UpTrajectory Review

The EU's special envoy for industrial AI, Jim Hagemann Snabe, used a private seminar with the 27 European commissioners to make an unfashionable argument: stop obsessing over who builds the biggest AI models and start obsessing over what AI actually does for Europe's dismal productivity numbers. The remarks were reported by Politico's Zoya Sheftalovich and picked up by The Next Web. The available excerpt is thin — it gives us the setting, the messenger, and the core message, then cuts off — but the framing alone tells you this is an intervention aimed directly at the commission's instinct to treat AI as a prestige race against the US and China rather than a tool for factories, logistics firms, and mid-sized manufacturers.

For a small-business operator, this matters because the EU's policy choices over the next year will decide whether AI shows up in your business as something usable or as something expensive. If Brussels follows Snabe's logic, expect funding, standards work, and compliance guidance to tilt toward applied AI — automation of back-office work, quality control, forecasting — the stuff that helps a 30-person firm compete. If it reverts to the model-building race, the money and attention flow to a handful of well-capitalized labs and hyperscalers, and the practical tools a small operator needs arrive late, priced for enterprises, and wrapped in compliance overhead designed for companies with legal departments.

What is genuinely notable here is who is making the argument and where. Snabe is not a critic on the outside; he is the envoy, delivering the message in a closed-door setting to the commissioners themselves, which usually signals a message the leadership wants aired but not yet owned publicly. We agree with the substance — Europe's productivity gap is the real economic wound, and model parity with OpenAI or Google was never a realistic near-term goal. Where we are skeptical: 'focus on productivity' is the kind of phrase every commissioner can applaud while disagreeing entirely about what it means in budget terms. Until there is money attached, it is positioning, not policy.

The second-order effects cut in two directions. Applied-AI emphasis would be good news for vertical software vendors, systems integrators, and the small manufacturers who adopt early — and mildly bad news for European foundation-model startups hoping for state-backed scale. It also reframes the AI Act debate: if the commission's own envoy is saying adoption over architecture, the compliance regime's drag on deployment becomes harder to defend internally. Watch for tension between the envoy's productivity message and the act's risk-based paperwork, which falls hardest on exactly the small firms least equipped to absorb it.

What to watch next: whether the commission's upcoming AI and competitiveness communications cite adoption and diffusion metrics — SME uptake, sector deployment — or revert to compute and model benchmarks. If you run a small business, the practical move now is to pilot one narrow, measurable AI use case in your own operations, because the policy wind is shifting toward rewarding firms that can show productivity gains rather than firms that can talk about them. The full Politico reporting behind this excerpt likely covers the commissioners' reactions; that is where the real signal will be.

“Racing to build the best models matters less, he told them.” — The Next Web

Takeaway: Pilot one measurable AI use case in your operations now — EU policy is shifting toward rewarding adoption and productivity gains, not model-building.