Our Blog

Introduction

The Autonomys Network has reached a new stage of maturity. With the foundations of mainnet secure and domains active, attention now turns to evolution — scaling the protocol, expanding its capabilities, and preparing for a fully permissionless ecosystem.

This technical roadmap outlines the next phase of that work: the application interfaces that make the network accessible, the protocol improvements that make it resilient, and the research that ensures it endures.

This roadmap is indicative, not absolute. Priorities may shift as the network evolves, usage grows, and new opportunities emerge from the ecosystem itself.

TL;DR

The Autonomys roadmap focuses on three parallel layers of progress — Application, Protocol, and Research — that together drive the network toward a fully permissionless and scalable architecture.

Application Layer:

• Pay with AI3 enables token holders to purchase storage directly, marking the first step toward financialized storage.
• Auto Portal evolves as a lightweight, modular interface hub with improved staking scalability and dedicated XDM and WAI3 interfaces.
• Auto SDK introduces MCP servers for agent interoperability, allowing AI frameworks to interact natively with network functions.

Protocol Evolution:

• Adds chain monitoring for real-time observability and offline operator handling to ensure domain liveness.
• Improves farming stability, introduces an XDM EVM precompile for seamless cross-domain transactions, and provides engineering support for Game of Domains.
• Prepares for permissionless operators and domain instantiation, gets ready for third-party runtime development and data domain scalability.

Research:

• Advancing Fast XDM to drastically reduce cross-domain confirmation times.
• Exploring post-quantum and information-theoretic security to safeguard the network’s cryptography.
• Designing data domain and scalability specifications to extend the protocol’s performance ceiling.

These developments enable the network to move toward a self-sustaining, permissionless, and community-driven architecture.

Application Layer

Auto Drive — Pay with AI3

“Pay with AI3” has been a consistently requested feature from both the community and partners we have been speaking to. It will allow token holders to purchase storage directly on the network using AI3, simplifying access for builders and users alike.

Beyond convenience, this marks the first step toward the financialization of storage — where network capacity can be programmatically accessed, monetized, and composed into higher-level products and services.

By enabling direct settlement in AI3, we’re laying the groundwork for new use cases that connect decentralized infrastructure to real economic activity. We are excited to see what can be built.

Auto Portal — Evolving the Interface Layer

Autonomys is designed for modularity — and lessons learned from our testnet initiatives reinforce the value of separation of concerns. The Auto Portal will continue to evolve as a lightweight interface hub, distinct from explorers and other dApps, focused purely on core network functionality.

Explorers: Now managed through Subscan and Blockscout, providing clear, specialized visibility into both consensus and Auto EVM layers.

Staking Interface: Live and functional, supporting operator/nominator delegation. Usability and scalability improvements are ongoing, especially to handle high-volume operator participation during initiatives like Game of Domains.

• We are actively collecting feedback to inform refinements.

AI3 Conversion Interfaces: Currently, executing Cross-Domain Messaging (XDM) transactions requires using the Substrate Portal. A dedicated interface will make this process simpler and more intuitive.

• See our XDM Guide for technical context.
• In addition, we plan to have an interface to simplify wrapping and unwrapping AI3 to and from its ERC-20 counterpart.

Auto SDK — Model Context Protocol (MCP) Servers

The Auto SDK will introduce dedicated MCP (Model Context Protocol) servers for staking and XDM, enabling integration with agents and AI frameworks via industry-standard interfaces.

• This allows intelligent agents to query network state, stake, or even execute XDM transactions on behalf of users.
• The MCP standard, championed by Anthropic, is rapidly becoming a common language for interoperable AI agents.
• As we progress toward AI-native infrastructure, Autonomys aims to be a decentralized network that agents can natively interact with. We see a future where users can instruct their agent of choice to stake or XDM on their behalf.

Third-Party Domain Support

One of the most powerful properties of the Subspace Protocol is the ability to support custom runtimes for domains. We plan to provide the tooling, infrastructure, and documentation necessary to make third-party runtime development frictionless.

While demand will shape prioritization, the long-term goal is to enable any developer to build domain types suited to their specific use cases, all while leveraging the shared consensus of the farmer network.

This work dovetails with our ongoing exploration of Data Domain implementations — specialized runtimes that target throughput bottlenecks and push the limits of scalability across storage and execution.

Protocol Evolution

Chain Monitoring

The introduction of chain monitoring brings real-time visibility into the network’s performance and security events. This tool enables automated alerts and observability for events such as:

• Reorganizations or abnormal block times
• Large fund movements and specific wallet activity
• Consensus performance anomalies

You can explore the early implementation in our repository.
The system is now being tested — we welcome contributions, feedback, and feature requests from the community.

Handling Offline Operators

Operators play a critical role in executing transactions and maintaining domain liveness. However, inactive or unresponsive operators can degrade performance, particularly when heavily staked.

• Operator Primer

The upcoming “Handle Offline Operators” feature adds logic to automatically detect, mitigate, and remove operators who are no longer participating correctly.

• Protocol Spec PR #72
• Implementation PR #3711

This improvement is a prerequisite for permissionless operators, ensuring domains remain stable and reliable once registration is open to all.

Farming Improvements

Several smaller issues affecting the farming client are scheduled for resolution in this cycle. None are critical but improving stability and reliability remains a core engineering goal.

• Cache hang issue
• Stack overflow (upstream fix committed)

Updates will be reflected in both the core client and Space Acres, ensuring a synchronized upgrade path for farmers.

XDM EVM Precompile

Each domain on Autonomys includes a Substrate layer beneath its execution environment. The new XDM EVM precompile introduces a direct method for cross-domain interactions from within Auto EVM.

Instead of calling consensus functions manually, developers will be able to interact with a precompiled smart contract that handles the Substrate transporter.transfer() function on the EVM layer.

This unlocks native XDM from EVM-compatible tools like MetaMask and ethers.js — making cross-domain interactions faster and more accessible while preserving Substrate’s underlying guarantees.

Support for Game of Domains

Now that Mainnet Phase-2 has been delivered we want to get back to the continuation of our domain and staking testing initiative. Engineering resources will continue to support Game of Domains as both a proving ground and a stress-test environment.

Focused Trials of Innovation will target particular facets of the network that our engineers are interested in giving a solid workout.

During the event, the team will monitor operator behavior, network performance, and bug reports — surfacing and resolving issues before mainnet deployment. This ensures future updates are validated under real-world conditions.

Permissionless Operators and Domain Instantiation

After Game of Domains concludes and any relevant fixes are implemented, the network will move toward permissionless participation.

• Permissionless Operators: Anyone meeting minimum criteria can register and operate domains.
• Permissionless Domain Instantiation: Once security assumptions are validated, it will be possible to launch new domains without governance intervention.

This transition marks a major step in progressive decentralization — shifting decision-making and ownership from managed governance toward open participation.

Governance Evolution

Autonomys, steered by the Switzerland-based Subspace Foundation, is entering a phase of progressive decentralization, where key decisions and upgrades will transition from managed oversight to on-chain, community-led mechanisms.

In early mainnet, governance remains partially centralized to safeguard network stability. Over time, these controls will give way to on-chain proposals, voting, and delegation frameworks tied to network participation.

Engineering work is required to design and build the core governance primitives that enable this shift — transparent proposal systems, verifiable decision records, and reputation-based delegation.

The end goal is a self-governing protocol — where upgrades, domains, and ecosystem initiatives are directed transparently by the community itself.

Implementation of Fast XDM, Data Domain, and Protocol Scalability

As the research for Fast XDM, Data Domain design, and the Protocol Scalability Specification mature, engineering work will focus on translating these outputs into production-ready implementations. This includes preparing prototype environments, defining specification-level changes, and sequencing the upgrades required across runtimes, clients, and tooling.

Implementation will begin only once the underlying designs are validated, ensuring each upgrade aligns with the security, performance, and architectural principles established during the research phase.

These efforts lay the groundwork for the next generation of Autonomys performance — faster cross-domain messaging, higher throughput capabilities, and a dedicated domain model optimized for large-scale data operations.

Research

Fast XDM

Cross-domain messaging is central to Autonomys’ modular architecture. Current confirmation times range from 10 minutes (Consensus → Domain) to roughly 1 day (Domain → Domain/Consensus).

The Fast XDM initiative aims to reduce these times to 10–20 minutes without sacrificing security.

This work builds on the fair exchange problem model, which requires either synchronous conditions (implying long confirmation time) or a trusted third party. The proposal suggests using the L1 beacon chain and operator economic security as the trust anchor — a novel design that combines efficiency with strong guarantees.

Research continues with an emphasis on formal verification and minimizing auditing costs, guided by principles of simplicity and robustness.

Post-Quantum Security

Quantum computing represents an eventual but existential risk to cryptography. Recent NIST guidance has accelerated the industry-wide shift toward post-quantum cryptographic (PQC) algorithms.

We are currently in an investigatory phase, exploring both PQC and information-theoretic (IT) signatures. IT-based approaches, while less common, offer strong security against quantum adversaries with lower overhead — potentially aligning with Autonomys’ performance needs.

This research is early-stage but crucial. The goal is to prepare a migration pathway for when PQC standards mature and to future-proof the network against emerging threats to key cryptographic primitives.

Data Domain Design

A data domain is a specialized runtime optimized for data throughput, providing a dedicated environment to handle large-scale storage and bandwidth demands.

Design work focuses on ensuring cryptographic integrity while maximizing performance.

Initial specifications are being refined and can be found here. These specs will be fed into the implementation process along with ongoing research support to ensure alignment with long-term project goals.

Protocol Scalability Specification

Scalability is not just about speed — it’s about sustainable flexibility. The upcoming scalability specification defines how the protocol can evolve to support high-volume workloads and dynamic execution across domains.

Unlike PoS (which is permissioned or “less” permissionless), the Subspace Protocol’s PoAS is fully permissionless — though there are hidden networking challenges for almost any scalability design.

This research involves complex trade-offs in data availability, transaction throughput, and validation efficiency. The outcome will form the blueprint for next-generation scaling within the Autonomys network stack.

Conclusion

This roadmap reflects the next phase of engineering progress — deliberate, methodical work that transforms a live network into a truly autonomous one. Each improvement, from Pay with AI3 to Fast XDM and post-quantum research, strengthens the foundation on which open participation will stand.

As Autonomys evolves, so too does its responsibility: to remain transparent, adaptable, and ready for the challenges ahead.

Autonomys isn’t built in isolation; it’s built through collaboration. Developers, operators, and researchers who want to contribute can follow progress on our forum, participate in Game of Domains, or apply for ecosystem grants.

The journey continues, as we scale toward a truly autonomous and enduring network.

‍

October marked a month of steady, deliberate progress as Autonomys deepened its focus on post-quantum preparedness and long-term network resilience. The Engineering team delivered key infrastructure upgrades, improved observability, and strengthened developer tooling across the network, while Research advanced both internal post-quantum analysis and external academic contributions. Ecosystem efforts centered on high-conviction partner discovery rather than volume, emphasizing integrations where permanent, verifiable data meaningfully improves protocol integrity. Marketing and Community initiatives reinforced this strategic direction through updated educational materials, refined brand positioning, and improvements to user-facing resources. Together, these actions reflect a disciplined approach: building quietly, strengthening fundamentals, and positioning Autonomys for the demands of an increasingly complex cryptographic landscape.

The Subspace Foundation continued strengthening the AI3 ecosystem by advancing mission-critical initiatives that reinforce Autonomys’ long-term resilience. In October, the team executed on core infrastructure for our post-quantum implementation, keeping the network ahead of an industry only beginning to grapple with quantum risk. In parallel, we refined how we communicate Autonomys’ competitive advantage, clarifying why our architecture is superior in a crowded market. Operational discipline also remained a priority, with ongoing improvements to internal processes and cost containment. Together, these efforts support sustainability in a challenging environment and further solidify Autonomys’ foundation for future growth.

Engineering focused on improving network robustness and enhancing the overall developer experience across Autonomys’ core tooling. Offline operator detection was deployed on the Chronos testnet, strengthening reliability across operator performance monitoring, while the rollout of Auto Drive Pay with AI3 introduced new payment functionality within the storage interface. The team also delivered improved XDM support in the Auto SDK and deployed a more comprehensive chain monitoring system to increase visibility across Domains. Research efforts continued progressing our post-quantum cryptography work, with team insights reflected in the recently published survey paper, “Navigating the quantum computing threat landscape for blockchains: A comprehensive survey” in Science Direct. Together, these initiatives keep Autonomys aligned with emerging standards and positioned for future upgrades.

The Ecosystem team continued advancing strategic partner discovery and technical scoping with projects that can leverage Auto Drive’s permanent, verifiable, on-chain data storage to strengthen transparency and integrity across decentralized networks. Rather than prioritizing volume, efforts remained focused on opportunities where permanent data is mission-critical and integration paths are clearly defined. In parallel, Autonomys has been mapping potential collaborations with projects and chains pursuing post-quantum resilience, with an emphasis on credible standards and long-term alignment. Together, these initiatives are shaping a high-conviction pipeline of potential integrations that can compound the value of Autonomys’ infrastructure over time.

[ Explore Ecosystem ]

Marketing and Community efforts throughout October emphasized clarity, education, and a focused refinement of Autonomys’ brand as the network moves deeper into post-quantum research. The team began a light refresh of the visual identity to reflect the post-launch phase, anchored by long-form articles examining centralized storage fragility and, in particular, the “Harvest Now, Derive Later” threat vector, which framed Autonomys’ stance on quantum-era risk as proactive rather than reactive. Redesigned educational assets were prepared for social channels, and key website content and links were updated to improve navigation and surface core resources. On the community side, we deployed the ERC-20 compliant wrapped AI3 (wAI3) contract, updated the faucet on the Chronos testnet, and released a Chronos developer ecosystem update. EternalMint Pro was upgraded for mainnet to distribute Auto Drive backed Dawn of Autonomy NFTs, and new documentation was published to help node runners sync Taurus quickly from a snapshot if they would like to validate Crossing the Narrow Sea results. Together, these efforts strengthened ecosystem understanding and kept the community aligned with Autonomys’ evolving technical priorities.

Autonomys is entering a phase not defined by rapid announcements, but by deliberate preparation. We have always believed that readiness is built ahead of time, not in response to crisis. As the rest of the industry begins to acknowledge quantum-era realities, we’re already planning for them. With AI accelerating quantum progress, strengthening our foundation now is essential to staying secure, credible, and ready for what comes next. Over the past six months, we’ve also streamlined the organization, reduced complexity, focused our resources, and strengthened our operations so we can move with intention rather than urgency. The pace is measured. Our direction is clear. And we’re heading exactly where we need to be.

— Todd Ruoff
[CEO, Autonomys]
[ X ] | [ LinkedIn ]

‍

The quiet threat beneath immutable ledgers
Blockchain is designed for permanence. Transactions, once confirmed, are recorded on a ledger intended to endure. That permanence is the source of trust, but it also creates a structural liability. What cannot be changed cannot be retroactively protected. As quantum computing advances, the cryptography that supports this permanence is approaching a point where the assumptions that secure today’s public-key cryptography could fail. The risk is not abstract. It is a timing problem. Anything recorded on-chain under today’s signature schemes may remain exposed to signature-forgery risk if those schemes become vulnerable faster than networks can migrate.

What “Harvest Now, Decrypt Later” means and how it maps to blockchains
The timing risk has a classical analogue. In traditional security, Harvest Now, Decrypt Later (HNDL) means adversaries capture encrypted data today and decrypt it in the future when stronger computation becomes available. Public blockchains are different because transaction data is not encrypted. Instead, they rely on digital signatures to anchor ownership and authorization. The analogous risk is harvest now, derive later. According to Navigating the Quantum Computing Threat, “by utilizing Shor’s algorithm, an adversary equipped with a sufficiently powerful quantum computer can derive the ECDSA secret key Ks solely from the public-key Kp” (Feng et al., p. 28). The authors add that this “poses a significant security risk to blockchain systems” because an attacker could “successfully forge the identity of any wallet, validate themselves on the blockchain, and potentially initiate the transfer of assets from the breached wallet” (Feng et al., p. 28). In practice, adversaries can record on-chain activity and any public keys that are revealed or can be recovered from signatures today, then attempt exploitation later if quantum computers can derive private keys from those public keys.

Migration is necessary and difficult
Addressing the risk requires two coordinated tracks. Even once a quantum-resistant replacement is chosen, migration is not simple. The paper cautions that “the process of directly replacing ECDSA with another compatible digital signature algorithm that is secure against quantum attacks, presents several challenges. The migration process itself is complex and time-consuming, raising concerns about compatibility and potential issues during the transition” (Feng et al., p. 28). In addition to protocol-level signature migration, the paper considers wallet-level mitigations. Many production wallets rely on BIP39, and while quantum attacks could target keys derived from public keys, “the security of the BIP39 Seeds or recovery phrases held by current users remains intact due to the quantum-resistant properties of hashing. If we are able to design an alternative post-quantum algorithm that leverages the Seed to accommodate inactive users and preserve the address mapping, it could partially mitigate the impact of migration” (Feng et al., pp. 28–29). These approaches are complementary: protocol changes protect future signatures, while BIP39-based tooling can help migrate existing users safely.

Time as an attack vector for authenticity, not secrecy
The core risk for blockchains is not decrypting private data. It is an authenticity failure. If quantum computers eventually enable private-key derivation from public-key material, signature validity collapses for affected accounts, and any resulting forgeries would be visible but not reversible on an immutable ledger. The National Institute of Standards and Technology (NIST) has issued post-quantum standards for digital signatures and key establishment to guide migration planning. In parallel, the Autonomys Research Team emphasizes efficiency, since “larger signature sizes contribute to increased network bandwidth and higher costs associated with the consensus algorithm,” and without care, validation can become “slow and impractical” (Feng et al., p. 30).

Conclusion: preparedness over promises
Cryptography will change. Resilient networks plan for change. Industry pathways under consideration include adopting NIST-selected post-quantum signature schemes and phasing key rotation. In parallel, the Autonomys Research Team is evaluating those options and exploring its own migrations, including preserving address mapping from BIP39 seeds to support safe user migration. Our focus is methodical readiness: establishing performance baselines, building migration tooling, and defining key rotation procedures so that, when standards and tooling stabilize, transition becomes execution rather than crisis.

About Autonomys

The Autonomys Network — the foundation layer for AI3.0 — is a hyper-scalable decentralized AI (deAI) infrastructure stack encompassing high-throughput permanent distributed storage, data availability and access, and modular execution. Our deAI ecosystem provides all the essential components to build and deploy secure super dApps (AI-powered dApps) and on-chain agents, equipping them with advanced AI capabilities for dynamic and autonomous functionality.

X | LinkedIn | Discord | Telegram | Blog | Docs | GitHub | Forum | YouTube

Sources

• Autonomys Research Team — “Navigating the Quantum Computing Threat”, research led by Chen Feng, Head of Research, Autonomys Network, to be published by Computer Science Review, 2025. Some paraphrased content referencing McKinsey Digital — “When and How to Prepare for Post-Quantum Cryptography”, May 2022.
• NIST IR 8547 (IPD) — “Transition to Post-Quantum Cryptography Standards”, November 2024.
• Global Risk Institute & Quintessence Labs — “Quantum Threat Timeline Report 2024”, 2024.

‍

September marked the first full month following the AI3 token listings and Phase-2 activation, as Autonomys transitioned from launch execution to assessing the quantum challenge and positioning for long-term security. With the network’s core layers of consensus, execution, and liquidity now live, Autonomys has achieved a level of technical readiness many blockchains never attain. This milestone validates the stability and maturity of the network and sets the stage for the next strategic focus, preparing for a world where advances in AI accelerate quantum breakthroughs, and create a non-zero risk to classical cryptography. We remain committed to the long term, refining what we’ve built, and strengthening the foundation for durable security in an evolving computational landscape.

The Subspace Foundation continued strengthening the AI3 token ecosystem by advancing opportunities for liquidity and accessibility. Following the successful exchange listings, liquidity across centralized markets remained strong. The Hyperlane bridge has been deployed on Auto EVM, enabling users to wrap and bridge AI3 to Ethereum. This paved the way for the listing of a wrapped wAI3/USDT pair on Uniswap. The Guardians of Growth boosted-staking initiative continued to attract participation, reinforcing network security and broadening token distribution. ~28MM AI3 have been staked thus far, representing ~25% of available supply. Together, these efforts sustain long-term ecosystem health and further solidify the foundation for Autonomys’ growth.

[ X ] | [ Official Markets ] | [ CoinMarketCap ] | [ CoinGecko ] | [ Forum ]

Engineering focused on network stability and infrastructure refinements, while Research advanced its post-quantum threat assessment. In addition to the Hyperlane bridge, we launched the new Chronos testnet, and deprecated its predecessor — Taurus. SAFE was also deployed on Auto EVM, expanding secure wallet options for multisig custody of digital assets. Concurrently, both teams continued the ongoing evaluation of NIST-approved standards and threat modeling intended to keep Autonomys aligned with emerging best practices in the post-quantum landscape.

The Ecosystem team continues to complete strategic integrations with projects that can leverage Auto Drive’s permanent, verifiable, on-chain data storage to strengthen transparency and integrity across decentralized networks. Simultaneously, Autonomys has begun mapping potential collaborations with projects and chains pursuing post-quantum resilience, prioritizing credible standards and long-term alignment.

‍

Onboarded Partner
Autonomys is pleased to announce its integration with Gaia, an open-source, decentralized AI network providing infrastructure, compute, and inference tooling for autonomous AI agents. Developers running Gaia nodes can now store full conversation histories on-chain through Auto Drive, while agents built with the Autonomys Auto Agents Framework can call Gaia’s LLM for inference, enabling autonomous applications that both reason with decentralized intelligence and retain verifiable memory.
[ X ] | [ Blog ]

[ Explore Ecosystem ]

Marketing and community activities throughout September focused on maintaining engagement and supporting awareness around Phase-2 developments. The monthly Docs Contest continued to highlight technical contributors improving developer onboarding, while the Spark & Seed activation concluded with strong participation and retention among early farmers. The Node Atlas initiative encouraged exploration of the global farming map, showcasing network growth and distribution. With the availability of the new Chronos testnet, testing will begin for Game of Domains — Stage 1, necessary for the instantiation of permissionless operators. Together, these efforts sustain community engagement as Autonomys transitions its focus toward addressing the quantum threat, and preparing for the next era of blockchain security.

Autonomys has reached a pivotal moment, one defined by technical maturity and strategic clarity. Over the past year, we achieved every milestone we set out to accomplish: launching the consensus chain, enabling token transfers, deploying Domains, and bringing the network to life through our global community. Each success has strengthened our foundation and prepared us for what comes next as we advance toward quantum resilience.

Just as we built an AI-ready network before it became fashionable to do so, we are once again preparing early for a technology that will reshape the way we think about computing, and more broadly, cryptography as a whole. AI is accelerating quantum computing breakthroughs faster than most anticipated, and the cryptographic foundations protecting Bitcoin, Ethereum, and virtually every major blockchain face a non-zero risk of being compromised.

Autonomys spent the past 4 years building its novel and sophisticated consensus from scratch. That same engineering acumen now positions us to implement post-quantum cryptography while others are still assessing the problem. This is about preparation, not panic, and readiness today will dictate which networks survive in an uncertain future. This is the work that will separate networks built for the long term from those built for the moment. We’re positioning Autonomys not just to survive the quantum transition, but to lead through it.
— Todd Ruoff
[CEO, Autonomys]
[ X ] | [ LinkedIn ]

‍

On October 20, 2025, a major Amazon Web Services (AWS) outage disrupted internet access and operations for thousands of businesses worldwide.

The incident began early Monday morning and took most of the day to resolve, with Amazon confirming that all services had “returned to normal operations” by 3:01 p.m. PT.

The disruption, traced to Domain Name System (DNS) resolution issues in AWS’s US-EAST-1 (Northern Virginia) region, temporarily brought down many of the internet’s largest platforms and applications.

Analysts have estimated that the financial impact could reach into the billions. The event once again underscored a fundamental weakness in the digital economy: when a single provider governs so much of the internet’s infrastructure, a localized failure becomes a global incident.

The Fragility of Centralization

Centralized cloud infrastructure has long been considered reliable. Yet centralized cloud reliability is statistical, an uptime promise based on averages, whereas decentralized permanence is structural, ensured by design. Even the largest centralized providers can suffer configuration errors, routing failures, or cascading internal dependencies that compromise critical services.

When a single provider controls global data storage, routing, and replication, data sovereignty and verifiability depend on operational continuity rather than mathematical proof.

The AWS outage exposed how fragile that model is, showing how concentrated trust can amplify failure across systems.

As AI systems, IoT networks, and autonomous agents become deeply embedded in society, these risks scale exponentially. A failure in centralized infrastructure doesn’t just interrupt workflows; it halts the continuity of machine learning pipelines, sensor data aggregation, and autonomous decision-making.

Decentralized Permanence as the Antidote

Autonomys approaches the problem from first principles.
Built on a novel Proof-of-Archival-Storage (PoAS) consensus mechanism, the Autonomys Network ensures that every byte of data is cryptographically verifiable and permanently replicated across a global network of independent nodes known as farmers.

Rather than renting ephemeral storage from a single cloud provider, users and applications can store data directly on-chain, where it inherits the same permanence, security, and decentralization guarantees as the network itself.

Through Auto Drive — the user-friendly gateway to Autonomys’ Distributed Storage Network (DSN) — developers gain access to:

• True on-chain permanence with cryptographic proofs of data integrity
• Automatic redundancy through erasure-coded replication across geographically distributed farmers
• Dynamic, market-driven pricing based on pledged SSD capacity rather than centralized control

This architecture creates a data infrastructure designed not just to perform but to endure, resilient to outages, independent of intermediaries, and verifiable by anyone.

Permanence Beyond the Quantum Horizon

Operational resilience is only part of the story. The coming decade will bring the rise of quantum computing capable of breaking today’s encryption standards, eroding traditional guarantees of data security and making the permanence and verifiability of information more critical than ever.

As discussed in research conducted by Autonomys’ Research Team, led by Head of Research Chen Feng, and drawing on analysis from McKinsey Digital (2022), “Sectors need to ready themselves for post-quantum cryptography based on the longevity of data and the lifespan of systems. Data with extended shelf lives, such as corporate trade secrets, personal health records, or classified government documents, will remain valuable even after the advent of quantum computers. If such data, transferred over public networks today, remain relevant for a long time, they may face the threat of being intercepted and decrypted by future quantum computers. For instance, life insurance plans with extended terms or 30-year home mortgage loan agreements could potentially be susceptible to quantum-related risks as they will still be in effect when quantum computers become commercially accessible.”

‍

Autonomys is pleased to announce our integration with Gaia — an open-source, decentralized AI network that provides the infrastructure, compute, and inference tooling needed to create, deploy, and monetize autonomous AI agents. Developers running Gaia nodes can now store full conversation histories on-chain through Auto Drive, while agents built with the Autonomys Auto Agents Framework can call Gaia’s LLM for inference — enabling autonomous applications that both reason with decentralized intelligence and retain verifiable memory.

Key Aspects of the Partnership

• Auto Drive for Gaia Conversations
  Through the Gaia–Autonomys starter kit, users interacting with Gaia nodes can store entire chat conversations permanently on Autonomys’ distributed storage network (DSN) using Auto Drive. This allows developers and users to preserve complete agent interactions with cryptographic verifiability and tamper resistance — accessible at any time through Autonomys’ gateways.
  Example outputs from this flow can already be seen here.
  Github Repo: https://github.com/harishkotra/gaia-autonomys-starter-kit
• Gaia LLM in the Auto Agents Framework
  Autonomys has plugged Gaia’s LLM into the Auto Agents Framework, giving developers the ability to design agents that directly leverage Gaia inference while also benefiting from Autonomys’ permanent memory and modular execution environment. Agents can both invoke Gaia models for reasoning and store their working memory, outputs, and contextual histories on-chain through Auto Drive.
  Gaia support has been added to the Auto Agents Framework and can be found at:
  https://github.com/autonomys/autonomys-agents/blob/a074d7d32290444b467bee5020d5635b64a82f8f/core/src/services/llm/factory.ts#L66-L77

Benefits for Developers and Agents

• Persistence: Conversations and outputs are stored permanently with Autonomys’ Proof-of-Archival-Storage consensus, ensuring verifiable and tamper-proof memory.
• Interoperability: Gaia inference is natively available to Autonomys agents, eliminating the need for developers to build custom bridges.
• Scalability: Auto Drive’s high-throughput architecture supports large datasets and full conversation logs, while Gaia provides distributed AI compute.
• Composability: Agents can now combine Gaia’s reasoning layer with Autonomys’ verifiable storage to build richer, more autonomous workflows.

Community Impact

This integration connects Gaia’s decentralized AI inference with Autonomys’ permanent, verifiable storage, allowing developers to build agents that both reason with decentralized intelligence and preserve tamper-proof interaction histories. It marks a practical step toward AI3.0 — where autonomous agents operate with transparent, scalable infrastructure that ensures their memory and actions remain provable.

“Gaia and Autonomys are proving what decentralized AI can deliver: inference you can trust and storage you can verify — a step toward AI sovereignty that’s private, scalable, and fair.” Matt Wright, CEO of Gaia

“By combining Gaia’s inference with Autonomys’ verifiable storage, we’re closing the gap between how agents think and how they preserve what they know — laying essential groundwork for AI3.0.” Jim Counter, Head of Ecosystem at Autonomys

About Gaia

Gaia is an open-source, decentralized AI network that provides the infrastructure, compute, and inference tooling needed to create, deploy, and monetize autonomous AI agents.