Recently, I have been chatting with many friends who work on AI and Web3 infrastructure, and everyone has a very direct feeling: the current AI models are really outdated. Open source models, various large models, and vertical agents emerge every day, but the common pain points we face are actually very down-to-earth: it is difficult to monetize good models, and the surrounding ecosystem is too scattered. If you are a model developer, besides working on several centralized giant platforms, it is difficult to directly reach the demand side and obtain stable returns; And if you are a developer or node, it is difficult to find stable, reliable, and incentivized inference services in a decentralized network. I just saw the official announcement of the launch of the Model Marketplace by @ dgrid.ai, and there is a very clear sentence in the original tweet: The next wave of AI value will not only come from better models, but from better infrastructure around them. ” DGrid's approach this time is very practical, directly opening up a commercial path for model providers and developers: "List Your Model. Set Your Price. Earn On Main. Its way of solving problems is not to build castles in the air, but to integrate the three things of "model discovery+service quality+user preferences" into a mutually interlocking closed loop: AI Arena: Stop using rigid technology rankings as a standard. AI Arena is playing with the "blind test preferences" of real users, using signals thrown by real users to determine model discovery and routing selection. ⚙️ PoQ mechanism: What is the biggest fear of decentralized reasoning? I am most afraid of node disconnections and unstable services. The PoQ mechanism is specifically designed to handle quality verification and long-term incentives for inference services, enabling stable and efficient services to receive long-term on chain value distribution. Based on DGrid's previously released six-month 23M ARR data, this is not just an "air narrative" that talks about vision, but truly a deep binding of model supply, inference services, and actual needs. In the upcoming TGE warm-up period, this growth signal supported by real business closed loop is worth paying special attention to. Opening up AI indeed requires an open market.