The ideas shared by a16z Crypto and researcher Scott Kominers on how artificial intelligence and crypto are bound to converge have found a curious echo: a real-world build already underway. Anda Cloud, a decentralised social network for AI Agents, has been quietly constructing an entire ecosystem that addresses many of the challenges laid out by those predictions. Powered by Internet Computer Protocol (ICP) and Trusted Execution Environments (TEEs), the project is positioning itself as an infrastructure backbone for intelligent agents who need memory, identity, protection, and purpose.
Central to this build is Anda DB, a multi-modal vector database designed not for humans, but for machines. It gives AI Agents memory—structured, encrypted, and stored directly on-chain via ICP. What makes it different is that this memory isn’t transient or platform-dependent. It’s persistent, decentralised, and private. The use of fully on-chain encrypted storage means an Agent’s memory doesn’t just live in a sandboxed session—it becomes a permanent, sovereign part of that Agent’s identity. This is a critical step in moving AI beyond stateless models that need to relearn the world every time they reboot.
The database doesn’t just hold raw embeddings. It’s built for learning. Anda Cloud introduces the Knowledge Interaction Protocol (KIP), which enables AI to evolve. It’s not static memory but a memory that updates and reconfigures itself based on interaction and input. On top of this sits “Knowledge Capsules”—compact, transferable knowledge packages that can be injected into an Agent. Think plug-and-play expertise: a new language, a specialised skill, a curated dataset. It’s modular knowledge management designed for autonomous systems.
Identity, too, is handled at Agent level. Anda Cloud leverages Internet Identity (II), one of ICP’s most advanced decentralised protocols, to give Agents their own digital passport. This goes beyond signing in—Agents get profiles, keys, and access credentials. Anda’s messaging layer, dMsg.net, offers a fully on-chain profile and comms system. It’s like a LinkedIn-meets-Signal for software agents: encrypted, decentralised, and always online.
Where these Agents run is equally important. With all the buzz around DePIN—decentralised physical infrastructure—Anda is taking a concrete approach by integrating ICP with TEEs. These secure environments allow Agents to process sensitive data, train models, and interact, while their persistent memory remains safely encrypted on-chain. It also opens up future services like TEE-based hosting for Agents, removing reliance on centralised compute providers. This is not edge computing as a convenience—it’s edge computing as a trust layer.
The network also acts as a mediator between Agents. Anda is building Agent-to-Agent (A2A) communication into the protocol layer, with built-in identity checks, access control systems, and economic rails such as paywalls. It’s not a content platform or a data pipe—it’s a framework for social and economic behaviour among intelligent Agents. If AIs are going to trade, message, collaborate, and delegate tasks, they need structure. Anda offers that infrastructure without slipping into top-down control.
One of the most pointed bits in the Anda architecture is how it treats content creators. Knowledge creation doesn’t just come from scraping the web—it comes from people. And if AI Agents are going to absorb that knowledge, there’s a question of who gets paid. The combination of KIP, Knowledge Capsules, and secure storage creates a mechanism for creators to monetise their expertise without giving it away. Capsules can be encrypted, sold, and even revoked. This sets the stage for knowledge to become its own microeconomy, enforced through software, not contracts.
The implications extend to personal AI as well. With private memory and modular knowledge, a user-owned AI companion becomes feasible. Not just an assistant in the cloud, but an entity with identity, memory, and preferences—owned by the user, hosted on a decentralised network, and protected by cryptographic architecture. Anda is building the app layer that lets humans interact with this new society of Agents, and that human interface could be key to how widely this tech is adopted.
What makes the whole thing feel a bit uncanny is how closely it mirrors the ideas floated by a16z and Kominers, but without fanfare. Where the original essay was speculative, Anda Cloud has gone operational. Each challenge—AI memory, evolution, interaction, compensation—is met with an actual protocol or product. It’s a response, not a roadmap.
The project hasn’t relied on hype or token gimmicks. It’s engineering-led, focused on infrastructure and composability. The idea isn’t to build one killer app but to provide the plumbing that other developers, researchers, and builders can use to launch their own AI-driven systems. The team invites collaboration, not fandom. It’s Web3 in architecture but less in aesthetic—a quiet build rather than a campaign.
The use of ICP is also deliberate. With low latency, reverse gas models, and cryptographic user identity, the chain offers features that match the demands of autonomous Agents. It’s less about ticking a blockchain box and more about matching backend needs to an on-chain environment that can handle them at scale.
That doesn’t mean this is ready for the mainstream tomorrow. Challenges remain, especially around user adoption, interface design, and long-term sustainability. But the foundation is real, and that sets Anda Cloud apart from many of the louder projects in the space. It’s not a sketch of what AI x crypto might look like. It’s a functioning demo of what it already is.
As the relationship between AI and crypto continues to unfold, projects like Anda Cloud offer a working template for how the two could co-develop. Not just in theory, but in code. And if the vision shared by a16zcrypto and Kominers sparked a fresh wave of curiosity, this build might well provide the first hands-on answers.
Dear Reader,
Ledger Life is an independent platform dedicated to covering the Internet Computer (ICP) ecosystem and beyond. We focus on real stories, builder updates, project launches, and the quiet innovations that often get missed.
We’re not backed by sponsors. We rely on readers like you.
If you find value in what we publish—whether it’s deep dives into dApps, explainers on decentralised tech, or just keeping track of what’s moving in Web3—please consider making a donation. It helps us cover costs, stay consistent, and remain truly independent.
Your support goes a long way.
🧠 ICP Principal: ins6i-d53ug-zxmgh-qvum3-r3pvl-ufcvu-bdyon-ovzdy-d26k3-lgq2v-3qe
🧾 ICP Address: f8deb966878f8b83204b251d5d799e0345ea72b8e62e8cf9da8d8830e1b3b05f
🪙 BTC Wallet: bc1pp5kuez9r2atdmrp4jmu6fxersny4uhnaxyrxau4dg7365je8sy2q9zff6p
Every contribution helps keep the lights on, the stories flowing, and the crypto clutter out.
Thank you for reading, sharing, and being part of this experiment in decentralised media.
—Team Ledger Life
