Dynamic Insight Beating AI News: Swiss AI lab Giotto has released two models, Alto and Solo, claiming to have created "the smartest sovereign AI in Europe." The official statement emphasizes that the models are hosted on European cloud infrastructure, with data remaining within Europe, avoiding reliance on U.S. cloud service providers.

However, shortly after release, users quickly uncovered issues. When testing Alto, the model claimed to be Kimi from China's Moonshot AI. Faced with community inquiries, Giotto co-founder and CEO Aldo Podestà admitted the model indeed inherited some weights from Kimi, but stressed the team had retrained it and incorporated their proprietary inference-time computation (TTC) technology. As for which version was used or exactly which weights were inherited, not a single word was disclosed.

Ironically, Podestà himself published an article titled "Open Source Is a Trojan Horse" on October 19, criticizing Europe for branding open-source models as "sovereign AI," often just another way of depending on others. Yet only eight days later, he ended up doing exactly the same thing—true hypocrisy.

This isn't the first time such incidents have occurred. On September 2, Spanish AI company Multiverse Computing promoted Quasar 438B as "Europe's leading AI model," without disclosing its underlying foundation in the initial announcement. It wasn't until the next day, in a technical article, that they revealed it was actually derived from compressed and optimized versions of Zhipu GLM-5.2.