美国和欧洲的 AI 新创公司与软体开发商为了降低高昂的运算成本，正逐渐从使用昂贵的美国顶尖 AI 模型（如 OpenAI 和 Anthropic）转向采用更便宜的中国 AI 模型（例如 DeepSeek、阿里巴巴等）。

中国模型如 DeepSeek 的使用量在开发者平台（如 Vercel 和 OpenRouter）上呈现爆发式增长，这表明中国在 AI 领域的竞争正透过高性价比和高采用率，在被忽略的「技术人员采用度」上迅速赶超美国。

虽然美国在技术研发与最强大模型的建构上仍保持领先，但历史表明开发者的习惯和依赖性具有长期优势，中国模型在初创生态系中的普及，正悄然侵蚀美国顶尖实验室的护城河。

To reduce skyrocketing computation costs, American and European AI startups and software developers are increasingly pivoting from expensive, frontier US AI models (like OpenAI and Anthropic) to much cheaper Chinese alternatives (such as DeepSeek and Alibaba).

The usage of Chinese models like DeepSeek has experienced explosive growth on developer platforms like Vercel and OpenRouter, indicating that China is rapidly gaining ground in the AI race through cost-effectiveness and developer adoption.

Although the US still leads in building the most powerful models, history shows that developer habits and dependencies carry long-term advantages, and the rising popularity of Chinese models among startups is quietly eroding the moat of leading US labs.