The Chips Joint Undertaking has opened two linked funding calls with a combined indicative EU budget of €80 million, aimed at pushing European AI chip designs from taped-out silicon to benchmarked, rack-level pilot deployments in EU compute infrastructure. Both calls now close Oct. 27, 2026 at 5:00 p.m. Brussels time, after a deadline extension.

The two calls sit within a broader initiative built on two operational streams. Stream A covers European AI chip and rack development through Topics A1 and A2. Stream B, under Topic B, provides the common evaluation and procurement framework. Only A1 and Topic B are currently open; A2, the rack-development phase, is not scheduled to launch in 2026.

€10M for AI chip demonstrators

The first call, "AI chip demonstrators for EU compute infrastructure" (DIGITAL-JU-CHIPS-2026-AI1-SG), is a Simple Grant with an indicative EU budget of €10 million. It funds projects that bring advanced European AI chip designs to the demonstrator stage. The Chips JU estimates an EU contribution of up to €1 million per project, though proposals may request a different amount.

The eligibility bar is deliberately high. Each proposal must include at least one EU-established AI chip design company responsible for the chip design and its associated intellectual property. System integrators and OEMs can join where needed for integration, software bring-up and benchmarking. Applicants must show that the AI chip has already taped out at a commercial foundry, with engineering samples available or expected within the specified timeframe. A PCB hosting the chip must be in production or expected during the six-month development phase.

Funding covers completion of the demonstrator — chip, PCB, cooling and interconnect — plus software integration and preparation for benchmarking. After development, demonstrators will undergo up to two months of performance-based assessment on a common evaluation platform.

That assessment carries consequences. The best-performing solutions can receive the AI Compute Excellence Label, which is a strict eligibility condition for the subsequent Topic A2 rack-development phase. In effect, the label is the gate between demonstrating a chip and getting it built into racks.

€70M for evaluation and procurement

The second call, "AI compute evaluation and deployment platform for EU infrastructure" (DIGITAL-JU-CHIPS-2026-AI-GfP), is a Grant for Procurement with an indicative EU budget of €70 million — seven times the size of the demonstrator call. It will establish the common evaluation infrastructure and services needed to validate Stream A systems against workloads and benchmarks, and it will procure pilot AI compute racks.

The call targets a buyers' group of EU-established legal entities. Proposals must include at least one contracting authority or entity operating an EU AI Factory or comparable EU public AI/HPC facility, and at least two EU-established cloud service providers. The consortium must span beneficiaries from at least three different Member States or EEA countries — a structure designed to make the procurement genuinely European rather than national.

The action runs through three operationally linked phases over an expected 42 months: first, benchmarking Topic A1 chip demonstrators; second, preparing and conducting the benchmark evaluation of Topic A2 rack systems; and third, procuring, installing and optimizing selected rack systems. The procurement will focus on scaled, multi-rack clusters rather than isolated single-rack deployments, so that distributed AI workloads can be tested under realistic conditions.

A connected pathway, not two isolated grants

The two calls are operationally linked rather than parallel. Demonstrators developed under the first call will be evaluated through the common platform established under the second, creating a pathway toward rack-level testing and pilot deployment. The target audience spans the AI compute value chain: AI chip design companies, system developers, AI Factories, public AI/HPC infrastructure operators and cloud service providers.

For chip designers, A1 offers funding to develop and benchmark advanced demonstrators and, for selected solutions, a route toward subsequent rack development. For AI Factories, public facilities and cloud providers, Topic B funds both the evaluation environment and the buyer framework, and pays for pilot AI compute systems.

The overall pathway runs from chip to cluster: European AI chips reach demonstrator stage, pass common benchmarking, progress to rack development and system-level testing, and then move into procurement and pilot deployment in European AI compute infrastructure.

Applications go through the EU Funding & Tenders Portal, with call details available on the Chips JU website. Whether the €10 million demonstrator budget is enough to move multiple European AI chip designs through the Excellence Label gate — and into contention for the €70 million procurement that follows — will shape how credible the EU's homegrown AI compute supply chain becomes.