SAP And AI Sovereignty
I watched the SAP (NYSE: SAP) keynote presentations from Sapphire Orlando to prepare for my visit to Sapphire Madrid. I should have realized that the keynotes in Madrid would be the same. And they were, except for ONE thing. SAP highlighted its European AI partners in Madrid and discussed AI sovereignty.
In Orlando, Anthropic was presented as a key partner and was given a speaking slot. At the Madrid event, Mistral and n8n were given time on stage to tout their capabilities.
In Phillip Herzig’s keynote, SAP’s chief technology officer said before introducing their European AI partners: “We are taking a look at sovereignty end-to-end with different levels of control and resilience.” Last year, SAP introduced the EU AI Cloud, which has “frontier models being served end-to-end out of European data centers.” The full Mistral platform is now generally available with sovereign model access and visual AI workflow orchestration inside Joule Studio.
Joule Studio is the software giant’s low-code/pro-code development environment for building, deploying, and managing custom AI agents and skills for Joule , SAP's enterprise AI copilot.
Journalists and analysts were given the opportunity to talk to SAP’s key executives in a question-and-answer forum. I asked about these sovereignty issues. What was driving this?
Christian Klein, chief executive officer, mentioned that many of their European customers are in the public sector or regulated industries. “They are concerned about the geopolitical situation in the world.” You never know what might happen tomorrow. “What if there are sanctions and you cannot extract data from one country to another? Or what if someone says this LLM (large language model), the latest version, cannot go to this country?”
Many of their European customers are concerned. “And yes, there is a high demand in the market for going more European.” Mr. Herzig added that ongoing deglobalization drives that demand. But sovereignty does not just apply to Europe, it applies to Canada, Singapore, Australia, and other countries.
Mr. Klein did say, “You will never have complete independence on the hardware side. But where it really matters is that the data is stored here and that there are European AI companies.”
In a one-on-one conversation, Jonathan von Rueden, the head of AI innovation at SAP, said that while Anthropic and OpenAI get most of the media attention, the capabilities of Mistral and n8n are strong, though perhaps not as robust as Anthropic's. Further, model development moves so fast that who is currently leading may not be particularly important depending upon the problem that needs solving. And the European AI is significantly less expensive!
At Sapphire Orlando, SAP introduced the Autonomous Enterprise as the next stage of its ERP ecosystem. The company’s core message was that AI agents would move ERP beyond systems of record into systems of execution, with agents embedded across finance, procurement, supply chain, HR, and customer experience.
At Sapphire Madrid, that message was supplemented with a discussion about sovereignty, trust, and the importance of a European AI infrastructure. SAP positioned the Autonomous Enterprise not only as an execution model for ERP but also as a response to the regulatory, geopolitical, and data-control concerns now shaping AI adoption in Europe and other regions.
Loading article...