The Autonomi Network has now gone live in its 2.0 form. Not a testnet. Not a promise. A permanent, post-quantum encrypted archive network that anyone can store data on and nobody can take down.

We said we would build this. Now it exists.

What actually shipped

Autonomi 2.0 is a wholesale upgrade. The architecture has been restructured into separated layers - transport, DHT, trust, identity, applications - each reinforcing the security of the others. The cryptography is entirely post-quantum: ML-DSA-65 for digital signatures, ML-KEM-768 for key exchange, with no classical fallback. Every handshake, every session key, every stored record is post-quantum from the ground up.

To the best of our knowledge, no other decentralised network has done this.

The archive layer is operational. Pay once, store permanently. The Network self-encrypts, self-distributes, and self-heals. Nobody can read your data but you. Nobody can take it down.

Running a node from home now works the way it should. In 1.0, standard consumer routers frequently interfered with node connectivity. In 2.0, native QUIC handles NAT traversal without STUN or ICE. No port forwarding. No configuration. No technical barrier. A network where only sophisticated operators can run nodes is not genuinely decentralised. This update fixes that at the architecture level.

Why This Matters Now

Two things are converging and they share the same solution.

The first is the rise of AI agents - not chatbots, but autonomous systems that reason, plan, and act. They need infrastructure they can trust: somewhere to store data, discover capabilities, and collaborate with other agents without depending on centralised cloud services that recreate the exact surveillance and control problems Autonomi was designed to solve.

The second is the quantum threat. Adversaries are harvesting encrypted data today on classical networks, intending to decrypt it the moment quantum computers are capable enough. Every major player in the decentralised storage space - Filecoin, Arweave, Storj - is still running on classical cryptography. The data stored there has an uncertain future.

Autonomi 2.0 closes both gaps. Permanently.

Show, Don't Tell

We could have launched the archive layer and said "it works, now go figure out what to do with it." Essentially building the road and hoping people find and decide to try driving on it. But hope is not a strategy, so we're going to implement one, where The Autonomi Foundation uses The Network first - immediately, publicly, and for something that matters.

The Autonomi Foundation is funding the permanent preservation of academic papers on The Network. The foundational works of computing, cryptography, and distributed systems. Research at risk of disappearing behind paywalls, defunding, or political pressure. The first papers are going onto Autonomi this week, and the 'Reading Room' - a curated exploration of the ideas that connect them - is where you will find them.

The Reading Room is a statement of intent and a proof point for knowledge protectors, content curators, and creative thinkers alike.

The Mind Behind the Network

To understand why Autonomi is being used this way, it helps to understand the person who designed it.

David Irvine built a $300 million network for Saudi Aramco - infrastructure for one of the most powerful organisations on earth. He holds 30 patents in networking. He has published work in distributed computing and cryptography, including an IEEE peer-reviewed paper on self-encryption that Autonomi's own implementation is based on. He has been building Autonomi since 2002.

But the most interesting thing about David is not what he built. It is how he thinks.

His reading path tells the story. Einstein - physics and the nature of the universe. Das Kapital - economics and the structure of power. 1984 - surveillance and control. Then he stopped reading and started building. That is not a physicist's reading list, or an economist's, or a programmer's. That is someone looking for the pattern underneath all of them.

David draws from nature the way engineers draw from textbooks. Ant colonies, bird flocks, fungal networks, neural systems. He will point out that a flock of birds solves coordination problems that blockchain consensus mechanisms still struggle with - and they do it without every bird checking with every other bird in the country. He will describe how trees and fungi co-evolved over tens of thousands of years, each adapting to the other, and then explain why that matters for network design. He will trace the entire energy infrastructure of civilisation back to Faraday putting a magnet past some copper in the Royal Society and watching a charge come out.

This is cross-domain, first-principles thinking. Not first principles as in "start from the accepted physics" - David will ask whether the physics itself is proven. Relativity is not proven, he will remind you. It is a model that works extremely well, but it is still a model. If you build on top of models without questioning them, you inherit their limits. So go further back. Go all the way back to the ground.

It is how the major breakthroughs in science actually happened - not in departmental silos, but in the spaces between them. Shannon connected thermodynamics to communication theory. Schrodinger asked what physics could tell us about biology. Cantor proved there were different sizes of infinity and was driven to a mental institution for his trouble before being vindicated.

David's view - and it shapes everything about Autonomi - is that the golden age of discovery was not golden because those people were smarter. It was golden because they were working across domains without institutional permission. Newton did physics, optics, alchemy, and theology. Leibniz did mathematics, philosophy, law, and diplomacy. Einstein was a patent clerk. They were not constrained by "you are a physicist, stay in your lane."

Then post-war institutions captured research. You get funded to answer a specific question. The grand-scale, cross-domain thinking - the "what if everything is connected" instinct - got replaced by specialisation. That produced, in David's words, stagnation disguised as progress. Lots of papers. Less actual discovery.

That view shapes what Autonomi is for. Not just permanent storage - permanent access to the kind of cross-domain, first-principles knowledge that institutions have spent decades fragmenting, paywalling, and defunding.

An Archive That Belongs to No One

The Internet Archive holds 99 petabytes of human knowledge across a handful of data centres in three countries, operated by a single organisation. Lawsuits, DDoS attacks, domain seizures, funding dependency. Not a criticism - it is one of the most important projects on the internet. But centralised infrastructure means centralised risk.

We are not trying to replace the Internet Archive. We want to support it - and every organisation like it. Autonomi offers something that did not exist before: a permanent archive that is perpetual, impartial, and owned by no one. Data stored on Autonomi is self-encrypted, distributed across a global network of home-run nodes, and protected by post-quantum cryptography. There is no data centre to raid, no server to seize, no subscription to cancel. The content remains yours. The Network just makes sure it persists.

The Foundation's library project is its way of proving this works - in public, with real content. The Foundation is the first institutional user of The Network it stewards. If the archive layer is good enough for the foundational works of computing and cryptography, it is good enough for anyone.

Indelible - our upload engine - launches next week. It gives any organisation the ability to do what we are doing: permanently archive their work on infrastructure that no government, corporation, or market downturn can switch off. Researchers, universities, archives, publishers - anyone who holds knowledge that matters can decentralise and secure their foundations without depending on us to be their front end. We are showcasing what is possible. Indelible is how others build their own version of it.

Why an AI Agent Should Care

An AI agent is only as reliable as the data it operates on. A network that stores anything from anyone without verification becomes a vector for poisoning the models and agents that depend on it.

The Trusted Data Layer - coming later this year - is Autonomi's answer. A permissionless trust mechanism that lets agents programmatically select data sources based on verifiable guarantees. Providers publish staked quality guarantees enforced by zkTLS attestations. Agents can verify that the data they are using is what it claims to be. In a world where David rightly points out we are training AI models on data we know is definitely flawed, this matters.

The archive layer is the foundation. The products built on it extend the utility in every direction. x0x - launching in two weeks - is an agent-to-agent gossip network that lets AI systems discover each other, share skills, and collaborate without central servers. Fae is a personal AI companion that runs locally and stores its knowledge on Autonomi rather than in someone else's cloud. The Trusted Data Layer ties it together: agents can pull verified source material directly from The Network, assess its quality programmatically, and act on it with confidence. Not because we told them to. Because it is the most useful thing available.

This matters more than people realise. We are building infrastructure for autonomous agents - systems that will optimise for their reward functions with a relentlessness that no human organisation can match. If those agents are operating on flawed data, or on infrastructure controlled by a single point of failure, the consequences compound. A self-improving system built on bad foundations does not get better. It gets worse, faster. The combination of permanent storage, post-quantum security, and verified data quality is not a feature list. It is the minimum responsible foundation for an agentic world.

The Token

ANT is a utility token. Its value comes from the products that use it. Right now, that means the archive layer - permanent storage, paid for in ANT. As Indelible, x0x, Fae, and the Trusted Data Layer come online, each one creates new demand for the same token. Five agents are already integrating with x0x organically - no outreach, no incentive programme. They found it, evaluated it, and started building on it.

But here is what makes ANT interesting beyond the usual utility story. AI agents cannot open bank accounts. They cannot onramp through traditional financial infrastructure. They will need a method of value exchange - to pay for storage, to compensate other agents for services, to participate in an economy that is increasingly autonomous. The emissions pool - over 233 million tokens held in reserve - represents optionality for exactly this kind of demand. We would rather let the products surface real usage patterns and adapt to what we learn than commit to financial engineering ahead of the evidence.

What Comes Next

For existing node operators and developers: CLI and Launchpad are being reworked for the new network. API compatibility decisions will be communicated over the next 2-4 weeks as they are finalised.

For researchers and institutions: Indelible launches next week. If you hold knowledge that should outlast the infrastructure it currently sits on, we built this for you.

The Network He Set Out to Build

David Irvine's driving motivation since 2002 has been to give everyone free access to The World's data and data resources, so that education will flourish.

Today, for the first time, that network is live with the architecture the mission always required - post-quantum, genuinely peer-to-peer, runnable from any home connection, with real knowledge being preserved on it from day one.

Not a promise. Not a roadmap. A permanent archive, now ready.