2025 annual report

In 2025 we entered a new phase of consolidation, strengthening our governance, core funding and financial infrastructure. Our scientists published research across all four themes, while we secured £5m in new funding commitments and created a forum of governors to complement our reformed board of trustees.

Strategy

Since its founding, the London Institute for Mathematical Sciences has passed through three broad phases. The first was about recognition: establishing our scientific credibility and gaining access to major science-funding agencies in the US, Europe and the UK. The second was about growth: recruiting scientists and staff, building international talent programmes, and moving into larger premises at the Royal Institution. In 2025 we entered a third phase of consolidation: making LIMS durable and financially secure. To do that, we focused on three things: governance, core funding and financial infrastructure.

Our first focus was our governing body. Since LIMS was created, our trustees played an advisory role rather than a fundraising role. In 2025 we changed this: financial support became an explicit expectation of our trustees, and we created the London Institute Council, combining the reformed board with a new, lighter-touch forum of governors. Both trustees and governors are expected to make or broker financial contributions. We describe the Council below.

Second, we needed to secure the Institute’s core costs. In general it’s easier to raise money for new fixed-term scientists than for the long-term scientists and staff already in place, or for rent and other running costs. We call the former extrinsic costs and the latter intrinsic costs. The distinction matters, because extrinsic costs are taken on after their funding has been secured, whereas intrinsic costs have to be met regardless. In 2025 we focused on gifts and grants that support intrinsic costs.

Third, we set out to strengthen our financial infrastructure. Over the years, we had consistently raised the calibre of our scientists, staff and trustees. But our financial services had not kept pace: in 2024 we were still using the same accountant, audit firm and bank we had used a decade earlier, when the Institute was smaller and our needs were simpler. In 2025 we brought in Ben Lynch to strengthen financial strategy and forecasting, appointed PEM as our new auditors, and began drawing on the financial expertise of our governors.

Research

This consolidation serves the Institute’s central purpose: research. In 2025, our scientists published work across all our four themes, including papers in Physical Review Letters, Physical Review X, Communications in Mathematical Physics and Nucleic Acids Research.

In “The elegant universe”, Andrei Stepanenko et al. published in PRL the fastest possible way to transfer a quantum state along a chain of qubits, reaching the quantum speed limit. Working with IBM Quantum and others, Juven Wang helped create and braid exotic quasiparticles known as Fibonacci anyons in superconducting qubits, a proof of principle towards fault-tolerant universal quantum computation. In Physical Review D, Evgeny Sobko introduced a new class of exactly solvable quantum field theories in which a fifth dimension emerges from the mathematics.

Work on “Life, learning and emergence” ranged from metamaterials to biology. In our first PRX paper, Oleksandr Gamayun and coauthors showed how non-reciprocity can sustain a new kind of soliton: a compact wave that travels in one direction while oscillating internally, and can persist indefinitely. Mikhail Burtsev led work on GENA-LM, a family of DNA language models able to analyse sequences up to 36,000 base pairs and recognise biologically important features, published in NAR.

In “Mathematics that unifies”, Oleksandr Kosyak found a new connection between two braid-group representations, showing how the Lawrence–Krammer representation can be derived from the Burau representation. Yang-Hui He and collaborators uncovered a common mathematical structure linking number theory, algebraic geometry and quantum field theory, in work published in CMP. Ilya Shkredov and collaborators answered a question of Endre Szemerédi, proving that sets of integers with relatively few prime factors must expand strongly under addition or multiplication. Their result appeared in Compositio Mathematica.

Finally, in AI-assisted discovery, Yang-Hui He and collaborators used reinforcement learning to search possible particle spectra in supersymmetric quantum field theories far more efficiently than brute-force scans, providing a general way to map the different regimes of these theories. The work appeared in Physics Letters B.

Gifts and grants

The year 2025 was our strongest fundraising year to date. Working with our development director, Sarah Myers Cornaby, we secured £5m in new funding commitments, around 50% more than in any previous year. Crucially, £2.1m of this is for intrinsic costs. Some of this funding arrived in 2025, with the rest committed for 2026 and beyond. Since our 2025 expenses were £2.4m, sustained fundraising at this level would put the Institute on a strong long-term financial footing.

Our largest gift was from the Khodorkovsky Foundation, which committed £2.5m over five years to continue our programme of Arnold and Landau Fellows. The gift also supports our new programme of Mendeleev Visitors, named after Dmitry Mendeleev, founding father of the periodic table and a regular visitor to the Royal Institution, where we’re based.

In early 2025, Mikhail Burtsev, one of our scientists, introduced us to the founder of Cognia AI. Our common vision developed into our largest commitment from a new donor that year. Cognia pledged £1.6m over three years to support junior fellows and visitors in the fields of life, learning and emergence, which together form one of our four research themes.

In late 2024, the fintech founder and philanthropist Ben Delo donated £100,000 to help us recruit the Harvard physicist Juven Wang. In 2025, he pledged a further £700,000 over five years to create the Ben Delo Fellowship, with Dr Wang as its first occupant.

Building on our longstanding scientific ties with America, we raised $250,000 from a US donor to support our programme to recruit American scientists. A further $130,000 came from a US foundation, brokered by our trustee Talulah Riley.

Alongside philanthropy, in 2025 we applied for over £4m in research grants, mainly to UKRI but also to Innovate UK, the Simons Foundation and Open Philanthropy. Most decisions remain outstanding because of long wait-times. One success so far is a $100,000 award from Open Philanthropy (now Coefficient Giving) to explore what the professionalisation of sport can teach us about research.

Our governors contributed cash and shares worth an estimated £300,000 during the year. Nebius provided compute time valued at $100,000. We also received smaller gifts from the Andrei Sakharov Foundation, Philip Bond, and Daniel Crispin, whose donation was backed by Rothesay.

Communication and events

In 2025, the London Institute broadened its public presence through articles, interviews and an expanding programme of events.

At our annual St Scholastica’s Feast in February, we announced the Ben Delo Fellowship. Mr Delo wrote in The Spectator about why he funded it, praising the Institute’s “blend of entrepreneurial spirit and academic excellence”. Then our writer Thomas Hodgkinson profiled him at length in Perspective magazine, tracing his path from mathematics to BitMEX to philanthropy and support for us.

Recruiting Juven Wang was part of a wider effort to bring on board American scientists. In The Spectator, Thomas Fink wrote about how Britain should respond to cuts in American science funding. In June, we welcomed Caltech president Thomas Rosenbaum and Ditchley Foundation director James Arroyo for a conversation about what those cuts might mean for Britain.

We continued to bring science to a wider audience. Ananyo Bhattacharya wrote in Nature about a breakthrough in the Langlands programme to unify mathematics. In The Idler, Mr Hodgkinson marked the 200th anniversary of the Royal Institution’s Friday Evening Discourses, the longest-running science talks in the world. AI-assisted discovery was a recurring theme. Yang-Hui He was interviewed about it in a feature in the Financial Times and appeared on the podcasts Theories of Everything and Quanta’s The Joy of Why.

At the Royal Institution, three of our scientists and trustees took to the stage of its famous lecture theatre. Mikhail Burtsev explored how large language models applied to DNA could transform biology. Yang-Hui He examined how humans and AI make discoveries together in mathematics. And Martin Reeves considered why the path to scientific and technological breakthroughs is often unplanned.

Toward the end of the year, we launched In Conversation, a new series in which our chief science writer Ananyo Bhattacharya interviews leading scientists and entrepreneurs about their life and work. In October we hosted the physicist and tech founder Stephen Wolfram, followed in December by the mathematician and author Edward Frenkel. We filmed both events for our new YouTube channel, with 21,000 views by the year’s end.

Council

As the London Institute has developed, so has its board of trustees. In our first phase, the board consisted of eminent scientists who lent their stamp of competence to a young organisation. In our second, it brought together leaders of learned societies and chief scientific advisers. In our third, we began building a governing body that not only advises the Institute but actively advances its mission by giving or helping to secure financial support.

To put this into practice, in 2025 we created the London Institute Council, combining our revised board of trustees with a new forum of governors. We took inspiration from Caltech, which made financial sovereignty an abiding principle and systematically reformed its board during its $3.4bn Break Through campaign. Governors advise and support the Institute, but unlike trustees they carry no formal legal responsibilities.

Our six inaugural governors put this new model into practice from the outset, broadening the Institute’s network of supporters. They are: Ammad Ahmad, a founding partner of Atheneum; Pinar Emirdağ, an entrepreneur in electronic trading and capital markets; Thore Graepel, Distinguished Researcher in AGI Futures at Google DeepMind; Amit Jain, co-founder of Signal Capital Partners; Nonna Materkova, founder of the Calvert 22 Foundation; and Tom Walton-Pocock, co-founder of the deep-tech investor Geometry.

Two longstanding trustees, Sir Roy Anderson and Sir John Beddington, stepped down in 2025, giving us room to reshape the board. We appointed two of the Institute’s existing supporters: Florian Schuster, co-founder of the Cambridge cell-programming company bit.bio, and fintech founder and philanthropist Ben Delo.

Together, these changes strengthened the Institute’s foundations and its prospects for long-term financial independence. By the end of 2025, an ambition that had once seemed distant was coming within reach: a permanent endowment for LIMS.

Main image
Main image
Main image
Main image
Main image
Main image
Main image
Main image
Main image
Main image
Main image
Main image
Main image
Main image
Main image
Main image