Many industry celebrities gather to understand the decentralized AI dark horse project Prime Intellect

Author: KarenZ, Foresight News

Against the backdrop of rapid development of AI technology, the centralization and high costs of computing resources are still the main bottlenecks in innovation and popularization. As a decentralized point-to-point computing power and intelligent protocol, Prime Intellect aims to provide a new path for open source intelligence by aggregating global distributed computing resources and promoting the development of collaborative AI models.

This article will introduce this project from three dimensions: investment background, team background and protocol operation mode.

Prime Intellect Team Background

Prime Intellect was founded in January 2024 by two co-founders Vincent Weisser and Johannes Hagemann.

CEO Vincent Weisser has relatively rich experience in Web3 and focuses mainly on the DeSci and AI fields. He is a co-sponsor of DeSci projects Bio Protocol, VitaDAO, and CryoDAO. He has previously served as the head of ecosystem and AI for DeSci Molecule. In addition, Vincent Weisser helped create the experimental community Zuzalu.city in 2023.

CTO Johannes Hagemann focuses on decentralized AI, semi-automation, energy optimization, clear technical optimism, brain-computer interfaces (BCIs), longevity and other fields. He worked as an AI research engineer at German AI system developer Aleph Alpha, and also served as a strategic consultant at VitaDAO from May 2021 to January 2023.

Financing background: Increased weights in the AI ​​field

Prime Intellect In April 2024, Prime Intellect completed a $5.5 million seed round, led by Distributed Global and CoinFund, with investments including Compound, Collab+Currency and Protocol Labs founder Juan Benet. In addition, the seed wheel has attracted well-known angel investors such as Clem Delangue, CEO of machine learning building tool Hugging Face.

At the beginning of 2025, Prime Intellect received another major financing progress, completing a new round of financing of US$15 million, led by Founders Fund and participated by Menlo Ventures and others. This round of financing attracted many heavyweights in the AI ​​and Web3 fields, including Andrej Karpathy (EurekaAI, Tesla, OpenAI), Clem Delangue (Chief CEO of Hugging Face), Dylan Patel (SemiAnalysis), Tri Dao (Chief Scientist of Generative AI Cloud Platform Together.AI), Balaji Srinivasan (Network School) and Emad Mostaque (Co-founder of Stability AI), Jake Medwell (Co-founder of 8VC), Brendan McCord (Cosmos Institute), Sandeep Nailwal (Polygon Co-founder) etc. After the completion of this round of financing, it raised a total of more than US$20 million.

It is worth mentioning that Prime Intellect investor Andrej Karpathy is one of the founding members of OpenAI. He has worked as a research scientist at OpenAI and has also served as the director of Tesla's artificial intelligence. Emad Mostaque was former CEO of Stability AI, and later resigned to focus on decentralized artificial intelligence and was also an advisor to Render, a supplier of decentralized GPU rendering solutions.

How to run?

Officially, Prime Intellect develops infrastructure and economic incentives based on all existing decentralized AI efforts to aggregate and coordinate global computing resources to achieve a truly sovereign open source AI ecosystem. Currently, the protocol is running on the internal test network and is fully integrated and launched in SYNTHETIC-1 operation. Prime Intellect has delivered several key jobs or products over the past year.

Compute Exchange: Aggregate, coordinate and schedule global GPU resources to integrate computing resources from data centers, cloud service providers and even personal GPU holders into a unified resource pool. Users can select the most cost-effective GPU resources based on chip type, quantity and rental time. This approach not only reduces the computational cost of AI training, but also promotes the democratization of AI technology through efficient supply and demand matching and improving resource utilization.

PRIME: A decentralized training framework that supports training large-scale models on global distributed computing resources, can achieve fault-tolerant training, and supports dynamic on/off of computing resources, and optimizes communication and routing in global distributed GPU networks.

In October 2024, Prime Intellect launched decentralized training INTELLECT-1 for a 10 billion parameter model. Leading open source AI participants such as Hugging Face, SemiAnalysis, Arcee, Hyperbolic, Olas, Akash, Schelling AI, etc. jointly contributed to this decentralized training operation.

GENESYS: Synthetic data generation framework that easily extends the open source library for synthetic data generation and verification as well as crowdsourcing tasks and validator calls. Prime Intellect announced in February 2025 that it will use DeepSeek-R1 to create a proven open source dataset, SYNTHETIC-1, an open source dataset for mathematical, coding and scientific inference trajectory.

TOPLOC: Prime Intellect introduces verifiable inference method TOPLOC, which uses compact locally sensitive hashing (LSH) technology to accurately detect unauthorized modifications to models, prompts or calculation accuracy. When generating a response, you need to perform a pre-filling on the input token once, and then perform multiple decoding on each new token generated. During verification, all tokens can be passed at once and pre-filled only once.

The Prime Intellect protocol forms a coordination layer that connects all of these components and provides them with economic incentives to attract the computing and capital needed to achieve scale.

The main components of the Prime Intellect protocol include Worker (users run software to contribute computing resources), validators, Orchestrator (the server used by the creator of the computing pool to manage the workers in the pool and coordinate the current workload running), and smart contracts.

Prime Intellect smart contracts are currently deployed on the Base Sepolia test network, but Prime Intellect plans to migrate to its own chain later. Prime Intellect's contract component RewardsDistributor also describes that the rewards of nodes can be calculated based on the active time in the computing power pool and tokens are allocated accordingly.

This means that Prime Intellect may issue tokens when it launches the Base mainnet or publishes its own chain. Prime Intellect plans to release a public testnet in the coming months, allowing anyone to contribute to computing power and make contributions without permission.

summary

Prime Intellect builds a sovereign open source AI ecosystem by aggregating large-scale computing power and providing a global GPU resource scheduling market, allowing anyone to create and contribute open models, agents, data sets, etc., and collectively obtain ownership of their contributions. This model is expected to drive global players to jointly build an open source AI model.

Prime Intellect says it is also exploring collective ownership and programmable licenses for models and agents to reward contributors and provide new revenue streams for open source AI.

As stated in its mission, Prime Intellect hopes to popularize AI development on a large scale through the power of decentralization, accelerate the progress of high-impact fields such as science, agency and code, and ultimately build an open, transparent and intelligent ecosystem that benefits more and more people.