforms are the core business model of crypto networks, and defensibility ultimately comes from network effects.