A European policy worker and a German AI founder's diverging views on China's DeepSeek and America's continued AI dominance highlight Europe's struggle to keep pace with the global AI race.

What happened: In January 2025, Chinese firm DeepSeek released a new AI model called R1 that is cheap and effective, with publicly available weights that anyone can run. Days later, OpenAI released o3-mini, which observers regard as more impressive than R1. European policymakers seized on R1 as proof that cutting-edge AI can be built without Silicon Valley's vast resources, while Silicon Valley largely remained unfazed and continued its massive AI investments.

Why it matters: The article presents a contrast between how European and American tech leaders interpret the same AI developments. European policymakers saw DeepSeek's R1 as a hopeful sign that Europe could catch up without needing the hundreds of billions of dollars pouring into American data centres. However, voices in Silicon Valley, including Christian Vogt (a German AI founder who relocated there), view the real story differently—that reasoning models work and that American progress is still rapid. This misalignment may shape Europe's strategic decisions about AI investment and competitiveness.

What to watch: The article frames this moment (January–February 2025) as the beginning of a five-year scenario extending through March 2031. Caroline Dubois, a French policy worker at the European Commission's digital technology office, is uncertain which interpretation is correct. Her internal conflict—whether to trust cautious voices about DeepSeek or Christian's Silicon Valley perspective—suggests Europe's policy response in the coming months will be critical to whether it can shape its own AI future.