EIC ePitching on AI and Software: eight deep tech teams, strong investor turnout and a win for quantum-inspired model compression

Brussels, July 17th 2025

Summary

›The EIC held an online ePitching for AI and software on 8 July 2025, featuring eight EIC-backed deep tech companies and 26 investors, generating 30 plus follow-up meetings.
›Investors voted Multiverse Computing as the best pitch with claims of 4 to 12 times faster model inference and 50 to 80 percent savings in operating cost and energy.
›Axelera AI placed second, pitching Europe-first semiconductor chips that the company says deliver 10x faster inference and 3 to 5x better energy efficiency.
›Presentations spanned model compression, sovereign AI silicon, AI security, generative AI for drug discovery, physics-based drug discovery, edge audio diagnostics, digital twins and new inference chips.
›Organisers say the ePitching is part of the EIC Investor Readiness and Outreach Programme under the EIC Business Acceleration Services, with more themed sessions planned in food and agritech, healthcare and energy transition.

EIC ePitching on AI and Software brings investors and deep tech startups together

On 8 July 2025 the European Innovation Council organised an online ePitching event under its Investor Readiness and Outreach Programme. Eight EIC-backed companies working across AI hardware, software, synthetic biology and diagnostics presented to a curated panel of 26 investors. The EIC reports more than 30 follow-up meetings and introductions resulting from the session. Investors included institutional VCs, corporate venture arms and strategic corporate investors active in industrial tech, energy and healthcare arenas.

The companies that pitched

Company	Core proposition presented at ePitching	Sector / notable claims
Multiverse Computing	Quantum-inspired AI model compression that reduces inference cost and speeds up inference by orders of magnitude	AI software, claims 4x to 12x inference speed and 50 to 80% lower operating cost and energy
Axelera AI	End-to-end AI compute platform and sovereign semiconductor technology using Digital In-Memory Computing and RISC-V cores	AI hardware, claims 10x faster inference and 3 to 5x better energy efficiency; 200+ TOPS family and 200+ employees
DeepKeep	Security and trust layer for AI including continuous risk assessment and runtime guardrails	AI security, focuses on LLMs and computer vision, targeting finance and insurance
Iktos	Generative AI plus robotics for drug discovery and automated DMTA cycles	Drug discovery, SaaS and pharma collaborations, 60+ projects and an internal drug pipeline
Kvantify	Physics-based, high-speed computational drug discovery methods to improve compound selection	Computational chemistry, promises faster, more accurate physics-based predictions
Neuron Soundware	AI, IoT and edge audio diagnostics for industrial equipment monitoring and predictive maintenance	Industrial monitoring, 130+ installations, claims top-five position in AI audio diagnostics
Twinsity	AI-powered digital twin inspection platform converting drone and robot imagery into 3D twins and audit-ready reports	Digital twins and inspection, claims to speed audits 10x and reduce inspection cost up to 90%
Vsora	Custom AI inference chip for data centre inference, autonomous driving, robotics and edge AI	AI hardware, positioning chips for inference workloads across cloud, edge and autonomy

Technical concepts explained

Model compression and quantum-inspired algorithms:Model compression covers techniques to reduce the size and computational cost of AI models while preserving performance. Quantum-inspired approaches borrow optimization techniques or mathematical formalisms from quantum computing without requiring quantum hardware. These may include novel factorisations or combinatorial solvers that reduce inference work. Vendor claims about multiple-fold improvements should be validated with independent benchmarks across representative workloads and deployment environments.

Digital In-Memory Computing:Digital In-Memory Computing places computation closer to memory to reduce data movement, which is a major source of energy consumption in neural network inference. Implementations vary from mixed-signal analog arrays to fully digital near-memory architectures. Claims of 3x to 10x energy or speed gains depend heavily on the workload, model size and end-to-end software stack.

Generative AI and retrosynthesis in drug discovery:Generative models propose novel molecular structures in silico. Retrosynthesis tools then determine synthetic routes from available reagents. Combining design, automated synthesis robotics and biological testing forms a Design-Make-Test-Analyse loop that can accelerate early discovery. Translating early hits into safe, developable drug candidates remains long, risky and expensive, especially for clinical development steps.

AI-native security and runtime guardrails:AI-native security platforms aim to detect and mitigate model vulnerabilities, hallucinations and adversarial attempts across the AI lifecycle. Runtime guardrails, red teaming and model provenance scanning are complementary measures. Enterprises require integration with existing security stacks and independently verifiable detection rates to trust such systems at scale.

Winners and investor voting

Investors present voted Multiverse Computing as the best pitch of the day, with Axelera AI placing second. Multiverse asserts its platform delivers inference speeds four to 12 times faster than existing solutions and reduces operating costs and energy use by 50 to 80 percent. The company positions its software as hardware and use-case agnostic, deployable at the edge, on-premise and in the cloud. Harshit Krisna, Head of Growth, Funding Strategy and CorpDev at Multiverse, emphasised the company’s focus on cheaper and greener inference and on European data sovereignty. The company said it has been an EIC beneficiary since its early days and credits EIC support with helping its growth and visibility.

Axelera AI pitched its end-to-end AI compute platform and RISC-V and Digital In-Memory Computing based chips. The company said it was founded in 2021, has more than 200 employees including 60 PhDs and serves over 230 customers. Louis Mather, VP of Strategy at Axelera AI, highlighted European sovereign supply chain arguments and said the event produced five high-potential investor conversations for the company.

Investor panel and impressions

The session assembled investors from Amadeus Capital Partners, Capricorn Partners, Crowberry Capital, DN Capital, eCapital, Equinor, Matterwave Ventures, GET Fund, NGP Capital, Open Ocean, Omnes Capital, Partech, SAP, Voima Ventures and others. Organisers report 26 investors attended and that the event led to more than 30 follow-up meetings and introductions.

Investor feedback highlighted a high calibre selection and sector diversity. Silviu Apostu, Principal at Matterwave Ventures, praised the EIC’s selection process and said the programme provides a minimum quality guarantee as companies have been tested through EIC processes. Apostu added that some pitches generated direct investment conversations and that verticals such as drug design and cybersecurity stood out. Steven Lambert, Investment Partner at Capricorn, noted that EIC gives access to deal flow that is otherwise hard to find and underlined the value of seeing a range of maturity stages from early product-market fit work to scaling teams.

What the EIC says and how this fits the European innovation ecosystem

The ePitching is run under the EIC Investor Readiness and Outreach Programme, part of the EIC Business Acceleration Services. The EIC BAS offers coaching, matchmaking, procurement and global expansion support to EIC awardees. Public materials from the EIC highlight large metrics for BAS since 2021 including more than 20,000 one-to-one meetings across programmes, several hundred deals, hundreds of millions of euros raised through investor outreach and a network of partners offering services to scale-ups. The EIC positions these interventions as de-risking steps that increase visibility and investor access for European deep tech innovators.

Investor Readiness and Outreach Programme explained:The programme provides dedicated investor readiness activities such as pitch deck reviews, benchmarking, valuation analysis and curated investor lists. It also organises pitching events, investor days and targeted marketing to increase exposure for selected EIC awardees.

A critical view on claims and outcomes

Vendor claims of 4x to 12x inference speed or 50 to 80 percent energy savings are notable and attractive to investors. Those figures are vendor statements and require independent validation in production-like settings, on representative models and datasets and with full stack measurements including software overheads, memory and IO. Artificially cherry-picked benchmarks or narrow workloads can overstate real-world gains.

Similarly, chip performance claims vary by metric. TOPS numbers are useful but do not directly translate to application-level throughput, latency or energy efficiency. Inference efficiency depends on model architecture, precision, batch sizes, software stack and peripheral system design. For buyers assessing hardware, long term supply, software toolchain maturity and roadmap commitments matter as much as peak performance numbers.

Follow-up meetings are an important early signal but do not equate to funding or procurement deals. Early investor interest often leads to diligence and pilot conversations, which may or may not convert to term sheets. For deep tech companies the path to revenue can be long, especially in regulated domains such as drug discovery and healthcare.

Implications for European policy and industrial strategy

The pitches reveal two consistent themes in European tech policy discussions. First is the emphasis on technological sovereignty. Hardware vendors and chip start-ups highlighted Europe as a differentiator, citing customers that value sovereign supply chains and transparency. Second is the strategic focus on vertical deep tech applications from energy and manufacturing to healthcare and drug discovery. The EIC's role in providing non-dilutive grants, visibility and investor introductions is intended to lower entry barriers for such capital intensive and long-horizon innovations.

However, sovereignty claims alone will not guarantee market adoption. Buyers will weigh price, performance, support, software ecosystems and long-term availability. Public funding and procurement can help bridge early adoption gaps, but commercial partnerships, standards compliance and integration work are required to scale.

Upcoming EIC ePitching sessions and how to apply

Themed session	Format and date	Application notes and deadlines
Food and agritech	Online, 24 September 2025	Applications open via EIC Community platform
Healthcare	Paris, 6 October 2025	Call specifically for EIC Accelerator companies in healthcare; apply before 9 July 2025 for the EuroQuity-Bpifrance investor day
Energy transition and sustainable energy	Online, 16 October 2025	Applications accepted until 3 August 2025, according to the call

Investors interested in participating or companies seeking to apply are invited to contact investments@eicfund.eu or to use the EIC Community Platform where open calls and eligibility details are published. The EIC calls indicate selection priorities may favour companies that have not recently participated in EIC events or that have not raised capital in the prior 12 months, so applicants should review call terms carefully.

Practical advice for founders preparing for EIC ePitchings

Prepare crisp, evidence-based claims. Include independent benchmarks, end-to-end metrics and clear buy signals from pilot customers. Explain the software stack and integration points for hardware claims. For healthcare or drug discovery, be explicit about regulatory milestones and timelines. Use EIC coaching and BAS services to refine pitch materials and to access investor-targeted benchmarking and valuation materials offered by the Investor Readiness and Outreach Programme.

About the EIC Investor Readiness and Outreach Programme

The programme sits within the EIC Business Acceleration Services. It aims to increase the investor readiness of EIC Accelerator beneficiaries through coaching, benchmarking, curated investor outreach and matchmaking events such as ePitchings and Investor Days. The wider BAS portfolio promotes market access, procurement matchmaking and global expansion services. These services are intended to complement grants by improving deal flow visibility and by connecting selected European deep tech companies with investor communities and corporate buyers.

Disclaimer: The information in this article is based on presentations and public materials from the EIC ePitching event and related EIC pages. Company performance claims are reported as presented by the companies and investors. Independent verification of technical performance and commercial traction is required before drawing investment conclusions.