Web3 is a term for the idea of a decentralised, user-owned internet built on blockchain infrastructure. The word was coined in 2014 by Ethereum co-founder Gavin Wood to describe a future where applications run on peer-to-peer networks instead of centralised servers, users own their own identity and data instead of renting it from platforms, and economic coordination happens through open protocols rather than corporate gatekeepers. The concept is positioned as a successor to Web2 (the 2004-2020 era of social media, search, and mobile apps dominated by companies like Google, Facebook, and Amazon) and Web1 (the early, mostly-static web of the 1990s).
Whether Web3 is a meaningful concept or a marketing term depends on who you ask, and the debate has been one of the more divisive threads in crypto-adjacent discourse since about 2021. At its most charitable, Web3 describes the set of applications and infrastructure that use blockchains to achieve properties (user ownership, censorship resistance, composability, permissionless access) that are hard to achieve with conventional web architecture. At its least charitable, Web3 is a rebrand of the crypto industry that lets speculative products claim they are “the future of the internet” rather than admitting they are financial instruments with a lot of JavaScript.
What Web3 Is Supposed to Be
The positive case for Web3 rests on a few specific promises.
User ownership of identity and data. In the Web2 model, your profile, your followers, your content, and your reputation are all owned by platforms that can delete them, restrict their reach, or sell their data. In a Web3 model, your identity is a wallet you control, your data is stored on-chain or in decentralised storage, and your social graph is portable between applications. The ENS (Ethereum Name Service) project implements the identity side of this — your .eth name maps to an address you control and can be used across any Ethereum application. Lens Protocol and Farcaster are social networks built on this model, where your profile and posts are records you own rather than entries in a corporate database.
Permissionless access. In a Web3 model, building on top of existing infrastructure does not require permission from a gatekeeper. Uniswap is a smart contract anyone can call; a new exchange or aggregator can build on top of it without asking. This contrasts with traditional web platforms where API access requires approval, can be revoked, and often comes with terms that limit what you can build. The composability that results from permissionless access is real and it has produced things (DeFi) that would not have existed otherwise.
Economic alignment with users. In a Web2 model, the platform captures most of the economic value that flows through it. In a Web3 model, users can own tokens that give them a share of the protocol’s success. When Uniswap launched UNI, early users received a substantial airdrop that gave them partial ownership of the protocol they had helped grow. This has not worked as well as early proponents hoped — most Web3 economic models have turned into speculation rather than genuine user-ownership — but the underlying idea is coherent and has produced some genuinely novel experiments in user compensation.
The Critiques
The criticisms of Web3 are substantial and come from multiple directions.
Most “Web3” is just crypto with extra steps. A social network where every post is a transaction, every like is a signature, and every follow is an on-chain action is much worse UX than Twitter. The underlying appeal of Web2 platforms is that they are fast, free, and easy to use, and adding blockchain friction to them rarely makes the user experience better. Many early Web3 social attempts (Steem, Voice, BitClout) failed not because their ideas were bad but because the implementations were unusable.
The decentralisation is usually superficial. A Web3 application that runs on Ethereum contracts but has a centralised frontend, a centralised RPC provider, and a centralised hosting layer is not meaningfully more decentralised than its Web2 counterpart. The smart contract might be on-chain, but if the only way users access it is through a single company’s website, that company can still censor, pause, or redirect users. Real decentralisation requires the whole stack to be decentralised, and most Web3 products are not.
The economic models are often extraction, not alignment. Many projects sold to users as “Web3” are actually token-issuance schemes where founders get a large allocation, retail buyers provide liquidity, and the economic value flows from users to insiders rather than the other way around. The “user ownership” framing is sometimes a marketing cover for what is really just speculative fundraising.
Moxie Marlinspike’s critique. In 2022, the founder of Signal wrote a widely-read post arguing that Web3 was less decentralised than its proponents claimed because the infrastructure it depended on (Infura, OpenSea, Metamask) was almost entirely centralised, and because users had no incentive to run their own nodes or manage their own infrastructure. His argument — that the trajectory of any decentralised system is toward recentralisation at the usability layer — has been broadly vindicated by subsequent events, and it is one of the more thoughtful critiques of the Web3 vision from someone who actually understands the technology.
Where Web3 Actually Lives
Stripping out the marketing, the things that work under the Web3 banner are a specific set: DeFi protocols that provide genuinely permissionless financial primitives, on-chain identity systems (ENS, Farcaster, Lens) that give users portable profiles, stablecoins as a payments layer that works across borders, NFTs for specific use cases where digital ownership is meaningful, and a handful of other applications where the blockchain substrate is actually pulling its weight. These are real, and they represent a meaningful expansion of what is possible on the internet compared to what existed before.
The rest of what gets called “Web3” — most of the NFT speculation, most of the play-to-earn games, most of the “decentralised” social networks, most of the tokenised everything — has mostly failed to deliver on its promises, and the gap between the Web3 vision and Web3 reality is where most of the skepticism comes from. Whether the vision will eventually catch up or whether the Web3 label will quietly fade is an open question, and the answer probably depends on whether the next cycle produces applications that are good enough to convince ordinary users, not just crypto-natives, that the decentralised model is worth the friction.