Does Venture Capital make sense for Web3?
November 9, 2022
This piece is a summary of a thesis I was developing in early 2022, during my stint working in Web3 (before the big crash). In meeting with early-stage Web3 companies, advising them, attending conferences, writing smart contracts, and running an accelerator, I began to realize that there were serious structural issues in the Web3 space and planned my exit. Since that time, negative market sentiment in Web3 has been sparked by bad actors committing fraud or creating unnecessary regulatory uncertainty. I believe that fundamental problems with the technology and economics of Web3 remain, and if unfixed, will eventually cause long-term problems for the ecosystem.
I have been curious about blockchain technology and Web3 ever since I first bought Bitcoin in 2013. Open protocols, permissionless access, and the ability to participate directly in the running of a tool, network, or app are attractive from an ideological standpoint and as a tinkerer who self-hosts many of my own digital tools. My background is in physics and engineering, and so I like to take a first principles “how does this work” approach to investing which has served me well in deep-tech and hardware investing. This approach highlights some potential problems with Web3 that have been overlooked by most investors and may change how we view Web3 companies for VC investment. I hope that these concerns can spark a conversation, and I encourage others with differing perspectives to try to poke holes in my arguments.
The biggest concern I have about Web3 is its economics. The same things that make Web3 exciting may also make most of the projects built on blockchain poor targets for VC funding. Most of Web1 and Web2 venture funding in software is based on the premise that if a company grows quickly and builds network effects, even if it does so while burning money, it can eventually build lock-in and extract value from its users at scale. Startups like Facebook, Yelp, and most B2B SaaS companies raised venture dollars and grew quickly while burning money in the hopes that once they had built sufficient network and scale effects they could begin extracting value from their users who were already locked-in to the network. They would be operating at such scale and in such a defensible position that they would effectively be monopolies. New entrants, even if they had superior technology, would be unable to effectively compete against those network effects. As Peter Thiel would say, “Competition is for Losers”. Are such defensible moats possible in Web3?
Web3 is open and permissionless, and while this is fantastic for users (who can freely move their data from one app to another), it poses two unique challenges to the venture investing model: Web3’s permissionless nature makes lock-in difficult by design–with user’s rewards, achievements, and other key metadata stored in the public blockchain and accessible to all to read and build upon. At the same time, the open-source nature of the smart contracts on fully decentralized apps make it trivial for competitors to clone a successful project. For example, following the success of CryptoKitties, clones immediately started emerging, and there are even guides on how to make a clone yourself. The combination of the ease of cloning an app and the portability of data means that users can easily move between competing apps if one offers a slightly better price or service. The only defense to cloning is to make code closed-source or off-chain. L1s and L2 networks can close-source the code that runs the network, and some apps can move their UX off-chain, but as they move more off-chain, these apps really become more like Web 2.5 than Web3.
Web3’s openness is great for consumers but terrible for blockchain-first businesses. In games, this means that not only game IP, but also player progress (a traditional lock-in effect) can be copied by a competitor. For example, a third party could build a tweaked clone of Axie Infinity, allow users to import their existing NFT game items and progress, and then play in a new parallel universe, issuing its own player power NFTs (that would presumably only be used in its own ecosystem). For a decentralized cryptocurrency exchange (DEX), the effects can be even more dramatic - a competitor could simply fork their own copies of the original smart contracts and operate with lower fees. Although the original exchange may have some defensibility through the scale of its network effects, users may quickly shift to a cheaper exchange because there are no switching costs. Aggregators exaggerate these problems as users aren’t even aware of what DEX they are using. Most DEXs have extremely low fees (~0.3%) which are designed to compensate liquidity providers for their risk of impermanent loss. A DEX’s reluctance to raise fees and return value to token holders is driven not only by a fear of slowed growth but even contraction due to users moving to their no-fee competitors.
The confluence of these factors incentivizes a race-to-the-bottom on fees (subsidized by VC dollars) and a constant cutthroat competition to improve product or risk mass emigration to competitors. In effect, Web3 is the closest thing the tech industry has ever seen to perfect competition. New retention techniques, such as “go-to-community” may allow a project to lean on the stickiness of communities as a moat, but it is not perfect, as well-organized communities can switch technology providers as well. For example, Decentralized Science (DeSci) is a vibrant and growing Web3 community, but it is unclear how projects in that space can generate large profits without risking their users building or switching to their own free tools. This is a far cry from the Web2 company strategy of grow fast and attract a broad set of users, then extract value from them at scale.
I believe that these differing business realities make it far more difficult for massive companies to build and maintain a foothold in Web3, with large-scale projects seeing profit margins competed away to near-zero. At the same time, it opens an opportunity for smaller niche businesses to flourish, leveraging the permissionless nature of blockchain tech to build unique experiences for users by pulling in their existing public data. These niche companies may not individually reach a scale or profitability where it is economical for new entrants to fork their code, rebuild a UI, and try to steal their users, but would provide good cash flow to their builders. In this world there would be some large-scale applications that act more as “public goods” than for-profit companies (e.g. Wikipedia) and a cottage industry of smaller companies building out their own sets of niche games, apps, or tools that (sometimes) interact with public user data on-chain. The “public goods” projects may be a great match for a different source of funding, however: community grants from L1s and L2s on which they operate to help ensure the best experience for all users of that platform. This is doubly true because most L1s and L2s use the number of active projects as a key metric for their success.
This is a world that is great for users, but not great for startups aspiring for hyper-growth and venture capitalists that fund them. With near-zero margins and no way to lock-in users, how can these massive-scale projects justify a high valuation? With limits to the scale a profitable business can reach before competitors catch wind, fork it, and eat their margins, how can a VC see growth on their investment?
There may be some lessons that can be learned from the hardware industry. In hardware investing, we always ask ourselves: how will the company survive when (not if) a competitor copies their design and floods the market with inexpensive clones? This is not so different from copying a company’s code and presenting users with a clone with lower fees. Successful hardware companies look for ways to build competitive moats and customer loyalty by combining hardware, software, and services. For example, Peloton integrates proprietary high quality spin classes with their smart bikes, building a moat against competing bike companies. A high-end mesh Wi-Fi router may work fine on its own, but for a monthly fee you can get enhanced firewalls, VPN, and other internet security features. Web3 companies should think about what additional value they can offer by using their platform that is difficult for a competitor to replicate, and how they can leverage that for customer loyalty.
Web3 companies should also think about network effects that are difficult to replicate. At the L1 and alt-L1 level, the physical network of validator nodes can be a competitive moat. But it is much more difficult to build competitive network effect moats higher in the stack for the reasons listed above. As a result, many Web3 projects, especially Dapps, are just “flashes in the pan” and have difficulty retaining users over the long term. Any company that cracks the problem of a network effect moat on an existing blockchain network will be a big winner.
Unfortunately, the easiest way to build a moat is to move more and more of a user’s data and company’s code off the blockchain. This makes it harder for a new entrant to “steal” users and to replicate the existing experience. A social network may choose to keep all the posted photos, videos, profile information, and other updates on the blockchain, but hold the map of how users are connected to each other privately. This grants it some defensibility through a network effect, and by keeping users reliant on it for long-term value. However, this kind of lock-in makes it more like a Web2 company in practice. It is possible that most Web3 successes may actually one day turn into Web2 companies with only minor Web3 functionality.
There may be some methods of building up defensibility that I have not thought up yet - apps that securely allow users to create backups of their wallets through trusted friends, for example, may have defensible network effects limited only by the size of the L1 they are built on. Likewise, identity solutions that reach large enough scale to become a de-facto standard have strong defensibility. Large-scale niches with strong interconnected communities can be a moat, but also a limit on market size.
VC funding will remain appropriate for some subset of Web3 projects, but certainly not for all of them, and VCs may need to adjust their return expectations even for successful Web3 projects. In hardware investing, most hardware companies (for example, a company that makes generic PC webcams or keyboards) are not a good venture investment, and even those that do hit a large scale (such as Logitech) have relatively low price/revenue multiples, as their margins are constrained by competing manufacturers. Big successes in hardware rely on some form of lock in and network effect that keeps users coming back, and Web3 companies will have to struggle with similar problems. I believe that the future of Web3 investing will look very different from Web2 investing and may look a lot closer to hardware VC than most would expect.
Lastly, two extential risks face the Web3 community - The first is the misaligned incentives in staking, that push individuals to run their nodes on the cloud, rather than on their own decentralized hardware. At the time of writing, 70% of all Ethereum nodes are hosted on centralized cloud services, which could easily collude to edit the chain or damage or destroy the network. The second is the threat of quantum computing, which may be able to crack the public/private keys on which blockchain networks are based. Networks can protect themselves by switching to quantum-resistant algorithms, but that may require re-issuing public/private keys, and as we have seen from the slow rollout of Eth2.0, major upgrades to the network can take years, possibly too long given the rapid speed of quantum development in recent years.