a16z: How to distinguish between network tokens and company-backed tokens

Author: Miles Jennings & Scott Duke Kominers & Eddy Lazzarin, a16z crypto; Translator: Jinse Finance xiaozou

Distinguishing between network tokens and company-backed tokens can be challenging, as both types of tokens may have utility and derive some of their value from the blockchain’s on-chain functionality and the company’s off-chain efforts. However, it is critical to distinguish between the two types of tokens: network tokens and company-backed tokens pose very different risks to holders and should be treated differently under applicable law. So, what is the difference?

Network tokens are unique from company-backed tokens in that they accrue value primarily from the blockchain or smart contract protocols. This is critical because these systems are able to operate autonomously and decentralized - without human intervention or control. As a result, blockchain-based networks are able to be truly open: the system’s network effects are captured on-chain and attributed to token holders, and in principle anyone can access and build on these network effects.

In contrast, company-backed tokens accrue value primarily from off-chain systems or sources that do not operate autonomously—these are centralized systems that rely on human intervention and control. This relationship is often very obvious, such as when the token price is tied to the profits of an off-chain application, product, or service, or when the token has utility in those systems. But it can also be subtle, such as when a token without purpose or utility uses a company’s brand, which may imply that the company will drive the token’s value. In either case, if the token is intrinsically tied to a system that does not operate autonomously and primarily derives—or is expected to derive—its value from it, then it is a company-backed token. The lack of autonomy means that any associated network, while it may appear to be public, is in fact closed—like a web2 social network that is solely controlled by a single company—and therefore the network effects of the token accrue to the company that controls the system rather than its users.

These differences in network design (closed vs. open) have real economic and regulatory consequences. Because network tokens are tied to an open network that no one controls, they are more akin to commodities—able to operate in a way that precludes any party from unilaterally influencing or structuring the risks associated with network tokens. This elimination of trust distinguishes network tokens from securities, and is further reinforced if the network directs value to network tokens through its functionality (e.g., through programmatic buying and burning). In contrast, company-backed tokens have a similar trust dependency to securities: if the value of a token is derived from a closed network controlled by a single entity, that entity can unilaterally change the expected value of the token. For example, an entity with control can change the utility of a token or increase the token supply at will, or even shut down the entire system. This strongly suggests that securities laws should apply when people invest in company-backed tokens. Here are two examples to further highlight this distinction: * ETH is a classic example of a network token. It enables holders to trade on the Ethereum network and gives holders an economic interest in the network. The network is decentralized and able to operate autonomously (no individual or management team controls it). As a result, the SEC itself has concluded that securities laws do not apply to ETH.

*FTT is a clear example of a company-backed token. Its value is entirely dependent on the continued operation of the FTX exchange, which is itself a centralized exchange operated and controlled by the company. FTX company extracts a portion of the profits from operating the exchange to buy back FTT, which drives its economic value. Therefore, FTT is essentially a profit interest of FTX - its utility and value are controlled by FTX - and should be subject to securities laws.

In between these two extremes, there can be ambiguity. But it is generally possible to determine whether a token is a network token or a company-backed token by answering three questions:

*Is the system's network design open?
*Do the system’s network effects accrue to the protocol and token holders?
*Does the system enable the protocol and token holders to accrue value independently? If the answer to all of the above questions is yes, then in theory the system should be able to continue to operate without the initial development team, even if with reduced functionality. This is critical because it means the system is able to operate without control.

Let’s look at a few more examples to better understand these concepts:

The tokens associated with most decentralized exchange smart contract protocols (DEX) are network tokens, even though the initial development team typically operates the front-end website and off-chain routing software for these protocols. Why?

DEX protocols are typically open networks, meaning that anyone, not just the initial development team, can operate a front-end website and their own routing software on top of the protocol. This means that the network effects of most DEXs accrue to the protocol and token holders, not the companies that built them. Specifically, the liquidity that is critical to the operation of a DEX is controlled by the protocol itself, not the initial development team.

Value accrual can occur in multiple places, but the key question is: Can value accrue to the protocol and token holders independently of the initial development team?DEXs often have their own programmatic economics embedded in the protocol (often called a “fee switch”), and front-end site operators also charge users regular fees. The key point is that if the economics of the network token value operate independently of and are not dependent on the initial development team’s off-chain products and services, then interface-level fees are not an impediment. In other words, if the protocol fee switch is on, the value of the network token will accrue to token holders independently of the interface operations of any single company (including the initial development team). This means that the network token is economically independent of and not controlled by the initial development team.

Thus, even if the initial development team abandons a DEX that meets all of the above criteria, the DEX will continue to operate. Therefore, the system's tokens should be correctly classified as network tokens.

Take games as an example. Although there are many full-chain games, most web3 games rely on off-chain services (such as servers) to run. But not being a full-chain game does not mean that the game cannot have a network token. If the core assets - items, characters, etc. - are issued and recorded on the chain and are not controlled by any single party, then the system can be said to operate like an on-chain network. Therefore, the network can be open, allowing anyone to build with the core assets of the network (that is, if this privilege is not reserved for the initial development team), which means that the network effects of the game can be attributed to token holders rather than the original game developers. Programmatic economic mechanisms designed to accrue value to game tokens will support the identification of tokens as network tokens - the system can continue to operate and create value without the initial development team.

Let's look at decentralized social media protocols. Many of these protocols use a combination of on-chain and off-chain components. For a social network to be open and for its network effects to accrue to token holders rather than to a centralized company, users must be able to find each other and communicate, even if the rest of the network tries to prevent this from happening. One way to achieve this is to keep authentication keys associated with user account registration and posting social media messages on-chain, and allow any developer to build their own clients on the network - while storing posts and other user interactions via an off-chain network of "hubs", with each hub replicating the network's data and state. Due to these properties, the network can be said to be open - anyone can create an account using publicly accessible blockchain smart contracts and then post content using that account. The network has strong protections against centralized control because hubs are maintained by multiple parties. Such a network can still have a network token despite the presence of centralized user applications that provide access to the network. The independence of the system can be further enhanced by adding programmatic economic mechanisms that accrue value to the network token.

Now let's imagine what would happen if Apple introduced an App Store token. Users holding tokens may receive App Store discounts, or they can use these tokens to pay for apps, and a portion of all blockchain-based app payments in the App Store can be allocated back to token holders through smart contracts. Despite the use of blockchain technology, Apple's tokens will be company-backed tokens because: the system is closed, and the use of blockchain technology will not allow third parties to exploit Apple's network effects and build competing app stores in the system. In addition, the value is derived from proprietary off-chain products and services (the App Store) controlled by Apple. Even if there are on-chain programmatic economic mechanisms, all value accumulation will stop once Apple closes the App Store. Therefore, the risk characteristics of the tokens will be closer to APPL shares, which are very different from network tokens and may be subject to securities laws.