Building an RL gym for training on-chain agents
We're building the infrastructure for verifiable reinforcement learning that will enable a new generation of reliable, trustworthy AI agents.
AI is failing us when it matters most. Despite impressive capabilities, today's systems can't be trusted in critical scenarios because there's no way to verify their training. They hallucinate, make errors in reasoning, and lack accountability—fundamentally limiting their real-world utility.
This verification gap is the single biggest barrier preventing AI deployment in medicine, finance, legal systems, and other high-stakes domains. When lives, money, or critical decisions are on the line, how can you trust an AI that was trained in a black box with no verification?
OpenBlock Labs is solving this by creating training environments with cryptographic certainty—where every reward signal, every data point, and every interaction is verifiably correct. Using blockchain's inherent properties of public data, shared state, and immutability, we create training grounds where AI systems can learn from transparent, provable interactions.
This isn't just a technical improvement—it's the difference between AI systems you can use for generating social media posts versus ones you can trust to diagnose disease, manage financial systems, or handle critical infrastructure. Cryptographic certainty means AI that can be audited, verified, and held accountable.
The market implications are profound. By solving the trust problem, we unlock trillions in economic value across every sector that requires high-confidence AI. Companies and institutions that have been hesitant to deploy AI in mission-critical applications will finally have systems they can rely on.
Our vision is to transform AI from impressive but unreliable technology into systems with verifiable guarantees. Through blockchain-powered reinforcement learning, we're building the foundation for AI agents that can be trusted with our most important decisions.
We're looking for exceptional engineers and researchers to join us in building this future.
High-impact roles for exceptional builders.
Creating the blockchain-powered gym for training AI through verifiable reinforcement learning
Blockchain-Verified Agent Training
OBL.dev is our first agent — an autonomous data analyst trained in blockchain environments that leverage four unique properties: rich public datasets, multi-agent interactions, deterministic execution, and verifiable reward functions.
Massive public datasets that capture real economic behaviors and interactions
Shared global state with many users interacting on the same database simultaneously
Cryptographically verifiable reward functions that guarantee training integrity
Advancing agent intelligence through reinforcement learning and verifiable data
We're OpenBlock Labs
A team of researchers and engineers at the intersection of AI and blockchain. Our mission is to build intelligent autonomous agents through reinforcement learning, with blockchains serving as verifiable environments for data validation and training.
Blockchains are uniquely suited to serve as the perfect training gym for AI. They provide massive public datasets, shared global state across many users, deterministic execution environments, and cryptographically verifiable rewards—all essential properties for advanced reinforcement learning. These characteristics enable us to train agents in environments where actions and outcomes are transparent, rewards are provable, and the complexity of real-world interactions is authentically represented. No other technology combines these properties in a way that solves the fundamental challenges of building trustworthy AI systems.
Backed by top investors, building from San Francisco
We've raised $20M from Foundation Capital, Amplify Ventures, Electric Capital, and prominent angels from both the AI and blockchain spaces.
We're in-person in a cozy office in San Francisco, focused on building a small, high-caliber team with expertise in AI, NLP, information retrieval, and blockchain technologies.