Europe's AI scene is buzzing with plans to create a homegrown rival to China's DeepSeek, especially as US-Europe relations strain amid Trump-era tensions over trade, NATO, and tech regs. US firms dominate the whole AI stack, from chips to models, grabbing most investments and leaving Europe hooked on their services, much like cloud computing dependency. But officials in the UK and EU aren't backing down; they've poured hundreds of millions into funds and deregulation to build local capabilities.
The spark? DeepSeek showed you don't need massive GPU farms or endless cash, just smart model designs and open collaboration. Take Germany's LS Research and the SOOFI project: they're teaming up on a 100-billion-parameter open-source language model, set to drop next year. Assumptions here lean on shared compute, transparent research, and tailoring for European languages like those in Apertus or GPT-NL projects. Hardware-wise, it's not about matching Nvidia's scale but efficient setups, dodging the energy-hungry mega-clusters. Experts like Oxford's Rosaria Taddeo say geopolitics flipped the script, making sovereignty urgent, while skeptics warn winner-takes-all dynamics favor top players.
Still, profs like Wolfgang Nejdl bet on catching up through openness: "We'll be the European DeepSeek." No firm timelines beyond SOOFI's 2027-ish launch, but momentum builds with AI factories and gigafactories in EU plans. Challenges? Vague policies on self-sufficiency levels and market openness versus protectionism. Overall, it's a pragmatic push, balancing collab with caution.
