Pond Raises $7.5 Million for Decentralized AI Model Layer

Pond, a crypto startup focusing on a decentralized AI model layer, has secured $7.5 million in seed funding led by Archetype.

Investors

Other investors in this round include Cyber Fund, Delphi Ventures, Coinbase Ventures, Near Foundation, and angel investors such as:
– Illia Polosukhin, co-founder of Near Protocol
– Chris Yin, CEO and co-founder of Plume Network
– Cynthia Wu, founding partner and COO of Matrixport
– Tesa Ho, head of market research at Flashbots

Pond closed this funding round, which began in April and concluded in July, with a simple agreement for future equity (SAFE) structure including token warrants. Co-founder and CEO Dylan Zhang declined to disclose the post-round valuation.

What is Pond?

Founded in 2023, Pond initially developed a user search engine using on-chain data that helps users explore blockchain connections. For instance, users can enter Ethereum co-founder Vitalik Buterin's ENS (vitalik.eth) to see who is connected within their network. Zhang mentioned that this was made possible through a proprietary graph algorithm tailored for blockchain data, calculating the strength of connections.

As time progressed, Pond recognized the need for a unified graph network across various types of on-chain data—social, financial, etc.—and pivoted to create crypto-native AI models for broader applications. Co-founder and CTO Bill Shi stated that in the AI industry, models serve as productivity engines, extracting value from data to make predictions. He cited the current AI wave being unlocked by models like GPT and Llama.

Pond aims to develop a comprehensive model ecosystem that covers data, model computation, training, and inference. The company plans not only to build its own models but also to assist others with model creation and commercialization.

User Perspective

While web2 AI models, like ChatGPT, aid in various tasks, Pond's web3 AI models focus on crypto-specific applications driven by on-chain data. Shi noted that on-chain data is expansive and messy, and AI can convert this complexity into understandable information, allowing users to maximize its utility.

Currently, Pond supports applications in the following areas:
– Security: A graph-based model predicts malicious addresses with accuracy of 0.936 and precision of 0.935, which will be utilized by an industry-leading security firm.
– Recommendations: Models suggest NFTs on Zora and DeFi protocols on Gearbox.
– DeFi: A dynamic fee model helps liquidity providers optimize transaction fees. This model can increase fees to compensate for potential losses and lower fees to encourage beneficial trades.

Pond intends to broaden its crypto-AI applications to include DeFi risk management, insider trading detection, and personalized recommendations. Currently, the team comprises 11 members and is actively hiring for roles such as chief operating officer, product manager, and ecosystem growth manager.

The Funding newsletter: Stay updated on the latest crypto funding news and trends with my bimonthly newsletter, The Funding. It's free. Sign up here!