400 people gathered at Funkhaus Berlin on 18 June to argue that AI is accelerating the centralisation of the internet — and that decentralised technology is the only answer. Yanis Varoufakis called it technofeudalism. Gavin Wood called it a design problem. Both agreed the current trajectory is unsustainable.
The Web3 Summit 2026 ran on 18-19 June at Funkhaus Berlin — and its timing was sharper than its organisers could have planned. Bill Laboon, vice president of technical operations at the Web3 Foundation, argued that personal data is one of the most valuable assets in the digital economy. “Over the course of a digital lifetime, a person gives about $162,000 (€141,000) worth of value to various companies without knowing it,” he said, adding that the rise of AI has made questions around data ownership even more pressing. “The danger in AI often is the data that it gets from you. What we don’t want is that it knows your particular information.”
That statement is against the week’s AI news. France fired Palantir and hired a domestic alternative, as TF covered in its VivaTech Day 1 article. The US government suspended Anthropic‘s most advanced AI model over national security concerns. The Web3 Summit is arguing that the events are not bugs. They are features of an internet architecture built for extraction, not for users.
What’s Happening & Why It Matters
The Three Pillars of Web3 Summit 2026
The Web3 Summit 2026 structured its entire programme around three guiding concepts. The first is privacy as a basic condition of digital life — positioning privacy as a default, not a feature. The second is self-sovereignty as meaningful control — giving individuals real ownership over their identity, assets and data. The third is usability as the bridge to adoption — closing the gap between decentralised ideals and everyday user experience.
Those three pillars are the movement’s specific critique. Privacy has been the stated principle of the internet since its founding documents. In practice, every major platform built its business model around the opposite. Self-sovereignty — the idea that users control their own digital identity and data — exists in theory inside GDPR and similar frameworks. In practice, a person gives $162,000 worth of value to companies without knowing it over a digital lifetime. Usability is the gap that has prevented Web3 from scaling for a decade. The tools are theoretically available. Most ordinary users cannot access them without significant technical knowledge.
Varoufakis: “Technofeudalism”
The Summit’s most politically pointed speaker was former Greek Finance Minister Yanis Varoufakis — economist, author, and the most prominent critic of what he calls technofeudalism. For Varoufakis, the debate is ultimately about who controls the infrastructure shaping modern society. He has long argued that technology giants have accumulated unprecedented influence over digital systems, creating what he calls “technofeudalism.”
The technofeudalism thesis is specific. Traditional capitalism involves markets — buyers and sellers transacting in a system where neither party has absolute power. By contrast, digital platform capitalism involves landlords and renters. The platforms own the infrastructure. Everyone else — sellers, creators, users, even competitors — rents access to audiences and services on the landlord’s terms. AI accelerates the dynamic because it requires vast amounts of user data to train — data that each platform has accumulated through years of extractive business models, and which no new entrant can replicate without starting from zero.
Wood: It Is a Design Problem, Not Inevitability
Gavin Wood — co-founder of Ethereum and creator of Polkadot — offered the more technically optimistic position. Wood’s argument is that centralisation is not an inevitable feature of digital infrastructure. It is a design choice. Alternatively designed systems — using distributed ledgers, cryptographic identity, and decentralised consensus mechanisms — can provide the same services without giving any single entity control over the underlying infrastructure. Polkadot‘s specific contribution is interoperability — allowing different blockchains to communicate and transact without going through a centralised intermediary.
Wood’s critique of AI is similarly structural. The problem is not that AI systems exist. The problem is that they are trained on data accumulated by platforms with no obligation to share access, and deployed on infrastructure owned by companies with no obligation to be neutral. A decentralised AI infrastructure — where training data, model weights, and inference capacity are distributed across participants who contribute and benefit proportionally — addresses that structural problem. Whether that infrastructure exists at commercial scale is a different question. At Web3 Summit 2026, it exists primarily in prototypes.
Web3’s Own Inequality Problem
The Web3 Summit 2026 attracted genuine critics alongside its advocates. The Euronews article covering the event noted the sharpest internal tension directly. As AI accelerates the race for control of digital infrastructure, supporters and critics of Web3 alike see decentralisation as an opportunity to redistribute power — though they differ sharply on whether the technology can deliver it without creating new forms of inequality.
That critique is well-founded. Bitcoin — the oldest decentralised network — is more concentrated in ownership than the US stock market. Early participants in most blockchain networks hold disproportionate stakes that give them structural advantages over later entrants. Decentralised governance systems frequently reproduce the same power dynamics they were designed to replace. Additionally, the technical complexity of Web3 tools creates a de facto wealth filter. The people who can navigate self-custodied wallets, decentralised exchanges, and cryptographic identity systems are predominantly those with technical training and financial resources. The people most harmed by data extraction are often least equipped to use the alternatives.
TF Summary: What’s Next
Web3 Summit 2026 concluded on 19 June. The Web3 Foundation continues its grant programme for decentralised application development. Polkadot‘s 2026 development roadmap focuses specifically on AI-adjacent use cases — decentralised identity, privacy-preserving computation, and data marketplaces. The EU Digital Wallet rollout — expected in 2026 across all 27 member states — represents the closest thing to government-mandated digital self-sovereignty currently in progress.
MY FORECAST: The Web3 Summit 2026 will not produce a single breakout technology. That is not its purpose. By contrast, it will produce something harder to measure — a growing technical and political coalition that agrees the current internet architecture is extractive and unsustainable. The AI acceleration of that extraction is the specific development that makes the moment different from previous Web3 cycles. When Anthropic‘s most powerful AI model can be shut down by a government directive, and when every major platform’s AI product is trained on data users never knowingly provided — the case for user sovereignty is easier to make. Varoufakis will speak to ever-larger audiences. Wood will continue building the infrastructure. Whether the infrastructure ever reaches the scale required to challenge the platforms is the open question. It will not be answered in Berlin. But Berlin is where the argument is being made clearly enough to travel.