Officials caution this is not a finished product and outcomes depend on delivering verifiably better, manufacturable chips.

The two-stage evaluation will, over twenty months, prioritize energy efficiency and performance gains, with potential standardization of AI-assisted design if results justify it.

Germany and the Netherlands will join forces, via SPRIND and NADI, to accelerate AI chip design as Europe seeks greater independence from US expertise.

A €40 million, 20-month program will fund multi-team efforts to develop AI-assisted chip design workflows aimed at rapid prototyping.

Success will be judged against clear criteria, including workload definitions, defined comparisons, software and manufacturing plans, and potential customer interest to validate prototypes.

The collaboration targets orders-of-magnitude faster design timelines for chips used in training and inference of AI models.

Stage 1, running Nov 2026 to Jul 2027, will select seven teams each awarded €2.6 million; Stage 2 in 2027–2028 will shortlist three teams for €7 million each plus coaching.

Jano Costard of SPRIND says the effort aims for more than incremental gains and seeks a strategic lead for Europe.

Program logistics: registration runs from September 23 to November 30; seven teams receive €2.6 million each, and eight months later the top three receive €7 million each plus mentorship.

Seven teams will start with €2.6 million apiece to develop concepts, advancing to three finalists who receive €7 million and additional support.

SPRIND uses a staged challenge model, offering coaching and collaboration opportunities beyond funding as teams progress through milestones.

The initiative targets two-stage activities around model training and inference workloads, encouraging rethink of core design logic to cut energy use while boosting performance.