Let’s face it, it’s an AI founders’ world and we’re all just living in it. Entrepreneurs with experience at leading research labs continue to raise mega seed rounds or sign high-profile commercial deals with big clients, and investors are hailing the technology as the phoenix lifting the startup sector from the ashes.
But, amid all the potentially highly lucrative wheat, there is also a lot of chaff in today’s “AI startup” sector, with many opportunistic companies just jumping on the hype bandwagon.
So, to try and find some of the younger, lesser-known companies that are worth being aware of in Europe’s AI scene, Sifted asked four leading investors in the space to pick some of the most exciting prospects they’re seeing (no portfolio companies allowed).
Here are Europe’s 11 AI startups to watch.
Rick Hao — partner at Speedinvest
Speedinvest in an early-stage firm based in Austria
Extensity.ai — Austria
Extensity.ai has developed a “symbolic AI” framework using a large language model (LLM). Neuro-symbolic AI combines the best of current deep learning and generative AI models with human-like “common sense,” to enable autonomous systems to intrinsically understand, think and reason.
Causa — UK
Causa is a company that provides causal AI, a type of AI that accurately predicts and shapes future outcomes, provides personalised recommendations and helps in making optimal decisions. It offers a solution, CausaDB, to manage workflow end-to-end, with features such as simulating actions, finding optimal decisions and conducting adaptive experiments.
Invariant Lab — Switzerland
Invariant Lab is building a novel platform for secured AI agents, systems designed to carry out complex workflows, to interact with data and APIs in a fully compliant way. With common LLM-based systems, AI agents cannot be guaranteed to act in compliance with internal company policies around data privacy. The company has built a novel approach called secure action models to sit at the intersection of LLMs and secure systems, to ensure all actions of AI agents are within the bounds of policies and security protocols.
Flavia Levi — earlystage deeptech investor at Octopus Ventures
Octopus Ventures in an early-stage VC firm based in London
Innatera — Netherlands
The company’s neuromorphic (brain-inspired) processors mimic the brain's mechanisms for processing sensory data in power-limited and latency-critical applications, such as “edge” applications (eg. when AI has to run on smaller devices). It's an outstanding team – one that brings industry and academic expertise.
Lumai — UK
Lumai is addressing the demand for AI capacity within data centres. These data centres are facing multiple constraints, including energy consumed by silicon-based GPUs. By leveraging 3D optical computing (i.e. overcoming the constraints of the 2D geometry of silicon chips), Lumai is enabling AI deployment at scale — with the fastest and most energy-efficient AI processor.
Fractile — UK
Fractile’s founding technology was spun out of the University of Oxford, after researchers Walter Goodwin and Yuhang Song observed a converging trend across multiple AI domains towards the use of a single type of model — the transformer (the underlying technology that powers AI models like ChatGPT). They started working on a new, ground-up approach to accelerated computing, designed to run these models at the fastest possible speeds by mapping the computational problem to the underlying hardware the model is running on.
Andreas Riegler, founder and general partner, Apex Ventures
Apex Ventures is an early-stage deeptech firm based in Germany
GEMESYS — Germany
GEMESYS is a pioneering German startup based in Bochum revolutionising the AI landscape by developing innovative AI chips tailored for edge computing applications. Unlike traditional methods that rely heavily on vast computing resources and centralised data processing, GEMESYS introduces a novel approach by creating fully analog hardware inspired by the human brain's processing capabilities.
Mindgard — UK
Mindgard is a UK startup that specialises in cybersecurity for AI. Founded in 2022 and spun out of the UK’s Lancaster University, its platform delivers automated security and rapid cyber threat response for AI, GenAI, and LLMs. Based in London, the company is led by cofounder and CEO Prof. Peter Garraghan — a global authority in AI security — and is backed by leading AI scientists and cybersecurity engineers.
Cogna — UK
Founded in May 2023 by Lars Mennen and Ben Peters, Cogna is at the forefront of revolutionising productivity across various industries. The two previously worked together at self-driving car company Five, which was cofounded by Peters and acquired by Bosch in 2022.
Based in London, it is developing an automated “software factory” designed to democratise the creation of custom software, enabling non-technical problem and process experts to easily articulate their challenges, explore and outline potential solutions and use AI-driven programme synthesis for rapid and cost-efficient software development.
Francesco Ricciuti — associate at Runa Capital
Runa Capital is an early-stage firm based in Luxembourg
Synthara — Switzerland
Synthara is offering chip makers the ability to seamlessly integrate in-memory computing into their existing platforms. This approach produces a 50-fold increase in computing performance without disrupting any chip architecture or software processes.
Nebuly — Italy
Hailing from Italy, Nebuly is a startup that's building user analytics for LLMs. The result allows enterprises to learn how users are interacting with models.
Ricciuti also nominated Fractile.
