Thing 1∙2∙3 Foundation Grant Program

This BMFTR funding guideline supports Leitprojekte - flagship, consortium-scale projects - on the use of artificial intelligence across industrial value chains, and it is one of the more substantial German AI calls currently open. The scope is deliberately narrower than general industrial AI: it targets the development of domain-specific AI models and systems, meaning the adaptation or creation of specialised architectures that encode industry-specific vocabulary, process knowledge and physical principles rather than the application of general-purpose foundation models off the shelf. Projects must address three things together. First, vertical anchoring in a real value chain, so the work is demonstrated in the context of an actual industrial sequence rather than a single firm's process. Second, data infrastructure integration with existing industrial IT systems, which is where most industrial AI pilots fail and where BMFTR is explicitly asking for engineering rather than research handwaving. Third, the establishment of open standards for data accessibility and governance across sectors, which signals that BMFTR is funding this as pre-competitive infrastructure and will favour consortia that commit to interoperable, publishable standards over proprietary stacks. Funding goes to consortia combining research institutions and industry. Project sketches (Projektskizzen) are due 1 November 2026, ahead of a second-stage full application for selected consortia - the standard two-stage German process, in which the sketch is the real competition.

Apply AI: AI-Driven Robotics for Industry is the robotics deployment topic in Horizon Europe's 2027 Cluster 4 digital and emerging technologies calls, sitting under the Commission's Apply AI strand and the AI, Data and Robotics partnership. It is an Innovation Action rather than a Research and Innovation Action, and that distinction is the most important thing to understand about it: Innovation Actions fund work close to deployment - prototyping, testing, demonstration, piloting and validation in real industrial settings - not foundational research. A consortium proposing to advance the state of the art in robot learning algorithms is pitching to the wrong instrument and should look at the RIA topics in the same call, such as the human-AI collaboration topic. What fits here is taking AI-driven robotics capability that already works in the lab and proving it out on real factory and industrial floors with industrial partners in the consortium. The indicative EU contribution is around 18 million euros per project, which places it among the larger single-project awards in the work programme and implies a substantial multi-partner consortium. Standard Horizon Europe consortium rules apply: at minimum three independent legal entities from three different EU Member States or Horizon Europe associated countries, with at least one established in a Member State. The deadline is 18 March 2027, part of a cluster of Cluster 4 digital and emerging topics closing the same day, including the EU Frontier AI Initiative at around 44 million euros per project and the Challenge-Driven AI Innovation Booster at around 14 million euros. Full topic conditions, the destination text and submission are on the EU Funding and Tenders Portal.

CyberAICorps Scholarship for Service (NSF 26-503) is NSF's response to what the solicitation calls a talent shortfall in artificial intelligence and cybersecurity, and it is worth understanding that CyberAI here means two distinct things: using AI in cybersecurity operations, and securing the AI systems themselves. Proposals should be explicit about which they address. The program runs two tracks with very different economics. The Scholarship Track, capped at $2,500,000, funds student scholarships that carry a post-graduation government service obligation - scholars complete formal degree programs with integrated AI and cybersecurity coursework and participate in internships and mentored research. The Innovation Track, capped at $500,000, is the education-research channel: curricula, teaching methods, faculty professional development, and communities of practice in AI and cybersecurity education, judged substantially on potential for wide adoption beyond the applicant institution. Both tracks have a $300,000 floor, and NSF anticipates up to 25 projects per fiscal year. Eligibility differs by track in a way that matters for community colleges: under the Scholarship Track they are eligible only as sub-awardees of four-year institutions, whereas under the Innovation Track they may lead. Scholars themselves must be U.S. citizens or lawful permanent residents. The Scholarship Track deadline recurs on the third Tuesday in July annually - the 21 July 2026 deadline has passed, so the next full proposal deadline falls in July 2027, with an Innovation Track target date in early April.