China’s Overlooked AI Strategy: Beijing Is Using Soft Power to Gain Global Dominance

Chinese firms DeepSeek and Moonshot AI are rapidly expanding their global AI footprint with state-of-the-art, low-cost open models like R1 and Kimi K2, demonstrating significant user adoption and customization, particularly in developing nations. This open-model approach allows China to gain substantial soft power and entrench its technology in global AI infrastructure, challenging the U.S.'s predominantly closed-model strategy and export control focus. The trend necessitates a strategic recalibration for U.S. policymakers and AI companies to promote their own open models, or risk ceding critical international AI influence and market share.

Analysis

A significant strategic shift is underway in the global AI landscape, driven by Chinese firms like DeepSeek and Moonshot AI, whose low-cost, customizable open-source models are achieving rapid global adoption. DeepSeek's R1 model, for example, saw its active user base triple to 97 million in just three months, while its derivative versions were downloaded 2.5 million times in a single month, underscoring intense demand for adaptable AI. This open-model strategy, focused on efficiency and real-world applications, directly challenges the predominant U.S. approach of developing powerful but proprietary, closed-source models. While U.S. export controls on advanced semiconductors have shaped China's focus, they have not prevented the development of competitive models that are closing the performance gap. This trend positions China to gain substantial soft power and entrench its technology in the AI infrastructure of emerging markets, supported by cloud investments from firms like Alibaba (BABA). In response, U.S. firms are beginning to pivot; OpenAI plans to release an open-source model and has cut prices by 80% on its o3 model, while Google (GOOGL) has released its Gemma family of models, though they currently lag in performance. The situation highlights Meta (META) as the current U.S. leader in the open-model space and creates policy uncertainty for chipmakers like Nvidia (NVDA) as the U.S. re-evaluates its export control strategy.

Key Decisions for Investors

• Investors should re-evaluate the competitive moats of U.S. AI leaders, favoring companies with a credible open-source strategy like Meta, while scrutinizing the reactive shifts by Google and OpenAI-backed Microsoft.
• Monitor U.S. policy regarding semiconductor export controls, as a potential strategic pivot towards restricting inference-specific chips could create new regulatory risks and opportunities for firms like Nvidia.
• Recognize the growing influence of Chinese technology in emerging markets; firms like Alibaba, which are building out foundational cloud infrastructure, are well-positioned to capitalize on the rapid adoption of China's open AI models.
• The shrinking performance gap between open and closed AI models suggests that proprietary technology may offer a less durable advantage than previously assumed, warranting a portfolio review to assess exposure to companies overly reliant on a closed-ecosystem strategy.