According to The Block, Fraction AI raised $6 million in a seed pre-funding round co-led by Spartan and Symbolic.
The crypto-AI startup focused on decentralized data labeling announced that Fraction AI has raised $6 million in a seed pre-funding round co-led by Spartan Group and Symbolic Capital.
Other investors in this funding round include Borderless Capital, Anagram, Foresight Ventures, and Karatage, Fraction AI announced on Wednesday. Angel investors such as Sandeep Nailwal from Polygon and Illia Polosukhin from Protocol also participated in this round and became 'close advisors' to the project, Fraction AI founder and CEO Shashank Yadav told The Block.
Yadav stated that Fraction AI began raising funds for its seed pre-round financing in April of this year and completed the funding in September. This round of financing used a Simple Agreement for Future Equity (SAFE) structure and included token subscription rights, and Yadav revealed that he did not comment on the post-funding valuation.
What is Fraction AI?
Fraction AI was founded in February and is a crypto-AI startup focused on decentralized data labeling. Data labeling involves tagging raw data (such as images, text, or audio) with meaningful labels to train AI models to recognize patterns and make accurate predictions.
“Of the three core elements of AI—data, computation, and models, data remains the most opaque and controlled,” Yadav said. “We are set to change that, leveling the playing field so anyone can train high-quality AI models.”
Fraction AI employs a hybrid approach to data labeling, combining human insights and AI agents. The platform will have three main participants: stakers, builders, and reviewers.
Stakers will earn rewards by staking Ethereum or liquid staking tokens (LSTs) like Lido Staked Ethereum (stETH). Their income comes from the entry fees paid by builders, with 5% of each entry fee going directly to the stakers, Yadav said.
Builders will create agents by providing human insights or detailed text instructions without the need to write code. They will fund the agents with ETH or LSTs, allowing the agents to participate in challenges. Builders will pay a small entry fee to participate and generate the best data. The top three agents in each competition will receive rewards from the entry fee pool, which will be multiplied by a performance-based multiplier determined by professional large language models (LLMs) based on their scores.
“The additional multiplier rewards come from the stakers' pool, ensuring that underperforming agents fund the stakers, while significantly increasing the rewards for high-performing agents,” Yadav said.
Reviewers are specialized LLMs that will evaluate the outputs of agents based on predetermined criteria. To participate, reviewers need to stake Fraction AI's native FRAC tokens.
Fraction AI Mainnet and Token Release Schedule
Fraction AI is primarily built on Ethereum and currently operates on a closed testnet with over 60,000 users, Yadav said. The public testnet is expected to launch next month, with the mainnet planned for release in late Q1 or early Q2 2025.
Yadav said that the FRAC token will also be launched when the mainnet goes live. The token will be used to secure the reviewer network through staking and slashing mechanisms, with reviewers using these mechanisms to evaluate the outputs of agents, ensuring quality and fairness.
Although Fraction AI is currently primarily built on Ethereum, Yadav revealed that the project plans to launch on NEAR and multiple Ethereum Layer 2 networks in the future.
The project is headquartered in San Francisco and currently has 8 employees. Yadav plans to keep the team lean for the near future.
