**[NVIDIA's Open-Source Model Strategy Unveiled: Shifting from Selling Shovels to Mining Gold]** BlockBeats News, August 23: Analyst qinbafrank published an article analyzing NVIDIA's latest deal with Poolside. NVIDIA has paid $6 billion for a non-exclusive license to Poolside's "Model Factory" technology, the core software system behind Poolside's Laguna series of open-source coding models. At the same time, NVIDIA has extended offers to over 100 Poolside engineers, who will join NVIDIA to focus on the Nemotron open-source weight model project. Since its launch in 2023, Nemotron has been developing larger and more advanced versions, rumored to have parameter scales reaching the trillion level, alongside approximately $1 billion in equity investment support. The analyst views this as a landmark move for NVIDIA, signaling a shift from "selling shovels" to "mining gold." Jensen Huang is leveraging this strategy to delve deeper into the model layer, building a full-stack "AI factory" capability that is deeply integrated with NVIDIA's ecosystem. The analyst outlined four key points: 1. NVIDIA aims to rebuild one of the world's strongest open-source weight models, benchmarking against China's open-source models while directly challenging U.S.-based closed-source frontrunners like OpenAI and Anthropic. Open-source weight models are less costly to operate and easier to customize. 2. By extending from selling GPU hardware to software and model layers, NVIDIA strengthens its control over the entire supply chain. 3. While this deepens NVIDIA's moat, it also intensifies competition with its customers. Enhanced open-source model capabilities will drive GPU demand, but Nemotron directly competes with the closed-source models of major clients, potentially creating tensions. 4. The trend can be summarized as major model companies developing their own chips, while chip companies develop their own models, further escalating competition. The analyst noted that Jensen Huang's public endorsement of open-source at the end of July was, in fact, paving the way for NVIDIA's own open-source models. This deal marks a critical step in NVIDIA's efforts in the U.S.-China open-source competition. If the new version of Nemotron catches up in performance, it will further benefit cloud providers and software companies that manage customer workflow data. [Original Article Link]