Web3 described an internet where users own their data and identity, applications run without central operators, and value flows to participants rather than platforms. Comparing that against what exists is instructive.
What was promised
Ownership of your data and content. Portable identity across services. Applications nobody can shut down. Governance by users rather than shareholders.
What was built
Decentralised finance, which works and serves a specialised audience.
Token-based governance, which mostly recreated shareholder voting with lower participation and concentrated holdings.
Decentralised storage, which functions and is used for a narrow set of purposes.
And a great deal of speculation on tokens attached to applications with few users.
Where the vision broke down
Most users do not want to manage keys. The recovery problem is fundamental: self-custody means no recovery, and the vast majority of people will not accept that for their identity and data.
The result was intermediaries offering custody, which is the same structure as before with extra steps.
Performance and cost. Applications that must run on-chain are slow and expensive, so most moved computation off-chain to servers – which are operated by someone.
And incentives. Token models rewarded speculation over use, which selected for projects optimised for token price rather than for anything working.
What survived
The financial applications, the settlement infrastructure, and specific tools where censorship resistance genuinely matters.
The honest assessment
A movement that produced real infrastructure while failing at its stated goal, which was changing who owns the internet.