SAP sees rising client scrutiny over AI model origins

Highlights

SAP CFO Dominik Asam reports increasing customer scrutiny of AI model origins, reflecting growing compliance, data governance, and geopolitical risk concerns in procurement decisions.
SAP embeds over 70 third-party AI models in its enterprise software and faces varying client acceptance, with stricter requirements for US defence clients and lighter standards for small businesses.
European debates on US cloud 'kill switches' are driving demand for sovereign alternatives, prompting SAP to sign deals with governments and defence contractors and broaden its AI sourcing across regions.

Client demand shifts AI sourcing priorities

As reported by Reuters, SAP Chief Financial Officer Dominik Asam says customers are increasingly examining the origin of AI models embedded in enterprise software as compliance, data governance and geopolitical risks gain weight in buying decisions.

Asam says SAP regularly hears questions from clients about whether data used by a model could be routed to places it should not reach. He adds that managing several AI models has made it more complex to understand each company’s exact technology governance requirements.

SAP, Europe’s most valuable technology company, embeds more than 70 models from multiple providers into software used for accounting and supply chain operations. The company has chosen to remain agnostic across the market rather than build its own models internally, but Asam says customer acceptance varies depending on where those AI components originate.

A U.S. customer may reject components developed with Chinese models, while some less sensitive back-office work, including at SAP itself, can run on open-weight software from a wider range of sources, he says. He contrasts the stricter requirements of defence-related clients with the lighter demands of smaller businesses.

Sovereign cloud debate shapes enterprise demand

Asam says European discussion around a possible "kill switch" in U.S. cloud infrastructure is also supporting demand for sovereign alternatives. He says SAP has recently signed deals with governments and defence contractors and wants access to models from every region, including Europe and Canada, so it can offer those clients a broader set of options.

On cybersecurity, Asam says SAP is using frontier AI models to attack and test its own products before external actors can do so. While declining to discuss the findings of those internal screenings, he says any company claiming to have no vulnerabilities would not be telling the truth.

In our earlier article on Cohere’s plan to acquire AI cybersecurity startup AISLE as part of its tie-up with Aleph Alpha, we outlined how the deal would add AI-driven software vulnerability detection and expand the group’s European footprint via a Prague engineering hub. We also noted AISLE’s early vulnerability research results and the focus on private, enterprise and public-sector deployments aimed at strengthening security and threat detection.

This material may contain third-party opinions, none of the data and information on this webpage constitutes investment advice according to our Disclaimer. While we adhere to strict Editorial Integrity, this post may contain references to products from our partners.