Participated · Equity · Jun 2026

Founded in 2022 as a spinout from University College London and the Faraday Institution, Gaussion develops an external magnetic charging approach that can be retrofitted to existing lithium-ion battery packs. Its two main products are MagLiB, a magnetic printed circuit board add-on for packs, and Aeon, a control chip and software platform that together form a smart control layer. The company claims faster charging and extended battery lifespan without changing cell chemistry or design and supports 14 active commercial programs across automotive, aerospace, grid storage and consumer electronics. Gaussion has filed 66 patent applications in 17 families and reported about 19 employees as of June 2025. In October 2025 it moved into a 40,000-square-foot R&D facility in central London. The company positions itself to capture demand from AI data centres, electric vehicles and grid-scale storage as the battery management market grows.

Participated · Seed · Feb 2026

Stanhope AI is developing a brain-inspired system called “Active Inference,” which lets machines learn and adapt on the fly rather than depending on large, static datasets like those used by current LLMs. This adaptive capability targets mission-critical applications across autonomous systems, defence technology, industrial automation, and embedded devices where efficiency and reliability are paramount. The company is led by CEO Professor Rosalyn Moran alongside Professor Karl Friston, both noted researchers in computational neuroscience. With its technology, Stanhope AI aims to bridge the gap between biological cognition and machine learning, offering more efficient and resilient AI solutions. The startup plans to use newly secured capital to accelerate product development and scale operations. To date, its external financing totals $8 million, and no additional operating metrics such as revenue or user figures have been disclosed.

Participated · Equity · Oct 2024

Oriole Networks builds advanced photonics technology to create networks of AI chips that combine processing power using light, enabling LLM training up to 100x faster while consuming a fraction of the power. The company positions its solution to reduce the energy consumption and latency of AI data centres, addressing growing grid strain in the US and Europe. Founded in 2023 out of University College London, Oriole commercialises two decades of UCL research from founding scientists Professor George Zervas, Alessandro Ottino, and Joshua Benjamin. CEO James Regan, who previously built EFFECT Photonics, has assembled a seasoned commercial and photonics team including hires from Lumentum and backing from specialists in high-performance clusters. Oriole is backed by investors such as XTX Ventures and others that operate large GPU clusters, giving it potential customer and infrastructure insight. The company plans to have early-stage products in customers' hands by 2025 and is expanding its team and supplier engagement to accelerate deployment. Oriole Networks is a London-based spinout from University College London, founded in 2023 by UCL scientists Professor George Zervas, Alessandro Ottino and Joshua Benjamin with CEO James Regan. The company has developed a novel photonic approach that uses light to connect thousands of AI chips so individual GPUs combine into a single “super-brain” for training advanced large language models. Oriole says this architecture enables training with a thousandth of the latency while consuming a tiny fraction of the energy. Its IP is licensed through UCL’s technology transfer company, UCLB. The company raised £10M in Seed funding and intends to use the proceeds to enhance AI performance and drive adoption of its technology. No operating metrics (revenue/users) were disclosed in the article.

Led · Seed · Jun 2024

EpilepsyGTx is a biotechnology company focused on creating adeno-associated viral (AAV) gene therapies for refractory epilepsy and other disorders of neuronal hyperexcitability. Its lead candidate, EPY201, is designed for single, minimally invasive administration directly to the epileptogenic focus and aims to provide durable seizure freedom for patients with focal refractory epilepsy (FRE). The therapy avoids systemic exposure and eliminates the need for brain tissue resection, ablation, or chronic antiseizure medication. The newly secured funding will support first-in-human Phase 1/2a clinical trials to assess EPY201’s safety and efficacy. Beyond this program, the company is building a pipeline of transformative, localized gene therapies addressing severe neurological conditions. EpilepsyGTx has raised $43 million to date, including a $10 million pre-seed and seed round led by the UCL Technology Fund with participation from Zcube, and the recent $33 million Series A. FRE affects approximately 10 million people worldwide, highlighting the significant unmet need EpilepsyGTx seeks to address.

Participated · Series A · Aug 2023

Phasecraft builds ultra-efficient, hardware-agnostic quantum algorithms that pair quantum and classical computing to make Noisy Intermediate Scale Quantum (NISQ) devices useful for real-world applications. Its algorithms are designed from insights in theoretical physics and computer science and are validated through extensive numerical simulation and hardware partnerships. The company works with major quantum hardware providers including Google Quantum AI, IBM, Quantinuum, and QuEra to increase impact and portability. Phasecraft is already deploying solutions with end users such as Johnson Matthey, Oxford PV, the UK National Energy System Operator (NESO), and BT. Its work claims orders-of-magnitude efficiency improvements in materials simulation and applies quantum methods to biochemical processes and energy-network optimization. The company plans to double down on R&D breakthroughs and expand industrial efforts to build commercial solutions. Phasecraft develops quantum algorithms and partners with quantum hardware companies, academics, and industry leaders to move quantum computing from experimental demonstrations to useful applications. The company focuses early efforts on using algorithmic advances to discover novel materials for the clean energy transition. Phasecraft works with hardware partners including Google, IBM and Rigetti to tailor algorithms to current machines. The team was spun out of UCL and the University of Bristol and aims to build a larger team of quantum scientists, researchers and engineers. The company intends to use new funding to further develop its algorithms toward practical quantum advantage. Phasecraft builds quantum software and algorithms designed to reduce the quantum computing resources required for complex simulations, aiming to accelerate practical quantum applications. The company has developed an algorithm that significantly lowers the quantum power needed for calculations, which could enable new materials discovery. Founded by academics from the University of Bristol and UCL, the team includes internationally renowned quantum scientists Toby Cubitt, Ashley Montanaro and John Morton and currently employs around ten people. Phasecraft has established partnerships with quantum hardware companies such as Google and Rigetti and industry partner Johnson Matthey to help bridge hardware and applications. The startup’s stated goal is to advance quantum algorithms to bring commercial quantum applications forward by years. Financially, Phasecraft just closed a £3.7M seed round and has raised a total of £5.5M in venture funding and research grants to date.