David Solomon at Goldman Sachs and Ted Pick at Morgan Stanley have both raised the same concern in recent commentary: the AI infrastructure build-out depends on a circular funding model where tech companies buy from each other while betting on future enterprise demand that has not yet materialised at scale. Solomon called it a question of whether the spending can continue without clearer returns. Pick pointed to the gap between infrastructure investment and actual productivity gains.
A Frankfurt-based asset manager rewrote a data platform role specification twice during a recent search. The original version asked for Databricks and Spark experience. The second added regulatory reporting. The final version required the candidate to present quarterly to the CFO on how the platform reduced operational risk. The role kept expanding because the hiring team could not justify pure infrastructure headcount to senior leadership without a direct line to business outcomes.
Searches in Zurich and Munich follow the same pattern. Companies want ML operations capability alongside the ability to explain model outputs to a trading desk or walk a compliance officer through an anomaly detection workflow. A data architect who has worked inside a financial services risk function moves through the process faster than someone with equivalent technical credentials from a different sector.
The searches that close quickest right now combine technical depth with business fluency. An ML engineer who can tie model performance to revenue impact, or a data engineer who has built pipelines inside a regulatory reporting workflow, attracts multiple offers. Pure infrastructure specialists still attract interest, and companies are asking for evidence that the person can connect their work to outcomes the CFO cares about.
Prompted by reporting from Diginomica.