Previously, Solana officially endorsed Gradient with two promotional tweets. The investor structure is very clean, with only three top institutions, Pantera, Multicoin, and Sequoia China, publicly announced. The seed round raised $10 million and has the potential to be listed on the platform. Among the many AI projects that have been hyped up, @ Gradient.HQ belongs to a relatively hardcore and focused team. Their recently released product Parallax is a decentralized and efficient LLM inference service, which focuses on breaking down LLM into multiple layers, allowing nodes with limited memory and computing power to participate in the inference process of large models across public network environments; And intelligently allocate the optimal computing path according to user needs, with better performance, cost, and privacy than traditional centralized data centers. Although the final presentation to the user is a simple dialog box, the actual experience details are very meticulous: during the model's response process, the right side will display in real-time which nodes are processing the current request simultaneously. Interested friends can experience the demo: https://chat.gradient.network/ . In addition to Parallax, Gradient also has the Echo distributed reinforcement learning framework, which completely separates inference and training, each performing its own duties, greatly improving efficiency and reducing costs. The Lattica point-to-point data communication protocol optimizes the data transmission path between nodes and feeds back the execution data to the Echo engine for further model parameter updates, forming a closed-loop decentralized infrastructure. Gradient Cloud packages multiple AI models into an out of the box service marketplace, where developers can access the models through APIs and pay per use to support the implementation of more AI applications. From the perspective of team and technical depth, Gradient is very hardcore. Its founder Eric comes from Sequoia, a top financial and technology composite background team in China, with top talents such as ACM Gold Medal and Tsinghua Yao Class members. Parallax has recently received endorsements from several top open-source AI labs, including SGLang, Kimi, Alibaba's Tongyi Qianwen, and Moon's Dark Side, demonstrating the hardcore power of distributed reasoning. Beyond technology, Gradient is pursuing its vision of decentralized AI: CZ previously emphasized that Al should be a public good and should not be monopolized by a few companies. Gradient significantly reduces the threshold for training and using AI, allowing everyone to train, expand, and deploy their own agents, whether it's personal trading agents, virtual companions, or personal memory management systems, all based on Gradient's Open Intelligence Stack. Both the Solana official and Binance ecosystem highly recognize DeAl, and the Gradient team's technical depth, investment background, and ecological resources make this project a benchmark in the DeAI field.