Story’s in-house incubated project, Poseidon, a full-stack decentralized data layer developed for AI training, has secured USD 15 million (approximately KRW 20.8 billion) in seed funding led by a16z Crypto, the cryptocurrency arm of Andreessen Horowitz. Poseidon addresses intellectual property (IP) challenges by providing high-quality, IP-cleared data for robotics, multimodal models, and next-generation AI development.

As large-scale language models (LLMs) become open-source and AI computational resources are increasingly commoditized, data has emerged as the critical competitive edge for AI development. However, AI training data often faces limitations in legal transparency and quality. To tackle these issues, Story introduced Poseidon, a blockchain-based platform built on programmable IP licensing, enabling secure and transparent data exchange for AI developers and data contributors.

Poseidon leverages Story’s IP licensing infrastructure to track and monetize data exchanges between AI developers and contributors. Collected data undergoes explicit consent processes and is registered on the Story blockchain, receiving licenses for commercial use. This approach resolves persistent AI training challenges, such as IP disputes and the scarcity of long-tail data.

The platform specializes in collecting and processing hard-to-source datasets, such as first-person view (POV) videos for robot training and multilingual voice data, transforming them into high-quality resources for AI development. Poseidon streamlines data discovery, licensing, and integration to enhance AI usability. By partnering with Decentralized Physical Infrastructure Networks (DePIN), Poseidon secures sensor data and enables secondary uses like remixing, recycling, and royalty sharing. This summer, Poseidon will launch a preview for AI developers and data contributors, rolling out contributor modules, software development kits (SDKs), and integrated licensing tools.

Sandeep Chinchali, Story’s Chief AI Officer and Poseidon’s lead scientist, emphasized the platform’s focus on differentiated data: “AI’s competitive edge now lies in high-quality, unique data rather than model scale. Poseidon delivers IP-cleared, high-quality data to build robust AI systems for complex real-world environments, from eldercare robots to multicultural sentiment analysis and public movement simulators.”

Chris Dixon, founder and managing partner of a16z Crypto, highlighted Poseidon’s potential: “AI models have largely exhausted easily accessible training data. Poseidon offers a new standard for data infrastructure, fairly compensating creators and contributors while addressing bottlenecks in next-generation AI development.”

Poseidon’s team includes globally recognized AI and Web3 experts. Sandeep Chinchali, a former Stanford AI researcher and lead data scientist at Uhana (acquired by VMware), co-leads the project with Sarick Shah, a machine learning engineer with experience scaling AI products in finance, telecom, and healthcare. Seungyoon Lee, co-founder and CEO of PIP Labs (Story’s developer) and Poseidon’s chairman, oversees strategic business development.

Seungyoon Lee stated, “Poseidon marks the dawn of Story 2.0, validating our vision that IP and data are the core of the AI economy. Built on Story’s programmable IP layer and tamper-proof IP registry, Poseidon’s datasets eliminate legal risks and ownership disputes, accelerating innovation in the AI training ecosystem.”