What happened
At HumanX Amsterdam 2026, Intellyx's Jason Bloomberg named 10 European startups — among them Distil labs, LangWatch, Orq.ai, Softr and UnfoldAI — as vendors closing the gap between AI ambitions and production.
Why it matters
The picks point to tooling for running AI in production, such as LangWatch's agent testing and Distil labs' specialized models that cut token costs versus large language models, suggesting buyers may prize deployment over model hype.
What to watch
Bloomberg notes every European AI conference also hosts U.S. and other non-European vendors, so the all-European list is his judgment rather than a regional rule. No dates, prices or funding for these vendors are given.
WHO IT HITSThis matters most to enterprise AI buyers and IT leaders evaluating vendors for production deployments, and to European AI startups courting them. Investors tracking early-stage AI tooling may also take note.
Summaries like this, in your inbox every morning.
The list arrives amid a wave of attention on frontier-model competition between the U.S. and China, but the columnist argues that is a distraction. Instead, he points to European vendors whose products solve narrower operational problems: tracking AI lineage (8wave Oy), automating legacy software maintenance (Bosun B.V.), capturing tribal knowledge (Clarifeye SAS), testing AI agents (LangWatch), governing agent development (Orq.ai), and connecting closed data systems (UnfoldAI).
A recurring theme is making AI deployable rather than merely impressive. Softr's survival is framed as a case study: generative AI undercut the low-code/no-code market, but Softr adapted by providing back-end infrastructure that safely runs vibe-coded applications. Elsewhere, bots test bots (LangWatch), and models are shrunk to cut costs (Distil labs).
The stakes hinge on whether European sovereignty requirements keep pushing local AI innovation independent of U.S. and Chinese frontier-model battles, and whether enterprises adopt these production-focused tools rather than waiting for bigger models to solve deployment problems.
For example, today's edition would include:
AI-summarized, only the topics you pick: one digest a day via Email, LINE, or Slack.
Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →
Ask AI anything about this article. The AI reads this article, earlier AIToday articles, and Wikipedia, and cites its sources. Q&As are published on this page for other readers too.
CoreWeave completed the industry's first bring-up and validation of Nvidia Vera Rubin NVL72 on CoreWeave Cloud…
Collibra BV unveiled Live Map, Maestro, Guardian Agents and Agent Contracts — runtime governance tools meant t…
Google Cloud revenue grew 82% year over year, and CEO Sundar Pichai highlighted a $514 billion backlog
Nscale, a British AI neocloud, announced $3.36 billion in convertible financing
Runway launched its first AI filmmaking course in 2024 at NYU’s Martin Scorsese Virtual Production Center, and…
McDonald's unveiled "Archy," an English- and Spanish-speaking AI drive-thru system it says exceeds 90% accurac…
