Shares of major European companies investing in AI, including Germany’s SAP and France’s Dassault Systemes, fell this week.

New, more advanced AI models sparked concerns about disruption in the software and analytics sectors.

The selloff followed the launch of OpenAI’s GPT-5 and Anthropic’s Claude for Financial Services, raising doubts about the future business models of firms such as the London Stock Exchange Group.

Since mid-July, shares in LSEG, Sage, and Capgemini have dropped 14.4%, 10.8%, and 12.3%, respectively.

They underperformed broader indexes like London’s FTSE 100 and Europe’s STOXX 600, which both rose over the same period.

🔗 Source: Reuters

🧠 Food for thought

1️⃣ European software firms enter AI disruption from a position of structural weakness

Europe’s software sector has long struggled with innovation gaps that now leave companies more exposed to AI disruption.

McKinsey research from 2019 found that only 1% of European firms qualify as “scale-ups” compared to significantly higher rates in the US, with European countries lagging in R&D investment compared to America and China 1.

This innovation deficit means European software companies like SAP and Dassault Systemes are defending market positions built on legacy advantages rather than cutting-edge innovation when facing AI threats.

The timing is particularly challenging because these companies command premium valuations. For example, SAP trades at 45 times earnings compared to the broader STOXX 600’s 17 times multiple, making them especially vulnerable to any narrative that questions their competitive moats [original article].

Germany, the UK, France, Italy, and Spain represent 86.4% of European software revenues, but this concentration in traditional software markets may increase vulnerability as AI models like GPT-5 and Claude begin automating tasks these companies have monetized for decades 2.

2️⃣ Early AI productivity data reveals the complexity behind disruption fears

Real-world studies of AI’s impact on software development show mixed results that help explain investor uncertainty about which companies will benefit or suffer.

A randomized controlled trial found that experienced developers using AI tools actually took 19% longer to complete coding tasks, contradicting initial expectations that AI would speed up work 3.