Four projects with great potential in the direction of AI decentralized computing power 0G Labs @0G_labs 0G Labs is a decentralized AI operating system focused on AI L1 blockchain. Financing scale: Total financing of approximately 75 million US dollars+node sales of 35 million US dollars. Interpretation: Comprehensive AI L1, but also decentralized computing power, published papers and models, but did not see anything related to open weights and code. The quality of the paper is quite good. There are many ways to participate, so I won't go into detail. Prime Intellect @PrimeIntellect Prime Intellect is a decentralized AI development platform dedicated to large-scale democratization of AI development. Aggregate global computing resources to assist researchers in collaboratively training state-of-the-art models. Multiple research papers have been published, as well as open weights, training datasets, and open code for research, supporting distributed reinforcement learning (such as the recently released INTELLECT-2 model). Financing of $20.5 million, led by Distributed Global and CoinFund. Interpretation: The research ability in this field is on par with DeepMind in this direction, and the research team is also the most promising team. But there are relatively few points to participate in, and the team is not more inclined towards web3 operations, but more towards computing power business operations. You can look for more angles. Especially in terms of node computing power contribution. The team whose technical ability I have the most confidence in. Open sourced multiple components https://(((github.com)))/PrimeIntellect-ai Gensyn AI @gensynai Project Introduction: Gensyn is a machine learning computing protocol for building machine intelligence networks. It connects all machine learning compatible computing hardware worldwide, provides decentralized computing resources, supports the training and development of deep learning models, and aims to enable machine intelligence and human intelligence to thrive together. Financing: The total financing is approximately $50 million led by a16z, with participants including Nat Friedman, Elad Gil, and SV Angel Interpretation: The core focuses on the training direction of reinforcement learning RL, with in-depth research on white papers and open-source code, demonstrating technical capabilities. Https://((GitHub))/gensyn ai/rl warm can participate in node contributions, whether there will be airdrops will be studied by everyone themselves. Pluralis Research @PluralisHQ Pluralis Research focuses on Protocol Learning, which is a decentralized, communication efficient model parallel training framework for basic models. Financing of $7.6 million led by USV and CoinFund Interpretation: Open research papers and codes, mainly on asynchronous parallel training (asynchronous is suitable for Internet computer instability and other characteristics) ICML 2025 paper "Nesterov Method for Asynchronous Pipeline Parallel Optimization" http://arxiv.org/abs/2505.01099 https://(((github.com)))/PluralisResearch/AsyncPP