Cryptocurrency zealots preaching the gospel of Web3 claim everything we do online will be powered by blockchains. Instead of large private companies managing speech and commerce on the internet, Web3 advocates say the next iteration of the internet will hand over much of that power to users.
How we communicate, exchange money, and organize ourselves in a Web3 world could be decentralized across networks of computers, rather than managed on corporate servers. Cryptocurrencies and decentralized finance (DeFi) will give people alternatives to traditional banks and financial institutions. Non-fungible tokens (NFTs) will track ownership rights of virtual goods within video games and let digital artists sell their work online. And decentralized social media could prevent content moderation by a top-down authority like YouTube or Twitter deciding what is acceptable.
But even if blockchains are decentralized, the Web3 services that interact with them are controlled by a very small number of privately held companies. In fact, the industry emerging to support the decentralized web is highly consolidated, potentially undermining the promise of Web3.
What is Web3?
The first version of World Wide Web consisted of static, isolated websites. Academics and internet critics refer to this phase, covering most of the 1990s, as Web1. The next phase known as Web2 in the early 2000s enabled interactive websites like Facebook, YouTube, and Wikipedia, where users could publish, edit, and re-share content, commonly called platforms. These websites typically relied on logins, profiles, embedded content. Crucially, they enabled the rise of user-generated content and gave power and revenue streams to individual content creators.
Web3 imagines the hardware and software of the internet migrating from corporate campuses of tech companies to distributed networks of computers that no single entity owns or controls. One of the first to promote this idea was Gavin Wood, the co-founder of the blockchain Ethereum, who outlined his vision for a crypto-powered internet in 2014. Wood wrote in a blog post that the Web should move toward a “zero-trust interaction system [where] all interactions will be carried out pseudonymously, securely and for many services, trustlessly.” He argued blockchains, the decentralized ledgers storing information across networks of computers, would almost eliminate the trust one placed in any one entity. Ultimately, he imagined Web 3.0 as the “Secure Social Operating System.”
Service with centralization
Blockchains work by decentralizing the storage of information. Instead of data being stored on a company’s servers, blockchains act as a permanent ledger, copying each new record on many different connected computers. Financial information, gaming data, and digital art pieces can all be stored and validated on these shared databases. Blockchains are designed for security (because they are encrypted), transparency (because they are public), and stability (outages are less likely because software does not rely on centralized servers, like Twitter running on Amazon Web Services servers.)
But there is a weak link in the vision of Web3, according to a widely-circulated essay by cryptographer Matthew Rosenfeld, better known as Moxie Marlinspike, of creator of encrypted messaging app Signal. Many activities that make blockchain-based activity on the web possible depend on a handful of private, centralized companies. “So much work, energy, and time has gone into creating a trustless distributed consensus mechanism,” Marlinspike wrote, “but virtually all clients that wish to access it do so by simply trusting the outputs from these two companies without any further verification.”
Marlinspike points to two weaknesses in Web3 infrastructure.
Blockchain-based apps rely heavily on application programming interfaces (APIs), which lets software communicate with other software. In Web2 sites, for example, APIs allow you to make a profile with a mobile game using your Facebook credentials. In Web3, many decentralized apps (dApps) use APIs to connect to blockchains like Ethereum rather than connect directly themselves.
The market has already consolidated around two providers, Alechemy and Infura, that power the bulk of dApps on Ethereum, the most popular blockchain for software developers. Instead of trusting Ethereum, in other words, one has to trust Infura or Alchemy.
But that’s not the only way the crypto industry is consolidated. Crypto exchanges, where users can buy and sell currencies like Bitcoin and Ether, are dominated by Binance, which handles about three times as much trading volume as the next most popular exchange. MetaMask, owned by Infura parent company Consensys, has also become the dominant crypto wallet for Ethereum—that’s where you can store the NFTs you’ve bought. And OpenSea essentially controls the NFT market, processing 84% of the $2.7 billion spent on peer-to-peer NFT transactions in December 2021.
More than half of all circulating stablecoins, a decentralized finance (DeFi) asset indexed to the value of fiat currency, are Tether (USDT), a coin controlled by the exchange Bitfinex. And crypto companies, many of which are funded by venture capital, are starting to buy up one another at a rapid pace.
This is not the decentralized Web3 that was originally promised. A few dominant companies are gaining massive market share by building user-friendly platforms to bring the next iteration of the internet to the masses. That sounds a lot like Web2.
Who’s afraid of a little centralization?
Ethereum co-founder Vitalik Buterin acknowledged many of Marlinspike’s critiques. But he predicted that the future will be different. “It’s easier to build things the lazy centralized way, and it takes serious effort to ‘do it right,'” he said. In the future, Buterin said, It will become easier and cheaper for companies and users to run programs directly on the Ethereum blockchain, even if users gravitate toward easy-to-use centralized options.
Centralization can even be a good thing for Web3, argues Marlinspike, if Web3 adopts the minimum amount of decentralization necessary to exist to prioritize users’ experience. If so, the adoption of Web3 technologies overall will grow.
“From the very beginning, these technologies immediately tended towards centralization through platforms in order for them to be realized,” Marlinspike wrote in his blog post. “This has [about] zero negatively-felt effect on the velocity of the ecosystem and..most participants don’t even know or care it’s happening.”