At Rise of AI 2025 in Berlin, a panel moderated by Stefanie Söhnchen of PIABO Communications, with Vanessa Cann (managing director for data and AI at Accenture and managing director of the German AI Association), Markus Hacker (senior regional director enterprise business DACH at Nvidia) and Carsten Kraus (founder and CEO of Casablanca AI), discusses whether Germany can overcome its reputation for AI pessimism. The panelists argue against constant comparison with the US and China, with Kraus urging Germany to move past a cultural 'angst' that slows adoption of imperfect but working technology, and Hacker urging a focus on domestic strengths such as DeepL, Black Forest Labs and Nvidia's digital-twin work with Sparkassen Finanzinformatik and KION. Cann reports that international investors increasingly see Europe as regulating rather than innovating, citing the EU AI Act's reputation, but also cites growth of the German AI Association (KI Bundesverband) to over 500 members and a shift among startups toward vertical SaaS. The panel identifies a persistent gap between academic research and commercialization, a lack of incentives for bold risk-taking inside companies and universities, and a cultural discomfort with visible success, and calls for more academia-industry connections modeled on Munich's UnternehmerTUM.
Key takeaways
- Vanessa Cann reports that Silicon Valley investors increasingly see Europe as regulating rather than innovating in AI, partly due to the EU AI Act's reputation.
- Markus Hacker cites DeepL as rated the third most popular AI tool globally by ZDNet, ahead of Gemini, Dall-E and Perplexity.
- Markus Hacker says Sparkassen Finanzinformatik is deploying 50 AI use cases across 200,000 employees and 50 million customers in its own sovereign data centers.
- Vanessa Cann says the German AI Association (KI Bundesverband) recently passed 500 member companies, with over 600 AI startups in Germany overall.
- Carsten Kraus argues breakthrough innovation cannot be incentivized through KPIs and instead requires removing bureaucratic and cultural barriers to give people freedom.
Detailed notes
Overcoming German grumpiness and angst
Söhnchen opens by naming the panel's premise: countering the narrative that Germany is falling behind in AI. Hacker argues the country should stop constantly comparing itself to the US and China on metrics like GPU counts and training efficiency, comparing the habit to his young children measuring who has more juice, and calls comparison 'the root of unhappiness.' Kraus reframes the goal as overcoming German angst rather than grumpiness, using the historical example of muskets that exploded in one of every 200 uses during the Thirty Years' War to argue that adopting an imperfect but critical technology was still necessary; he compares this to China deploying AI doctors in hospitals despite occasional errors, and argues Germany should turn its angst into FOMO instead.
Where Germany is already strong
Cann reports that on a recent trip to San Francisco, investors questioned whether Europe is a place to build or invest in AI companies at all, a narrative she ties to the EU AI Act creating an impression that 'the USA is innovating, Europe is regulating, and China might be copying.' She counters with growth figures: the KI Bundesverband (German AI Association) recently passed 500 member companies, and Germany has more than 600 AI startups, increasingly focused on vertical SaaS rather than general-purpose applications. Hacker names DeepL, rated by ZDNet as the third most popular AI tool globally and the top-rated AI translation tool ahead of Google Translate, and Black Forest Labs' image-generation models as German successes. Cann adds an Accenture-Nvidia project with a company called KION, building a digital twin of a shop floor on Nvidia Omniverse to simulate robot movement seconds before it happens.
From pilots to enterprise-wide deployment
Cann says most DAX-40 companies she works with already have collections of 200 to 400 AI use cases; the bottleneck is not ideas but change management, as companies shift from siloed business-unit or IT-led AI efforts toward a hub-and-spoke model. Hacker cites Sparkassen Informatik, now Finanzinformatik, as a company seriously scaling AI: it is rolling out 50 use cases across 200,000 employees and 50 million customers inside its own sovereign data centers. He argues the German Mittelstand, hidden champions with billions in revenue but reluctance to put proprietary IP into public clouds, needs comparable support. Kraus recommends a concrete practice: in any meeting longer than an hour, reserve ten minutes for every participant to propose an AI-based solution, so the habit spreads beyond the usual few people who already think about it.
The gap between academia and industry
Cann argues Germany's basic AI research, including work behind the transformer paper and Stable Diffusion, is strong but rarely translates into commercial value locally, and calls for closer, more practical academia-industry links, pointing to Munich's UnternehmerTUM as a model that Berlin has not matched. Hacker agrees research is fragmented across federal states and adds that researchers lack incentives to transfer projects into products once grant funding ends. Kraus separates two functions of university: training the workforce that companies need, versus producing people who invent something new, for whom the key resource is not curriculum but a peer group of similarly ambitious people to exchange ideas with, which he thinks universities underinvest in.
Incentives, failure culture and jealousy of success
Cann says employees are rarely incentivized to take bold, potentially failing action, and often receive contradictory instructions when new AI tools are rolled out, leaving them afraid to use them. Hacker urges copying what already works and shifting from a 'yes, but' mindset to 'why not,' referencing a DeepL culture of embracing failure described earlier that day. Kraus argues real breakthrough innovation cannot be incentivized through KPIs at all, since nobody could have set an incentive for Einstein to invent relativity; what is needed instead is removing bureaucratic and cultural barriers so motivated people can work freely, as he did by refusing homework from eighth grade onward. He closes by naming a further obstacle: Germans who become visibly successful or wealthy often face negative press and social suspicion, which he links to rising political polarization, and argues society needs to stop treating financial success as automatically suspect.
Quotes
Comparing is at the root of unhappiness.
Markus Hacker
I think the most important thing is not to overcome grumpiness, but to overcome German angst.
Carsten Kraus
The USA is innovating, Europe is regulating, and China might be copying.
Vanessa Cann
Questions answered in this talk
Why do international investors see Europe as behind in AI?
According to Vanessa Cann at Rise of AI 2025, investors in San Francisco increasingly believe Europe regulates rather than innovates in AI, a perception she ties to the global reputation of the EU AI Act; she describes being shocked by this and argues the AI ecosystem needs to actively counter it, pointing to Germany's growing number of AI startups as evidence.
What is holding back enterprise AI adoption in large German companies?
According to Vanessa Cann, most DAX-40 companies already have hundreds of AI use cases identified; the bottleneck is not technology or ideas but change management, organizational silos and how business units and IT departments coordinate on AI.
What German AI companies were named as success stories at Rise of AI 2025?
According to Markus Hacker, DeepL was rated the third most popular AI tool globally by ZDNet and the top AI translation tool ahead of Google Translate, and Black Forest Labs was named for its image-generation models, both cited as German-headquartered successes.