Why do American technology companies strive for development space for open source models (which account for the vast majority in China)? The biggest news last night was undoubtedly that Lao Huang @ Jensen Huang opened his X account for the first time. The first tweet was an open letter initiated by a16z and jointly signed by Nvidia, Microsoft, PLTR, PerplexityAI, Olama and other technology companies, advocating that "cutting-edge open source models and cutting-edge closed source models should coexist and develop". Everyone is talking about the open letter itself, but there is not much discussion about the motives and intentions behind it. From a personal perspective, the core reason is the following image and the content discussed in the previous long article that was cited: 1. The proportion of Chinese open source models used by American companies is approaching 60% per week, and open source models account for such a large proportion of token consumption in American companies. Once banned or restricted, the cost for companies will increase dramatically. 2. In this long article in mid July (the cited article is), it is discussed that the adoption of enterprise AI has entered the era of engineering, and the adoption of AI in the enterprise market is increasingly becoming a stage of segmented application models 1) Continue to use high-value tokens from closed source cutting-edge models in high-value scenarios; 2) In high-frequency, low complexity, and high volume task scenarios, enterprises are increasingly inclined towards high cost-effectiveness tokens (mostly derived from open-source models). 3. This trend is inevitable: For chip manufacturers, it is necessary to train and reason with whoever's model they use; For cloud vendors, the more model choices they offer to enterprise customers, the higher their gross profit margin; For deployment layers (such as PLTR) or application-oriented companies, there is a natural motivation to choose cost-effective tokens. The joint ventures of this open letter are basically these three categories, so there is actually a strong commercial demand and huge commercial value behind the open letter. 4. Authropic's call in Congress to restrict or ban the development of open-source models in China is essentially a clear indication of the threat and impact of this trend. Kimi K3 is just a catalyst. Prior to K3, the share of China's open source in token consumption had been rapidly soaring, and even without K3, Authropic had the motivation to do so. This article is sponsored by @ bitget_zh, titled 'Bitget Buying US Stocks: Instant Entry, Smooth Trading'