🔗The Cloud

The Cloud forms the foundation, a 100% Kubernetes-compatible compute and storage layer that operates as a global, distributed cloud.

Built on top of the ThreeFold Grid, it draws from independent capacity providers around the world. Each contributes compute, storage, and network resources, forming a resilient, peer-to-peer infrastructure that’s already live today.

This layer makes decentralized Kubernetes real and accessible. Developers can spin up clusters directly across the grid using familiar declarative workflows, autoscaling, monitoring, and logging tools – all without reliance on centralized cloud vendors.

The result is a production-grade environment that offers the usability of conventional clouds combined with the sovereignty, security, and resilience of decentralization. Developers gain the freedom to deploy applications anywhere, anytime, without gatekeepers.

While ThreeFold’s community-driven open-source cloud has been live for years with builders already commercializing its resources, this new Kubernetes-based solution marks a major evolution: simplified, commercial-ready, and purpose-built to host the next generation of autonomous agents. It’s the necessary foundation for the sovereign agentic cloud.

Stay tuned this month for the release of the Kubernetes Cloud.

🔗The Network

The next step extends this foundation into a living, intelligent network layer.

This step introduces a full suite of decentralized network services – including VPN, DNS, CDN, and LLM gateways – all designed to make the underlying grid fully reachable, programmable, and autonomous.

These gateways transform ordinary grid nodes into public-facing endpoints, solving one of the hardest problems in decentralized hosting: achieving global accessibility while providing greater independence from scarce IPv4 addresses and centralized ISPs.

Traffic moves dynamically across the mesh. Every connection is encrypted end-to-end. And instead of relying on fixed routes or central ISPs, the network forms a peer-to-peer overlay that automatically discovers and maintains the most efficient path between nodes.

When a node or connection goes down, the network self-heals, rerouting traffic in real time to maintain uptime. This allows workloads running anywhere on the grid, even on devices without public IPs, to communicate securely and act as if they were hosted in a single, unified data center.

What sets this network apart is not just its autonomy, but its evolutionary behavior. As the grid grows, it adapts, reconfiguring routes, optimizing performance, and expanding reach.

Parts of the network are already operational, with ongoing developments making it ever more powerful, efficient, and resource-rich.

🔗The Agents

The final phase brings the architecture to life: the Agent Layer, the cognitive and operational tier of the Sovereign Agentic Cloud.

Within this layer, every person can have one or more Personal Agents – intelligent digital counterparts that manage tasks, workflows, and communication while maintaining full sovereignty over data and identity.

These personal agents can coordinate networks of “Worker Agents,” each capable of handling specialized functions such as research, coding, analysis, or transactions.

Supporting them is a secure foundation of semantic search, private active memory, and long-term, zero-knowledge storage, allowing agents to learn, reason, and operate within encrypted, self-contained sandboxes.

🔗Agent Layer Evolution: From AGI to ACI

While much of the world is racing toward AGI (Artificial General Intelligence) – massive centralized systems aiming to replicate human cognition – we’re charting a different course.

Today’s AI giants control nearly all of the compute, data, and infrastructure required to train and deploy large models. This concentration of power brings enormous risks such as surveillance-driven data collection, bias embedded at scale, and energy consumption rivaling entire nations.

ThreeFold’s approach is rooted in Augmented Collective Intelligence (ACI). Not a machine trying to mimic humans, but a system that grows through shared human experience while protecting each individual’s privacy.

There’s no hidden monetization model, no surveillance layer, no opaque black box where your information disappears. Each agent learns locally, contributes globally, and respects the sovereignty of its user. It’s a model of intelligence that grows organically, not hierarchically.

🔗Building The Vision

That is the vision of what we call the Sovereign Agentic Cloud: a fully decentralized ecosystem where the Cloud provides resilient compute and storage, the Network connects nodes into an intelligent, self-healing mesh, and the Agent layer brings autonomous, personal intelligence to life. Together these layers form a self-sovereign foundation for the next generation of digital life.

Development toward this vision is well underway, with the Cloud and Network systems already live and evolving, and agents coming in Q1 2026. Stay tuned via our news channel and chat as the rollout unfolds.

The world is standing at the edge of a technological revolution. Quantum computing, once confined to theoretical discussions in labs, has begun breaking into reality. With advancements like Google Willow and the race towards more powerful machines, quantum computers are an approaching storm, and most of us are not even close to being ready.

🔗The Quantum Threat to Security

For decades, the bedrock of digital trust has been cryptography. Public-key encryption (RSA, ECC), digital signatures, and hashing algorithms are what secure our bank accounts, blockchain transactions, confidential files, and online identities.

But here’s the problem: quantum computers are tailor-made to break these very defenses. Shor’s algorithm allows quantum systems to factor large numbers and break asymmetric encryption at a speed classical computers could never match. Once a quantum machine with sufficient qubits (the basic unit of information used to encode data in quantum computing) and error correction is operational, today’s standard cryptography becomes obsolete.

This danger unfolds in multiple ways:

• Harvest-now, decrypt-later: Encrypted data stolen today can be stored and easily decrypted by quantum machines in the future.
• Compromised blockchains: Bitcoin, Ethereum, and other cryptocurrencies are secured by vulnerable cryptography. A sufficiently advanced quantum computer could forge transactions or drain wallets.
• Broken trust in communications: VPNs, HTTPS, and secure messaging would be exposed, jeopardizing global internet security.
• Critical infrastructure risk: Power grids, defense systems, and financial networks could all be destabilized by quantum attacks.

In short, quantum computing threatens the very assumption of “safety” on which the digital world has been built.

For businesses, governments, and individuals alike, this presents both unparalleled opportunity and unprecedented danger.

🔗Why Existing Systems Can’t Adapt Quickly

Crypto networks, cloud providers, and enterprises face a brutal truth. It is absolutely critical for them to get quantum-safe.

For blockchains, transitioning means changing the very essence of their algorithms, requiring contentious hard forks that could split communities and collapse valuations.

For cloud storage providers, petabytes of vulnerable data are already locked in centralized infrastructures, making retrofitting costly and slow.

For governments and enterprises, complacency is the biggest risk, assuming quantum is “decades away” when in reality, technological breakthroughs are accelerating faster than anyone expected.

What’s needed isn’t retrofitting. It’s rethinking.

🔗Enter ThreeFold’s Quantum Safe Storage (QSS)

Zero-OS provides several core storage primitives like disks, volumes, and 0-DB. Quantum Safe Storage is constructed on top of these primitives, adding the encryption, encoding, and dispersion mechanisms that deliver a storage layer that is unbreakable, sovereign, and efficient.

QSS is the productized form of Zero-OS storage, designed specifically for the quantum age. QSS packages its core advantages into a clear value proposition:

• Quantum-Resistant Security: Mathematically encoded fragments and post-quantum cryptography protect against future quantum attacks.
• Immutable & Append-Only: Data can’t be altered or erased, ensuring integrity forever.
• Self-Healing: Automatic detection and reconstruction of lost or corrupted fragments.
• Sovereign & Compliant: Data placement can be controlled geographically for governance and compliance.
• Efficient: Achieves redundancy with only ~20% overhead, compared to 400%+ for replication systems.

🔗The QSS Storage Stack

The storage stack is made up of three core layers:

• ZDBFS (Zero-DB FS): A filesystem abstraction layer that makes the system accessible to applications, developers, and end-users.
• ZSTOR (Zero-Stor): The encoding, encryption, and distribution engine that applies error-correcting codes and disperses fragments.
• ZDB (Zero-DB): The append-only, immutable storage daemon that directly interfaces with physical drives.

🔗#1 ZDBFS (Zero-DB FS)

At the surface, Zero-OS offers the ZBDFS. In this layer, users and applications can interact with each other. To the user, ZBDFS feels like a familiar mounted filesystem. Beneath the surface, every action taken is translated into operations that the deeper architecture understands.

When applications interact with ZDBFS, it passes those instructions down to ZSTOR, which performs the real work of encoding, encryption, and dispersal. On retrieval, ZBDFS orchestrates the reverse process, seamlessly reconstructing files from distributed fragments.

The filesystem itself is robust at scale:

• It can hold petabytes of data
• Manage up to two million files per instance
• Transfer at speeds of 100 MB/s
• It is optimized for larger objects.

Metadata is redundantly maintained with configurable consistency, typically within a fifteen-minute window, ensuring durability without compromise.

🔗#2 Zero-Stor (ZSTOR)

In ZSTOR, raw input becomes mathematically encoded resilience. When data enters ZSTOR it goes through the following stages:

• The data is compressed to conserve space, then encrypted with user-defined keys
• Forward error correction transforms it into fragments redundant yet efficient equations that can survive node failures without loss.
• These fragments are distributed across multiple Zero-DB backends, with metadata carefully managed to map locations, redundancy, and integrity.

ZSTOR also watches over data. A built-in monitoring layer continuously checks the health of every connected device, automatically rebuilding missing fragments if one goes dark. ZSTOR can even synchronize across clusters in real time, ensuring redundancy and failover across geographies.

🔗#3 Zero-DB (ZDB)

Unlike conventional storage engines, ZDB is designed for simplicity. It is append-only by default. New data is added sequentially, references update to the latest version, and old data remains as history. Mutation or deletion requires a compaction process, which is possible but not part of normal operation.

ZDB is a general-purpose data store. It can hold entire unencrypted objects or the fragmented shards used by ZBDFS, and the engine itself remains indifferent to the content.

🔗Scalability in QSS

QSS is built to scale outward, not upward. Each use case runs on its own dedicated instance of ZBDFS and ZSTOR, which connect to a constellation of ZDBs spread across the grid. A typical configuration uses around twenty or more, but the exact number is configurable by the user.

These components can be distributed across geographies, giving the system reach without sacrificing integrity.

There is no ceiling here. The architecture can support an unlimited number of ZBDFS and ZSTOR instances, each operating independently yet harmoniously. Since data is always processed and consumed at the place of its creation, latency is minimized and sovereignty preserved.

The result is a storage fabric that grows horizontally, node by node, instead of demanding ever more powerful centralized machines.

🔗The Maths Behind QSS

Replication is the crutch of conventional storage. QSS leaves it behind. A file, once ingested, is broken into sixteen fragments. From those fragments, twenty equations are born. Any sixteen of those equations can rebuild the whole. Four nodes may vanish from the network, and the data remains untouched, its integrity intact.

Where traditional systems bloat themselves with 400% overhead just to promise redundancy, QSS achieves it with only 20%.

Example Implementation

Imagine a piece of data split into three fragments with values:

• a = 1
• b = 2
• c = 3

Instead of storing these fragments directly, the system encodes them as equations:

• a + b + c = 6
• c − b − a = 0
• b − c + a = 0
• 2b + a − c = 2
• 5c − b − a = 12
• …and more.

At first glance this looks like overkill: why so many equations for just three unknowns?

Here’s the math behind it:

In a system with three variables (a, b, c), only three linearly independent equations are needed to uniquely recover them.

Any set of more than three equations will be linearly dependent—they don’t add new information but can be expressed as combinations of the others.

From a storage perspective, that means only three equations are strictly necessary.

So why generate extra equations? For resilience. If some equations are lost or corrupted, the system can still pick any valid set of three independent ones and reconstruct the original fragments. This redundancy ensures the data remains recoverable even under failure.

In production, the model scales. A typical configuration splits each object into sixteen fragments, then produces twenty equations, each stored on a separate device. To rebuild the data, only sixteen of the twenty are required. The system can lose four entire nodes and still recover the file without corruption.

This mathematical method achieves redundancy with just 20% overhead, compared to the 400% overhead demanded by traditional replication systems.

🔗Deploy Quantum Safe Storage

The Quantum Daemon is designed to manage all aspects of QSS that don’t fit neatly into the roles of the core components. Its goal is to simplify deployment, operation, and recovery of QSS into a single binary with a single config file.

🔗Responsibilities

The daemon takes care of:

• Backend deployment: Initial deployment of backend zdbs and replacement of failed backends during operation.
• Service management: Installation of all required binaries and creation of system services (supports both systemd and zinit).
• Event handling: Processing hook events from zdb (replacing legacy shell scripts).
• Data integrity: Running checks and retries for failed uploads.
• Config management: Providing a single central config file for end users (QSS component configs and CLI args are generated automatically).
• Recovery: Automated restoration of QSS onto a new machine in case of frontend failure.
• Simplified operation: One binary and one config file can bring up the whole system.

🔗Usage

A subset of functionality is already implemented, with heavy changes expected as the project evolves. Today, QSS can already be deployed with quantumd using either local or remote backends.

• Local backends are useful for testing, and can be spun up quickly with a single command.
• Remote backends are used for real deployments, requiring a TFChain account and configuration details.
• Daemon mode handles retries, hooks, and caching in the background once set up.

For step-by-step installation and deployment instructions, see the official deployment guide.

🔗Comparing QSS With Traditional Storage Solutions

Here is a comparison between QSS and traditional storage to show how it outperforms conventional systems.

🔗In Closing

ZDBFS, ZSTOR, and ZDB form a storage stack that balances performance, sovereignty, and quantum resistance. Whether deployed for simple file storage, global applications, or blockchain-grade data integrity, QSS empowers organizations, developers, and individuals alike to move beyond the vulnerabilities of conventional storage and embrace a future where security, resilience, and sovereignty go hand in hand.

That’s exactly what our team is building with KubeCloud, a platform that brings Kubernetes to the decentralized ThreeFold Grid with simplicity, sovereignty, and scale.

🔗So, What Exactly Does Kubernetes Do?

At its core, Kubernetes helps teams run applications smoothly, no matter where they live, on laptops, in data centers, or in the cloud. It takes the heavy lifting out of day-to-day operations by automating what would otherwise be slow, error-prone, and manual:

• Tames complexity: The main benefit is simplification. Kubernetes can simplify the chaos of microservices, making large, distributed systems manageable for multi-team organizations.
• Boosts reliability: Built-in recovery, replication, and load balancing keep applications running with minimal downtime.
• Delivers consistency: Standardizes deployment and operations across clouds, virtual machines, and bare metal. This makes sure that apps run the same everywhere.
• Automated deployment: Roll out applications consistently across physical servers, virtual machines, and clouds without having to configure each by hand.
• Elastic scaling: Add or remove capacity when demand spikes or shrinks, keeping performance steady while optimizing resources.
• Self-healing resilience: Identifies system issues and resolves them automatically.
• Service discovery & load balancing: Gives each connected app a fixed name and address so they can always find each other, while also spreading traffic out so nothing gets overloaded.
• Storage orchestration: Works with many types of storage systems so apps can safely save their data and get it back whenever they need it.

In short, Kubernetes takes the chaos out of running modern applications. That’s why it’s become the global standard.

🔗Why the World Adopted K8s

K8s didn’t become the de facto standard by accident. Kubernetes adoption was fueled by two converging technology shifts: the rise of containers and the rise of the cloud.

🔗The Rise of Containers

Containers changed how software was built and shipped. These are lightweight “boxes” that package an app with everything it needs to run. They’re faster and more portable than traditional software setups. But once teams started running hundreds or thousands of containers, they needed a smart system to organize them. Kubernetes became that system.

🔗The Rise of Cloud

Cloud computing redefined infrastructure. These platforms allow everyone to rent computing power instantly. However, every cloud worked a little differently. What was needed was a universal system that worked across providers. Kubernetes stepped up by becoming a unifying layer for hybrid and multi-cloud strategies.

This combination made Kubernetes the go-to tool for modern infrastructure.

🔗Limitations of the Centralized Kubernetes Model

There is no doubt whatsoever about the practicality of traditional Kubernetes. As explained above, it has become the standard for running containerized applications. The issue, however, is that most clusters today live inside centralized cloud providers. This introduces several problems

• Vendor lock-in: Dependence on a single provider limits flexibility and leverage.
• Rising costs: Data-intensive workloads become expensive to run at scale.
• Reduced sovereignty: Organizations have little control over where and how data is stored or processed.

For developers, this means reduced flexibility. For enterprises, it means risk, especially when compliance, data privacy, or infrastructure resilience are on the line.

A decentralized approach to Kubernetes can potentially address these gaps by distributing workloads across independent, globally available nodes, while still preserving the ease and power of cloud-native operations. In this decentralized environment, developers can enjoy the same features, without having to surrender control to hyperscale platforms.

For AI and next-generation applications that demand both performance and sovereignty, decentralization is essential.

This is where KubeCloud comes in.

🔗KubeCloud: Decentralized Kubernetes on Steroids

KubeCloud is designed to make decentralized Kubernetes both real and accessible. Instead of relying on centralized cloud providers, it provides clusters directly on the ThreeFold Grid (TFGrid). The TFGrid is a decentralized, peer-to-peer cloud made up of independent capacity providers distributed worldwide. The Grid offers compute, storage, and networking resources contributed by individuals and organizations.

🔗What Powers KubeCloud

Under the hood, KubeCloud integrates Mycelium. Mycelium forms an adaptive, encrypted mesh across TFGrid nodes, routing traffic securely and efficiently. Unlike traditional networking, which often depends on centralized ISPs and data centers, Mycelium self-heals and reroutes around failures, creating resilient connections between workloads no matter where they run.

KubeCloud also supports multi-master clusters for high availability. Control plane nodes are distributed, and if one fails, the others automatically elect a new leader to keep the cluster running smoothly.

Key capabilities include:

• Zero-downtime upgrades: Apply updates without interrupting running workloads.
• Built-in failover: Automatically recovers from node or service failures.
• Elastic scaling: Expand or shrink services on demand to match workload needs.

For observability, users can track cluster health through integrated Prometheus and Grafana dashboards. Exposing applications to the public internet is equally streamlined. A simplified web gateway replaces the usual complexity of ingress controllers with built-in domain and prefix-based routing.

🔗What It Means for Users

So, how does KubeCloud help users? Let’s take a look.

🔗For Developers and DevOps

• Familiar workflows which includes declarative deployments, autoscaling, monitoring, logging, and collaboration workflows.
• Runs in a decentralized environment, not a centralized vendor cloud
• Users retain full control of compute, storage, and networking

🔗For Enterprises

• Production-grade Kubernetes, without lock-in. Run clusters confidently, free from dependence on a single vendor.
• Made for data-heavy teams. Tap into next-gen hardware on the ThreeFold Grid, built for AI and high-performance workloads.

🔗In Closing

Kubernetes has reshaped how the world builds and runs software. By automating deployment, scaling, and recovery. Yet, as powerful as Kubernetes is, most of its deployments still run on centralized clouds, bringing challenges of lock-in, cost, and sovereignty. That is why the next leap is decentralization.

KubeCloud combines the familiarity of Kubernetes with the resilience and sovereignty of decentralized infra. It takes the power of the TFGrid and Mycelium and packages it into a platform that anyone can adopt with minimal onboarding friction.

Kubernetes changed the world once. By leveraging decentralization and ThreeFold, KubeCloud may do so again.

Join our Telegram and stay up to date with the latest developments on ThreeFold.

Watch the demo here with timestamps for easy navigation, and leave your feedback here.

Please take a moment to review the GEP addressing June’s minting.

🚨 Voting is now open and will run until July 7th at 10:00 AM CEST. Please cast your vote via:

• dashboard.grid.tf under “TFChain - DAO”
• or the TF Connect app

ℹ️ On a separate note: According to the latest discussion on April’s re-mint, it is expected to be implemented along with June’s payout.

As always we look forward to and appreciate community feedback.

At the core of this design is Zero OS, an autonomous, minimalist operating system purpose-built to deliver compute, storage, and network primitives without relying on centralized infrastructure, traditional sysadmin practices, or hyperscaler capital.

Zero OS has been specifically designed to ensure that:

• A high level of computer expertise is typically not required to set up a node.
• There is virtually zero maintenance after initial setup. All updates happen automatically.
• Recovery from power or network loss happens without human intervention.

🔗What is Zero OS?

Zero OS is a stateless, bare-metal runtime environment, employing a minimalist design built from scratch around the Linux kernel. It exposes a tightly scoped set of primitives for compute, storage, and networking.

• Headless by Design: Runs entirely autonomously after boot. There’s no need for user intervention post-deployment.
• Ephemeral & Immutable: The root filesystem is read-only, and all system behavior is stateless and declarative.
• Ledger-Driven Provisioning: Nodes listen to a blockchain-based reservation system and respond by deploying the workloads described in those reservations.

It has no package manager, no local user accounts, no shell access, and no remote management via SSH. This drastically reduces the attack surface and operational complexity.

The OS acts as a waypoint, converting instructions stored on a distributed ledger into real-world compute deployments.

🔗Composable Primitives

Instead of exposing bloated APIs or fully fledged orchestration platforms, Zero OS provides a narrow but composable set of system primitives:

These primitives can be composed by developers or workload orchestrators (like ThreeFold Labs or TF Grid SDKs) into more complex services, such as Kubernetes clusters, storage grids, VPNs, or decentralized databases, all without a central coordinator.

🔗Programmatic Trust via Blockchain Coordination

Instead of relying on centralized control systems like AWS Control Tower or Google Cloud IAM, Zero OS nodes operate through a decentralized model driven by TFChain and peer-to-peer communication.

Each node is associated with a Twin ID, a unique digital identity stored on TFChain. When a user wants to deploy a workload (like a VM or database), they submit a reservation smart contract to TFChain. This contract includes a cryptographic signature from the user’s Twin and a hash of the workload deployment.

Here’s how it works:

• The blockchain confirms billing authorization and verifies the hash of the deployment. This creates an auditable, tamper-proof record.
• The full deployment instructions (including specs like CPU, memory, storage, and network) are transmitted off-chain via RMB (Remote Message Bus), end-to-end encrypted, from the user to the node.
• The node monitors TFChain for reservations tied to its Twin ID, verifies the signature and hash, and then automatically provisions the workload. No human intervention needed.

This setup ensures:

• Auditable Infrastructure: Every change is logged on-chain.
• Trustless Automation: No need for sysadmins or manual approval.
• Sovereign Ownership: Each workload is cryptographically linked to its creator.
• Efficiency: By transmitting heavy deployment data off-chain via RMB, blockchain usage is minimized, reducing on-chain storage and compute load.

Just like Ethereum smart contracts execute code, ThreeFold nodes execute real-world workloads. Driven by a hybrid architecture that blends blockchain trust with encrypted peer-to-peer efficiency.

🔗Grid Networks for Safe Testing and Rollouts

Zero OS runs across several different network environments, each designed for a specific stage in the development and deployment process. This setup helps the system evolve safely without risking the stability of the live grid.

Here’s how the networks work:

This tiered model makes sure that:

• New features are tested step by step
• Production remains stable and secure
• Developers can safely build and improve the system without risk to users.

🔗True Sovereignty at the System Level

ThreeFold, through Zero OS and the TF Grid, decentralizes the cloud and rebuilds it from the ground up, with sovereignty, automation, and composability at its core.

Zero OS isn’t just an operating system. It’s a self-operating system.

And in a time when data, identity, and computational infrastructure are geopolitical battlegrounds, that changes everything.

🔗What Does it Enable?

Think about this:

• A school in Uganda can run its own local cloud with no data ever leaving its premises.
• A nation-state can build a national cloud grid without relying on a centralized provider.
• A developer collective can launch resilient apps that survive even if core internet services go down.

This is infrastructure-native sovereignty. Not just ownership of software, but autonomy over the full lifecycle: deployment, operation, upgrade, and recovery.

With Zero OS and the ThreeFold Grid, you move from dependency to autonomy.

That’s what makes ThreeFold truly sovereign tech.
Not just infrastructure. Not just decentralization.
But total autonomy, by design.

The future is self-owned, self-operated, and borderless.

Whether you’re an individual, an enterprise, or a grassroots community, the tools to build your own cloud, on your own terms, are finally here.

Start building sovereignty today.

Explore how Zero OS and the ThreeFold Grid can power your next step.