How Beijing transformed a cement plant into a magnet for global mathematics

From factory floor to a campus by Yanqi Lake

On the northern edge of Beijing, a former cement plant now houses a thriving center for mathematics. The Beijing Institute of Mathematical Sciences and Applications, known as Bimsa, sits beside Yanqi Lake with mountain views and, on clear days, a glimpse of the Great Wall. Inside, research groups gather in bright rooms to work through geometry, algebra, quantum fields and string theory. The setting is peaceful, but the ambition is large. The institute has become a meeting point for mathematicians from many countries who want time, collaborators and resources to pursue difficult problems.

Bimsa has grown fast. It employs more than 200 full time researchers, with about one third coming from outside China. Mathematicians from nations at odds, including Ukraine and Russia, work side by side and share blackboards. The draw is clear to visiting scholars, who describe a scale of activity, a mix of disciplines and a density of expertise that is rare in a single campus.

The institute is jointly supported by the Beijing municipal government and Tsinghua University. That funding model has backed competitive packages for senior and junior researchers, with salaries that, according to institute materials and recruitment notices, range from 700,000 yuan to 3.5 million yuan a year. The resources are paired with a structure that minimizes administrative tasks so faculty spend most of their time on research and seminars.

The vision reflects the leadership of founding dean and president Shing-Tung Yau, a Fields Medal winner whose work reshaped differential geometry and mathematical physics. His international standing has helped recruit established researchers and rising talent, while connecting Bimsa to a global network of institutes and conferences.

What is Bimsa and how does it work

Bimsa positions itself as an open research campus focused on both basic and applied work. Roughly half of the institute is focused on applied mathematics, one quarter on pure mathematics and the rest on cross disciplinary areas such as artificial intelligence and data science. The campus hosts frequent seminars, workshops and reading groups that push collaboration across groups and fields.

Research mix at scale

Research spans algebraic geometry, algebraic topology, analysis and geometry, number theory, mathematical physics and general relativity, quantum fields and strings, quantum symmetry, cryptography, computational mathematics, statistics and probability. On the applied side, faculty work on machine learning, optimization, algorithms for scientific computing and mathematical models of complex systems, including projects related to the digital economy.

Weekly gatherings aim to keep researchers conversant with progress outside their specialties. In these sessions, faculty from different areas present the most important recent results and techniques. The aim is to generate new ideas and collaborations, and to shorten the time from a breakthrough in one field to its uptake in another.

South Korean theoretical physicist Kimyeong Lee, who has worked in the United States and Asia, underscored the scale and intensity of daily exchange at Bimsa. He noted that large schools in the US have wide resources, yet smaller departments can feel isolated. At Bimsa, he said, the everyday experience is different.

In the US there are many excellent universities, but smaller institutions can offer limited opportunities. At Bimsa, interaction is everywhere.

Teaching and the talent pipeline

Many Bimsa researchers also teach at Tsinghua University’s Qiuzhen College, which opened in 2021 to train students aiming for research careers in mathematics. Instructors describe the students as highly motivated and quick to absorb new material, a sign that Beijing is building a deep pipeline for future faculty and industry researchers.

Russian mathematician Semen Artamonov, who teaches at the institute, offered a straightforward description of the pace in the classroom.

It is as if they already know the material and came for advanced level material.

A global stage: the International Congress of Basic Science

Since 2023, Bimsa has hosted the International Congress of Basic Science, a two week gathering that brings together Fields Medal recipients, Turing Award winners, Nobel laureates and scholars across mathematics, physics and computer science. The program includes lectures, panel discussions and prize ceremonies that celebrate influential papers and lifetime contributions. Organizers and committee members have included leading figures in both mathematics and physics.

The congress awards Frontiers of Science prizes, which recognize standout papers in recent years across mathematics, theoretical physics and computer science. Among the recipients at the inaugural congress were mathematicians Máté Matolcsi and László Végh for work that drew wide attention in their fields, including advances on long standing problems and algorithmic breakthroughs. Lifetime honors went to scholars whose contributions shaped entire branches of research.

Iranian theoretical physicist Mohammad Hossein Yavartanoo, now a Bimsa professor, described the scale of the event in plain terms.

It was a very big event. I had not seen anything like it anywhere in the world.

Gatherings like the congress give young researchers concentrated access to leaders in their areas, and they strengthen the campus culture of open problem sessions and intensive courses. For Bimsa, they also signal to the wider community that Beijing intends to be a crossroads for basic research, not only applied projects with near term commercial payoffs.

Collaboration without borders amid conflict

Bimsa’s daily life includes collaboration among mathematicians from countries that are otherwise estranged. The institute has made a point of welcoming researchers from Ukraine and Russia, and it has extended that ethos to institutional partnerships. One example is the Geometry and Physics project launched with HSE University in Moscow, a joint effort focused on homological mirror symmetry and birational algebraic geometry. That cooperation has included conferences in Beijing and Moscow, with mixed in person and online formats that bring senior scholars and early career researchers together.

In November 2024, a joint conference on algebraic geometry and mathematical physics gathered more than 30 mathematicians, with three Fields Medal recipients among the speakers. A substantial portion of the program featured doctoral students and postdoctoral researchers from China and Russia who presented new work and learned from senior figures. The approach has been to seed long term collaboration through repeat events and shared projects, rather than single visits.

Valery Gritsenko, head of HSE’s International Laboratory for Mirror Symmetry and Automorphic Forms, summed up the teaching value of these programs for rising scholars.

It was a great learning experience for the new generation of mathematicians.

For individual researchers, the pull is as much about time as it is about talent. Ukrainian mathematician Yevgen Makedonskyi highlighted the freedom to focus that he found on the Yanqi Lake campus.

Bimsa is one of the few places in the world where I can do maths and almost not worry about anything else.

Funding, pay and the race for talent

Bimsa’s compensation and support have helped it compete for established names and promising young faculty. Public recruitment calls describe salary ranges from 700,000 yuan to 3.5 million yuan per year, with additional research support, housing assistance and grant opportunities. The institute continues to expand its roster of research groups and positions, covering appointments from postdoctoral researcher to professor.

The campus environment has also been a draw. Proximity to Tsinghua’s Yau Mathematical Sciences Center enables shared seminars, joint supervision of students and cross campus projects. The location near the mountains gives a quiet setting for research, and the surrounding Huairou science park has grown into a cluster for basic science institutes and laboratories.

China has been increasing resources for fundamental research over the last decade, with more university rankings success in natural sciences and engineering and new long term philanthropic programs for basic science. The landscape still relies heavily on government support, yet non state programs are becoming more visible. Those trends have made it easier for institutes like Bimsa to plan multi year agendas in pure and applied mathematics and to recruit internationally with competitive offers.

Why mathematics, AI and industry are converging

In the AI era, many of the most complex problems hinge on mathematical insight. Wu Rongling, deputy head of Bimsa, has argued that control over advanced AI depends on much more than scaling up datasets or compute. He connects the institute’s research aims to China’s broader strategic goals in technology.

Wu’s view is that a deep base in mathematics is essential for independent innovation in AI frameworks and tools. He has expressed the argument in simple terms.

Only by mastering mathematics ourselves can we develop our own AI frameworks.

Modern AI leans on linear algebra, probability and optimization. Those tools are a start, but many industrial and scientific problems involve structure that is not captured by standard models. For example, rigid constraints in molecular dynamics, symmetry and conservation laws in physics or topological constraints in materials cannot be learned easily without guiding mathematical structure. When those structures are built into model architectures and training objectives, systems can be more sample efficient and less fragile.

Wu describes the current challenge as a gap between the complexity of real world problems and the simplicity of many deployed models. In his view, the next step is to bring advanced mathematics into the core of AI systems to handle those constraints and to reduce the heavy compute that brute force training often demands.

AI now uses basic math, but real world problems are complex. Complex problems need complex tools.

The practical payoff could include faster drug discovery through better molecular modeling, more reliable predictions in climate and fluid systems, or improved design in materials science and engineering. Those areas align with the interests of several Bimsa groups and with applied programs at Tsinghua and partner institutions.

Outreach, museums and public engagement

Bimsa has tried to build a research culture that starts early. Faculty participate in student contests, summer camps and thematic workshops run through Tsinghua’s Yau Mathematical Sciences Center. The campus opened the Yanqi Lake Mathematics Museum to give students and visitors hands on exhibits that explain core ideas and famous problems, an effort to make pure mathematics approachable and to inspire future researchers.

On the applied side, Bimsa’s laboratories have pursued projects with industry and government partners. Examples include indices and modeling work in the digital economy, where measurement and statistical methodology matter for policy and business. Those programs give students exposure to real datasets and constraints while keeping research grounded in first principles.

Challenges to watch

Bimsa’s rapid rise also faces challenges. International collaboration in science continues to face political tension, export controls and visa barriers. Conferences, joint labs and student exchanges need stable channels and predictable rules to keep momentum. Beijing’s broader academic diplomacy has given the institute partners to work with, including universities across Asia and Europe, but sustaining that openness requires continued attention.

Campus leadership has put a premium on keeping researchers free from heavy administrative burdens and from political interference. Maintaining that standard while the institute grows will be a test of internal governance. Hiring diverse faculty, integrating newcomers smoothly and balancing pure and applied work at scale are practical issues that come with size.

Stable funding is another variable. Government budgets have supported the institute’s growth, and new philanthropy programs in China are increasing support for basic research with multi year grants. Building a wider base of long horizon funders would make it easier to plan cycles of recruitment, major conferences and long projects that do not fit annual or short grant timelines.

Key Points

• Bimsa in northern Beijing has turned a former cement plant into a busy mathematics campus with more than 200 full time researchers and a strong international presence.
• The institute is backed by the Beijing municipal government and Tsinghua University and offers competitive packages that help attract global talent.
• Research spans pure mathematics, applied mathematics and interdisciplinary work in areas such as AI, data science and mathematical physics.
• Since 2023, Bimsa has hosted the International Congress of Basic Science, a large gathering of leading scholars that includes Frontiers of Science awards.
• Partnerships, including a Geometry and Physics project with HSE University, support joint conferences and training tracks for early career researchers.
• Leaders at the institute emphasize that a strong base in mathematics is key for independent progress in AI and complex real world modeling.
• Outreach programs, a mathematics museum and student contests build a pipeline for future researchers and connect pure theory to public engagement.
• Key challenges include maintaining open international collaboration, ensuring research independence and broadening long term funding sources.