Joe Tsai Urges Europe to Leverage Industrial Data and Open Source for AI Sovereignty
Alibaba Chairman Joe Tsai argued at the Wave by Vento conference that Europe should build AI sovereignty through local compute infrastructure and proprietary industrial data, citing open source as the best defense against geopolitical supply disruptions. He contrasted China's focus on open models with US closed labs like OpenAI and Anthropic, criticizing the latter for ceasing academic publications to maintain secrecy.
Related Entities
Full Report
On the 7th, at the Wave by Vento conference in Turin, Italy, Joe Tsai stated: "No country wants to be entirely dependent on another nation's technology. Once a government changes or circumstances shift, the other side can simply cut off the supply of technology. Europe must seriously invest in computing infrastructure so that models can run locally, keeping both model training and inference on European soil."
In his view, while Chinese and American labs are competing fiercely to recruit AI researchers, Europe has yet to fully leverage their potential. He suggested that Europe should focus on training models with proprietary industrial data to build a competitive advantage. "Chinese factories possess incredibly valuable manufacturing data for model training. Many European countries still have a massive industrial base," he noted.
According to reports from IT Home, Tsai is optimistic about a new wave of emerging "cloud-native providers" in Europe—companies specializing in cloud services and operating data centers equipped with AI chips.
He also mentioned that China intends to prioritize open-source models and publicly available research to drive global adoption. In contrast, US companies like OpenAI and Anthropic primarily offer proprietary models.
Tsai was blunt in his assessment: "The closed-source labs in the US have stopped publishing papers because they don't want to share. They want to remain closed."