OpenGradient raises $9.5M
OpenGradient announced $9.5 million in funding to build a network for open, verifiable decentralized AI focused on on‑chain inference and agent deployment. (x.com) The round highlights investor interest in decentralized architectures for hosting and verifying AI agents. (x.com)
OpenGradient said on April 14 it has raised $9.5 million to build a blockchain network for running and verifying artificial intelligence models and agents. (prnewswire.com) The New York company said a16z crypto backed the round, with Coinbase Ventures, SV Angel, Foresight Ventures, Pragma, SALT, Symbolic Capital, Canonical Crypto, Black Dragon, NEAR, Celestia, and Thanefield Capital also participating. OpenGradient also listed angel investors including Balaji Srinivasan, Illia Polosukhin, and Sandeep Nailwal. (prnewswire.com) OpenGradient says its network is designed to host models, run inference — the step where a trained model produces an answer — and deploy agents onchain with proofs that the computation happened as claimed. Its documentation describes the product as infrastructure for secure and verifiable artificial intelligence execution, model hosting, and application deployment. (opengradient.ai, docs.opengradient.ai) The pitch targets a problem that has become more visible as artificial intelligence agents spread into finance and software: most model outputs are generated inside closed systems that outside users cannot independently audit. OpenGradient says its network is built around “verifiable AI inference,” or cryptographic checks that let others confirm a model run without trusting a single operator. (prnewswire.com, opengradient.ai) The company has been building toward that pitch for more than a year. In October 2024, it announced an $8.5 million seed round for decentralized infrastructure for model hosting, secure inference, and onchain application deployment; the new announcement puts total funding at $9.5 million. (prnewswire.com, prnewswire.com) Since then, OpenGradient has rolled out pieces of that stack in public. Its site says the testnet is live, its March 2026 devnet post said developers could deploy models and run secure inference, and its portal shows more than 2 million inference transactions and more than 3,800 models. (opengradient.ai, opengradient.ai, portal.opengradient.ai) The company has also pushed beyond infrastructure into applications. Its portal lists BitQuant, a quantitative finance agent, and the company last year launched MemSync, a memory layer meant to preserve context across tools such as ChatGPT, Claude, and Perplexity. (portal.opengradient.ai, prnewswire.com) OpenGradient’s latest funding lands as crypto investors keep backing projects that try to move artificial intelligence workloads onto decentralized networks rather than centralized cloud providers. For OpenGradient, the next test is whether developers will use those verification features for real agents and applications, not just demos on a new chain. (prnewswire.com, docs.opengradient.ai)