A curated list of 25 deep tech signals. Useful intelligence, not a promise.

The European Innovation Council has published the EIC Tech Report 2026, a horizon scanning exercise that distils 25 emerging deep tech signals from EIC portfolio evidence. The signals sit at low to mid maturity and may shape Europe’s innovation, industrial and market capabilities. They originate from 2021 to 2025 EIC data across funded projects and high quality unfunded proposals in Pathfinder, Transition and Accelerator. The analysis was guided by EIC Programme Managers and external experts with support from the European Commission’s Joint Research Centre. The report is positioned as structured intelligence that can inform strategy and foresight. It is explicitly not a prediction, ranking or a funding priority list.

The signals are organised under three domains aligned with the Strategic Technologies for Europe Platform. They reflect where the EIC is seeing technical movement among European teams. They also expose familiar European constraints. Many items need standards, certification pathways and scalable manufacturing that are still immature. Several rely on materials or components where the EU is strategically dependent. That is precisely why they appear here as signals rather than markets.

Scope at a glance

The report clusters signals in three domains. Examples include semiconductor materials, secure and distributed AI, quantum repeaters and in orbit servicing in digital and space. In clean and resource efficient technologies, signals include secondary raw materials recovery, electrochemical treatment for persistent water pollutants, advanced heat to electricity materials and energy active buildings. In biotechnologies and health, signals range from novel food production and computational protein design to advanced therapy manufacturing and next generation diagnostics and intervention devices.

How the EIC built the list

The EIC covers the innovation lifecycle from early science to market deployment. That vantage point generates portfolio data that can indicate where technical novelty is converging with potential relevance for resilience and strategic autonomy. The 2026 scan mined applications and funded projects between Q2 2021 and Q1 2025. The selection criteria emphasised breakthrough potential and likely relevance to EU resilience and autonomy objectives consistent with STEP. The Joint Research Centre supported the methodology, which combined portfolio data mining and scientometric analytics with expert panels steered by EIC Programme Managers.

Digital and space technologies: foundational layers under competition

2D materials for advanced memory and memristive devices

Why it could matter: Atomically thin materials can improve energy efficiency and density for memories and in memory computing, relevant to automotive, industrial automation and IoT. This aligns with the European Chips Act and STEP ambitions on advanced semiconductors. The sovereignty angle is clear. Europe wants options that reduce dependency at advanced nodes.

What it is: Two dimensional materials such as TMDs, h BN, metal oxides and van der Waals heterostructures can host resistive memories and memristors based on mechanisms like redox reactions, vacancy migration and ferroelectric switching. Emerging optoelectronic memristors and 2D spintronics open MRAM and FeRAM concepts in atomically thin stacks.

What is new: Reports of higher on off ratios, lower set voltages, faster switching and better endurance. Vertically stacked devices and dense crossbar arrays using MoS2, HfSe2 and h BN. Photon sensitive memristors and 2D ferroelectric behaviour extend function. The real test remains scalable, reliable integration in foundries.

Scalable MXene manufacturing for industrial electromagnetic applications

Why it could matter: High performance electromagnetic shielding without rare earth intensity would support telecommunications, automotive electronics and sensing. It also fits the Economic Security Strategy and the Critical Raw Materials Act by reducing dependencies.

What it is: MXenes are transition metal carbide, carbonitride and nitride nanosheets with metallic conductivity and tunable surface chemistry. They enable EMI shielding, energy storage, sensing and flexible electronics.

What is new: Kilogram scale production with better control of flake size and surface terminations. Films of 6 to 15 micrometres achieving order of magnitude higher shielding effectiveness at lower weight versus incumbents. Improved passivation for stability. Processing routes to transparent conductive films and flexible components. Computation assists structure property prediction. Industrial qualification and long term stability remain gating.

Quantum repeaters for trusted-node-free quantum networks

Why it could matter: Moving beyond trusted nodes is central to the European Quantum Communication Infrastructure strategy for long term sovereignty. Repeaters address distance limitations and could underpin entanglement based secure services for government and critical sectors.

What it is: Devices that create, store and connect entanglement across nodes using memories and entanglement swapping rather than amplifying signals. They are essential for long distance quantum key distribution without trusted relays.

What is new: Better quantum memories using rare earth doped crystals and cold atom ensembles with longer coherence and multiplexing. Advances in single photon sources, frequency conversion and integrated photonics. One way repeater concepts with photonic cluster states and error correction promise higher rates and simpler hardware. Integration with software defined control planes is emerging. Europe will need interoperable components and certification methods.

Embedded Zero Trust architectures for distributed and federated AI

Why it could matter: Distributed AI that does not require centralising sensitive data lines up with the AI Act and EU cybersecurity policy. Hardware anchored security with post quantum cryptography addresses Cyber Resilience Act objectives and reduces reliance on non European cloud stacks.

What it is: Zero Trust assumes no component is trusted and requires continuous verification across chips, software, networks and users. For AI, that means secure execution fabrics for federated training and inference with isolation, attestation, encryption and scalable key management.

What is new: Co design of AI accelerators and runtimes with embedded isolation and attestation from memory to interconnect to network. Post quantum key management introduces latency and scaling trade offs. Orchestration layers are adapting to privacy preserving computation and AI driven threat detection. The complexity lies in integration and assurance, not point features.

Bio-inspired AI for self-organising and resource-efficient systems

Why it could matter: Scaling based AI faces efficiency and controllability limits. Bio inspired methods aim for learning efficiency and adaptive structure which aligns with EU goals on trustworthy AI. They could reduce dependence on very large proprietary models.

What it is: Approaches that bring mechanisms from neuroscience, cognitive science and evolution into artificial systems. Focus areas include synaptic plasticity, temporal learning, structural diversity and continual learning.

What is new: Greater use of formal principles such as the Free Energy Principle, evolutionary search for architectures and new temporal learning insights using eligibility traces and reward signals. Cognitive evaluation protocols probe planning and causal reasoning in large models. Interpretability and verification remain hard as systems become more dynamic.

Embodied AI for adaptive agents in open environments

Why it could matter: Foundation models for robotics and embodied AI will be shaped by whoever controls environments and simulation platforms. Europe will need its own training and evaluation infrastructures for physical AI to avoid dependency on opaque benchmarks.

What it is: Agents that link perception, priors and action in three dimensional, dynamic settings, including simulated and hybrid digital twin environments. They require continual decision making, long horizon planning and social interaction.

What is new: Shift from task specific training to environment centric strategies that combine large multimodal priors, world models and online adaptation with hierarchical memory. Intrinsic motivation and memory based exploration support zero and few shot generalisation. Lifecycle assurance and safety in human environments remain open.

Edge computing for scalable and loss-tolerant satellite operations

Why it could matter: European constellations and deep space missions cannot rely on continuous ground control. Onboard processing supports operational continuity, space safety and debris mitigation and aligns with the EU Space Programme and the Security and Defence Space Strategy.

What it is: Processing data on orbit to filter, classify and decide locally under latency and bandwidth constraints. Relevant for Earth observation, debris avoidance, swarm coordination and on orbit servicing.

What is new: Elevating onboard processing from pre processing to mission critical autonomy at constellation scale. This demands new verification, responsibility and behaviour safeguards in shared orbits.

Graphene coatings and composites for performance-critical space systems

Why it could matter: Mass reduction, thermal control and multifunctional structures are core spacecraft constraints. European capacity to produce and qualify graphene for space strengthens supply resilience and supports STEP on advanced materials.

What it is: Graphene in thin films, coatings and as fillers in composites offers high strength, thermal and electrical conductivity and barrier properties for structures, radiators, antennae and radiation protection.

What is new: Maturing production via liquid phase exfoliation, CVD and mechanical methods with better consistency. Shift from lab to system level integration is beginning. Qualification for space environments and repeatability are the hurdles.

In-space servicing robotics for maintenance and reuse

Why it could matter: Non serviceable, non recyclable spacecraft are a strategic vulnerability in congested orbits. Autonomous inspection, repair and end of life handling support resilience and space safety. This links to EU thinking on space traffic management.

What it is: Robotic manipulation, docking, tool use and assembly in microgravity, with potential to reuse materials from defunct satellites and upper stages. Coupled with in space assembly and manufacturing and in situ resource use.

What is new: Moving from demos to repeatable maintenance and reuse concepts. Exploring clustered manipulators and additive manufacturing methods like direct energy deposition in orbit. Most enabling technologies are low to mid readiness with autonomy and qualification as key gaps.

Clean and resource efficient technologies: closing loops under real constraints

Microbial biomining for secondary metal recovery and bioremediation

Why it could matter: Access to critical metals increasingly depends on extracting from low grade ores, tailings and waste streams. This intersects with the Critical Raw Materials Act and the Circular Economy Action Plan.

What it is: Extremophilic microbes oxidise metal sulphides under very low pH, releasing metal ions. Acidophilic microalgae and bacteria can capture and remediate in harsh industrial sites. Targets include copper, nickel, cobalt, gold, lithium, zinc and more.

What is new: Extending to e waste, industrial residues and wastewater with a focus on selectivity and efficiency using synthetic biology, adaptive evolution and de novo enzyme design. Concepts tested under microgravity and Mars analogs underline long term interest, but industrial economics and permitting will decide adoption.

Capacitive deionization for low-energy water treatment

Why it could matter: Brackish water and variable effluents need low energy, modular treatment in a more water scarce Europe. This aligns with the European Water Resilience Strategy and the Water Framework Directive.

What it is: Applying low voltage across high surface area electrodes to adsorb ions into electrical double layers then regenerating by reversing or removing the voltage. Operates without high hydraulic pressure and can be selective.

What is new: Better electrodes using carbon nanotubes and graphene, membrane CDI and flow electrode configurations for continuous operation and redox active materials for higher selectivity. Increasing reach to higher salinity streams, but robust long term operation and fouling control at scale remain decisive.

Electrochemical destruction of persistent water contaminants

Why it could matter: PFAS and micro or nanoplastics are difficult to remove and pose health risks. Electrochemical destruction supports Drinking Water Directive compliance and REACH restrictions.

What it is: Ambient temperature electrochemical systems designed to defluorinate PFAS and degrade plastics in water rather than transferring pollution to secondary waste.

What is new: Advanced three dimensional graphene based electrodes to enhance electrocatalysis while limiting by products like chlorinated species. The promise is modular and reagent free operation, but energy intensity at trace concentrations and by product management need independent validation.

Advanced thermoelectrics for low and mid grade waste heat

Why it could matter: Converting low temperature waste heat to electricity would help meet Energy Efficiency Directive targets. Reducing reliance on tellurium or lead aligns with STEP and material criticality goals.

What it is: Heat to power via the Seebeck and Nernst effects with performance set by the figure of merit zT. Materials include Bi2Te3, skutterudites, half Heuslers, magnesium compounds and organic systems.

What is new: Incremental zT gains via doping, nanostructuring and interface engineering plus exploration of topological and correlated materials for unconventional transport. Organic thermoelectrics advance for very low grade heat. The bottleneck is durable module engineering and cost effective integration.

Spin-caloritronic materials for heat to electricity and sensing

Why it could matter: Spin based thermal effects diversify options for low grade heat harvesting and precise sensing where classical thermoelectrics underperform. This remains research grade but could open niche applications.

What it is: Using spin currents and magnons induced by temperature gradients such as the spin Seebeck and anomalous Nernst effects to generate electrical signals without relying on high charge mobility.

What is new: Three dimensional composites like YIG with platinum for better conductivity and robustness. Magnetic doping and phase engineering to amplify transverse thermoelectric responses. Flexible and printable devices are being explored. Efficiencies are modest and mechanisms still being clarified.

Inverse design with digital twins for predictive materials manufacturing

Why it could matter: AI driven inverse design can shorten discovery to production cycles for materials tailored to EU regulatory and environmental contexts. This supports the Net Zero Industry Act and the Circular Economy Action Plan.

What it is: Generative models, topology optimisation and physics informed neural networks search material formulations for specific targets then validate in digital twins of buildings and infrastructures under realistic operating conditions.

What is new: Closing the loop between design, manufacturing constraints and system level performance. This can reduce trial and error and support transparent traceability. The dependency on high quality data, interoperable models and industry adoption is significant.

Passive cooling and gravity storage for energy-active buildings

Why it could matter: Reducing peak cooling demand during heatwaves and smoothing local energy use supports EPBD revisions and the Renovation Wave. Reduced dependence on imported HVAC equipment and batteries has autonomy angles.

What it is: Nanophotonic coatings that reflect solar radiation and emit heat through the atmospheric window for passive daytime radiative cooling, combined with gravity based storage that lifts and lowers masses in buildings or shafts to store energy mechanically.

What is new: Practical coupling of passive cooling materials with building scale mechanical storage, with optional thermoelectric harvesting of waste heat and personal thermal management. AI assisted inverse design accelerates tuning to local climates. Urban deployment will face heritage, safety and cost constraints.

Biotechnologies and health: translation, manufacturing and regulation front and centre

Mycelium-based hybrid fermentation for whole-food production

Why it could matter: Whole food outputs from locally available residues reduce reliance on imported proteins and align with the Bioeconomy Strategy and Farm to Fork. The approach also touches industrial biotech and materials.

What it is: Fungal mycelia used as a structural scaffold with bacteria, yeasts, algae or plant components in one bioconversion process to create whole food products with complex texture and functionality. Uses liquid fermentation, 3D structuring, bioinks and fogponic inspired cultivation with AI assisted control.

What is new: Moving from single organism mycoprotein toward engineered multi kingdom consortia and transferring fabrication methods from materials to food. Regulatory classification, safety assessment and consumer transparency will be decisive in the EU’s Novel Foods regime.

Biotech-enabled perennial crops for regenerative systems

Why it could matter: Perennial grains can reduce inputs and soil disturbance, supporting CAP, the Soil Strategy and Farm to Fork while lowering reliance on volatile fertilisers.

What it is: De novo domestication of wild perennials and wide hybridisation to obtain multi year harvests. Perennial rice shows what is possible, but major cereals face chromosomal barriers.

What is new: Genomic selection, TILLING, FIND IT and AI assisted breeding can accelerate domestication timelines and trait stacking. EU policy on new genomic techniques and coexistence frameworks will shape deployment.

Novel microbiome therapeutics

Why it could matter: Shifting toward preventive and personalised interventions aligns with the Pharmaceutical Strategy and could leverage Europe’s strengths in cohorts and regulated clinical research.

What it is: Ecosystem therapies such as live biotherapeutic consortia, engineered strains, bacteriophages and metabolite based interventions that reshape microbial communities to improve health.

What is new: High throughput cultivation unlocks previously unculturable microbes. Spatially resolved sampling informs site specific functions. Engineered sense and respond strains and colon targeted delivery enter trials beyond GI disorders. Manufacturing standards and regulatory clarity for live products are core issues.

Computational protein design for faster drugs and enzymes

Why it could matter: AI assisted protein engineering can reduce cycle times for therapeutics, vaccines and industrial enzymes. This strengthens technology sovereignty in biomanufacturing.

What it is: ML models plus biophysical insight to design sequences that fold into desired structures and functions. The SKYCovione vaccine based on a designed nanoparticle signals regulatory viability.

What is new: Integrated design build test cycles combining deep learning, rapid expression and screening. The frontier is moving from structure to function prediction with iterative experimental feedback. Open, interoperable toolchains will matter for EU skills and diffusion.

Automated manufacturing for scalable CAR immune cell therapies

Why it could matter: Europe needs reliable, cost effective and decentralised manufacturing for ATMPs to meet supply security and access goals under the Pharmaceutical Strategy and STEP priorities on cell and tissue engineering.

What it is: Chimeric Antigen Receptor T cell therapies are approved in the EU for blood cancers but are constrained by variable starting material and centralised, labour intensive production.

What is new: Miniaturised, automated, semi or fully closed platforms that support near clinic production. Process analytics, standardised vectors and cell engineering strategies improve consistency and safety. Vector supply, comparability and quality by design remain regulatory priorities in the EU.

Other health signals to track

The list also includes biohybrid microrobots for cellular scale interventions, autonomous robotic systems for integrated surgical workflows, noninvasive and minimally invasive brain interfaces for adaptive therapeutic modulation and portable ultra low field MRI for distributed clinical uses. Each of these will require strong evidence on safety, workflow integration and cost effectiveness in EU health systems.

Context from inside the programme, with a note of caution

EIC Programme Managers frame these signals as levers to keep Europe in the game while paradigms are fluid. Isabel Obieta argues that supporting proof of concept developments in novel semiconductor materials and architectures can help retain critical know how in Europe as device paradigms shift. Samira Nik says quantum architectures are fluid and early positioning in high uncertainty components may determine Europe’s role in configuring secure quantum infrastructures. Hedi Karray highlights that novel AI architectures are proliferating and that Europe should preserve experimental diversity while embedding responsible AI principles before standards consolidate. Stela Tkatchova points to next generation space infrastructure requiring onboard autonomy, innovative materials and in space robotics to improve resilience. Carina Faber underlines that biological and electrochemical conversion approaches could expand domestic recovery and clean up capacities. Paolo Bondavalli stresses early performance assessment of advanced materials in emerging energy systems. Franc Mouwen ties digital twins and inverse design to measurable lifecycle emission reductions in construction. Ivan Stefanic frames agrifood deep tech as part of a broader resilience and health agenda. Orsolya Simmons sees convergence in computational design, programmable biology and advanced bioprocessing and argues that public funding can bridge high risk experimentation to industrial and clinical platforms.

These statements describe intent more than outcomes. The gap between portfolio level promise and market level proof still includes reproducibility, scale up economics, certification and global competition. The report’s authors are careful to call these signals and not priorities. That caveat should guide interpretation.

Key EU policy linkages and the hard parts ahead

The signals align, on paper, with several EU policy instruments. Examples include the Chips Act, EuroQCI and the EU Quantum Strategy, the AI Act, the Cyber Resilience Act, the Critical Raw Materials Act, the European Water Resilience Strategy, the Energy Efficiency Directive and EPBD, the Farm to Fork and Bioeconomy strategies and the Pharmaceutical Strategy and ATMP framework. Alignment does not guarantee delivery. Progress depends on standards and certification from ETSI and CEN CENELEC, open testing and evaluation infrastructure, patient procurement and long term finance. Value chains for advanced materials, semiconductors and bioprocess inputs remain exposed to non EU suppliers. Skills and talent mobility will be a binding constraint in several of these areas.

Publication details and reuse

EIC Tech Report 2026. General publication dated 30 March 2026. English PDF of approximately 3.21 MB. Reuse is authorised with attribution under the European Commission reuse policy. The views expressed do not necessarily reflect the Commission position.

What to watch next

Signals become trajectories when three things happen. First, credible independent validation of performance and durability in operational settings. Second, evidence of scalable, bankable manufacturing and service models in Europe. Third, standards, certification and regulatory pathways that remove adoption friction. If the next EIC scans track those three, then this year’s promising list will have matured into a strategy conversation about industrial capacity, not just technology potential.