5Y Talk | Next-Gen Computing + Experimentation: The Birth of New Application Categories

Information technology is penetrating every field at an unprecedented scale, reshaping the future of every industry. Medicine — the sector most intimately tied to each of our health and daily lives — is also facing a paradigm shift under IT's influence.

As one of the earliest investors to bet on the "ITBT" track, over the past few years we've discovered and supported many entrepreneurs in digital healthcare. They've thrown themselves into this field without hesitation, using technology to break through industry bottlenecks. From a longer time horizon, IT's penetration into medicine has only just begun — much more change lies ahead.

What new possibilities can next-generation computing and experimentation bring to the industry? How have entrepreneurs in this space explored and practiced? Last month, 5Y Capital and VB Data jointly hosted the "Digital Pharmaceutical R&D (ITBT)" forum.

During the roundtable, three founders from 5Y Capital's portfolio — Wen Wen, founder & CEO of Revolone Biotech; Yan Tan, founder & CEO of Unknown; and Yuchong Wang, founder & CEO of Cellnovo Biotech — were joined by two industry experts, Xuwo Ji, founder of PuriBenchmark Technologies, and Weiwen Qian, deputy general manager of Wugan Pharmaceutical, for an in-depth discussion on "Next-Generation Computing + Experimentation: The Emergence of New Application Categories."

We've compiled portions of the roundtable. Hope you find it insightful :)

Wen Wen: I'm Wen Wen from Revolone Biotech. Our company focuses on continuously optimizing an AI + systems immunology technology platform. We do this by collecting large-scale multi-omics immunological data, performing engineered mechanistic modeling, and accelerating biomarker and target discovery through computation.

Please introduce yourselves — how are you applying next-generation computing and experimentation, and why did you choose this field?

Yan Tan: I'm Yan Tan from Unknown. Unknown is a gut microbiome drug development company. We're a pharmaceutical company whose drug active ingredients are live microorganisms or molecules that modulate live microorganisms. We use high-throughput automated experimentation combined with computational and AI methods to discover drug molecules or medicinal live bacteria.

In 2007, American scientists launched the Human Microbiome Project, a large-scale scientific initiative comparable to the Human Genome Project completed around 2000. It aimed to elucidate how gut microorganisms affect human health and disease. Based on this, we believed the gut microbiome would spawn an enormous industry, with pharmaceuticals being the largest sector capable of closing the entire value chain loop. That's why we entered this industry.

Yuchong Wang: I'm Yuchong Wang, founder of Cellnovo Biotech. Cellnovo Biotech is among the earliest teams developing single-cell technology in China. In ITBT (IT: Information Technology; BT: Biotechnology), we lean more toward the BT side. Life science is a field propelled by observation technologies — the invention of microscopes let us discover microorganisms, which led to antibiotics.

One of the biggest breakthroughs in life science observation in recent years is single-cell technology. For the first time in history, we can analyze molecular biological processes at single-cell resolution. This not only greatly advances our understanding of life processes but also enables systematic comprehension of previously intractable, complex endogenous diseases.

Cellnovo Biotech built China's first mass cytometer, and around this platform we've developed diagnostic-grade single-cell technology for R&D, manufacturing, and clinical applications. Starting last year, our equipment gradually moved from labs to production lines. Early this year we began commercialization, and we'll have more commercial applications — especially clinical ones — in the coming period.

Xuwo Ji: I'm Xuwo Ji from PuriBenchmark Technologies. Our company mainly uses advanced bioinformatics algorithms and AI to mine massive multi-omics data, hoping to achieve deeper understanding of disease biology and thereby more effective therapeutic strategies. We've built AIBERT, a multi-omics data mining system for new drug R&D, and established an open collaboration platform. We now have partnerships with over twenty leading domestic and international pharmaceutical companies.

Weiwen Qian: I'm Weiwen Qian from Wugan Pharmaceutical, a Suzhou-based pharmaceutical manufacturer with 34 years of history. I joined in 2017 and was quite shocked — looking at the industry through an IT person's lens, it felt like medicine was still in the farming-with-fire stage, with all R&D being experiment-driven. My first thought was that computers could definitely be introduced into the entire experimental process. Over the past four years we've built an AI R&D team, and we're continuing down the path of combining a traditional enterprise with artificial intelligence.

Wen Wen: Many of you come from academic backgrounds, but academic research and entrepreneurship are completely different domains. What challenges arise when translating an academic result into commercialization? What have you learned from this process?

Yan Tan: Since we're all IT-plus-BT or AI companies, an important aspect of AI is the optimization objective. A major difference between scientific research and business is that the target optimization function may differ. Scientific research requires extreme innovation — going from zero to one in foundational theory or the most cutting-edge directions, requiring genuine innovation. When moving into industry, finding a viable path to implementation alongside innovation is equally important. This is a crucial distinction.

I often share a concept called "sci-tech entrepreneurs" internally. China's commercial history hasn't really seen this group before. Previously there was more commercial entrepreneurship; it's only in our generation that large numbers of scientists have become company founders. Going forward, how to truly apply scientific research and innovation to business and achieve industrialization is the biggest challenge for sci-tech entrepreneurs.

Yuchong Wang: The journey from science to technology to product to application requires simultaneous consideration of how technology and business models complement each other. When we chose the single-cell technology path, we had to consider how to better meet clinical customers' needs in terms of cost, speed, throughput, and applicability.

This touches on a more fundamental issue: when designing products, you must base decisions not only on current industry needs but future needs as well. How to build cognitive advantage around future industry needs involves methodology and accumulated time — rapidly building understanding while exploring existing or potentially viable technologies globally, integrating, reconstructing, and even breaking through these technologies to create products that truly match users' future needs. This may be entrepreneurship's greatest challenge.

Xuwo Ji: Traditional tech transfer follows basic research first, industry second. "Omics data mining-driven R&D" follows the same pattern — bioinformatics and multi-omics data mining made rapid progress in research first, then found application in industry.

Our biggest challenge on the AI+BT path now is that on one hand, we need to ensure R&D leadership and keep pace with academia's latest advances; on the other, we must clearly define the boundary between research and industry — where research results can deliver maximum value in industry, where to draw the conversion line, whether to push it ten meters forward or pull it five meters back. This is the hardest part.

What I often worry about isn't whether our technology is advanced enough, but whether its application is grounded enough. When the wind is blowing, people may not consider this "boundary" — the more advanced the better. But all trends rise and fall. Even in a clearly upward trajectory, "going from one high to another," there are relative valleys in between, and that's when grounding matters. This thinking permeates everything from strategy to tactics. For example, the best choice from a research perspective may not be the best choice for industry. Abalone and shark fin are great, but you need to be able to eat plain tea and simple rice too — because it also contains information that can't be casually discarded.

Weiwen Qian: Since our company has existed for over thirty years, we don't face the early startup conversion problem so much as the challenge of enterprise longevity. Regarding how to delineate research and industrialization, I personally believe they've always been integrated — the ultimate goal of research is industrialization, and research results all eventually face conversion. I don't think there's a separation issue; it's about how to do it.

Yan Tan: I may disagree somewhat. I think scientific research may lead to industrialization, but industrialization shouldn't be the purpose of doing research. If basic scientific progress takes industrialization as its ultimate goal, it may compromise research purity. However, when industrializing, companies may have many opportunities to do research. A company can't operate with pure research interest or basic research as its purpose; it must confirm the company's goals, tolerance levels, and technical capabilities to conduct certain research and industrial conversion.

We hope China can, in 5, 10, or 20 years, have basic research not guided by industrialization goals, while also having sci-tech entrepreneurs constantly finding potentially commercializable points in research and breaking through barriers. In the future, we may see more great technology-oriented enterprises emerge.

Wen Wen: The guests are all scientists and entrepreneurs. I actually come from a business background, but our company has many PhDs — some in basic research, some in translational research, and some even in clinical medicine. We also discuss together the process from research to commercialization to application.

At the company level, we need to see long-term commercial value, and our strategy is to do R&D planning based on 5-10 year goals. Life science differs from other industries — if you only see what's coming next year and start R&D now, you're already too late; newer technologies will emerge next year. So you need to look 5-10 years ahead, have a general vision for healthcare, R&D, and life science overall, then plan accordingly. Especially when company resources are limited, you do need to make certain trade-offs, considering cost, feasibility, and future commercial value.

The next question: next-generation computing and experimentation are very cutting-edge. Among your clients or partners, what do they value most about you? Please share your experiences.

Yan Tan: We mainly do drug R&D, and our biggest partners are hospitals and research institutes. Taking hospitals as an example — from a drug R&D perspective, a given indication may previously have had no treatable drugs, which you find exciting. But for doctors, they don't first see a disease category or new target; they see living patients, families behind them. Doctors prioritize a company's professionalism and product nature, then technological cutting-edgeness.

In our earliest days we often talked with doctors about our research capabilities and technological leadership — this was our first foot in the door. But truly building long-term partnerships hinges on implementation and whether you can genuinely bring therapeutic benefit to patients.

Yuchong Wang: The fortunate thing about doing technology is "in culture there's no first or second, in martial arts there is" — in this sense, the best support clients give us is support for product strength.

On the other hand, the medical industry itself has extremely high requirements for compliance and safety, especially for diagnostic-grade products where there can be no lapses. Therefore the entire R&D chain and trial-and-error cycle are very long, so fundamentals in healthcare matter greatly — you can't take detours. That's why we explore the 1-to-100 process with top clinical institutions at very early product stages.

Xuwo Ji: Since PuriBenchmark itself is an open collaboration platform, in working with pharmaceutical companies we can tangibly advance new drug R&D progress, which gives us great satisfaction. Even more satisfying is when not only pharmaceutical partners recognize us but regulatory authorities do too — when our technical capabilities truly enable these drugs to enter clinical trials, proceed phase by phase, and eventually gain approval and reach patients.

Weiwen Qian: We're mainly not doing innovative drugs but more industrialization projects, where the core concerns are project timeline, efficiency, and clients' target goals, so understanding client needs is key. This is also what clients value about us.

Wen Wen: To summarize everyone's points — whether clinically, with different users, pharmaceutical companies, or other client types, everyone solves pain points through products and services. And in life sciences, being able to solve even a bit of a pain point is quite fulfilling. We've had this experience ourselves.

The next question: for this field, cross-disciplinary team building is also critical. Could you share your experiences in talent recruitment?

Yan Tan: Basically for companies like ours, R&D staff concentrate mainly in IT and BT. For recruitment, finding computational talent who deeply understand biology is very difficult; simultaneously, finding people with computational or high-throughput thinking among experimentalists is also hard. We say IT and BT need to integrate — IT people need to talk BT, BT people need to understand IT results. The company's R&D thinking must reflect integration, and every individual in the team needs this mindset. IT and BT were separate in the past; now they need to merge. This requires extensive training, and training is itself a challenge.

Yuchong Wang: Because we're in equipment manufacturing, our entire R&D chain is long — from physics to electromechanics, materials, to molecular biology, cellular immunology, and also algorithms and software development, requiring multi-disciplinary integration. In our internal innovation system, we often emphasize being "tight in one aspect, loose in another."

"Tight" means being core-product-oriented with clear R&D boundaries. People doing technology easily fall into "when you have a hammer, everything looks like a nail" — wanting to fix everything everywhere. Excellent products require precise definition and clear paths to be done well.

On the other hand, for much innovative work you can't set too many restrictions; you must allow individuals maximum creativity within frameworks. Simultaneously, you need an effective internal information coordination system so people can better collaborate. There will definitely be various collisions in this process — that's daily life at startups.

Beyond R&D, the bigger challenge is on the commercial side. Healthcare commercialization and trial-and-error cycles are both very long; any time savings are extremely valuable for startups. How to build rapid cognitive chains and quickly achieve commercial output also requires the right talent structure.

Xuwo Ji: Since my co-founder and I have been in this field for 20 years, relatively scarce bioinformatics talent is actually easier for us to recruit. What we currently seek are high-level talents with translational medicine awareness and substantial new drug R&D experience. Additionally, we hope to recruit people with excellent business sense.

My experience is that talent isn't trained; it's attracted and selected. What we specifically do is put our vision "on the public screen," light a torch, find the right direction, and people with the same beliefs naturally follow. You must give everyone a vision while also watching your feet — having a implementable goal while avoiding pitfalls along the way as much as possible.

Weiwen Qian: I very much agree with what Mr. Ji said. For traditional enterprises, high-end talent is indeed hard to recruit. My understanding is that bringing in high-end talent should inherently be about having shared vision and mission, recognizing the company's culture, and doing things together.

Wen Wen: Early on I thought recruitment would be very difficult because the company was very small and unknown. But because what we do is systematizing and digitizing immunity, this vision attracted many excellent talents — which was unexpected at the time. Having such a goal does attract good people.

The next step is: after people from different fields come in, how do they communicate with each other? People usually assume cross-disciplinary work has high cognitive barriers, that what the other person does is very difficult and I definitely won't understand. Much of this is psychological barrier. I often tell my team not to think it's so difficult — you're also excellent talent, an expert in your field, and understanding another field is a matter of learning. Once psychological barriers are lowered, mutual exchange and learning become much smoother.

Additionally, among cross-disciplinary talents, there may indeed be very few who are truly proficient, but they do exist. There are many abroad as well, which is why we decided to recruit an international team. There really are people who understand biology, computation, AI, and drug development. You might consider recruiting international teams too — science has always been an internationally collaborative endeavor.

For the final question, I'll give everyone time to speak freely. Looking ahead, what do you think your field will look like in 5-10 years in terms of products or applications?

Yan Tan: I believe that in the next 5-10 years, organic life drugs will emerge in large numbers, including microbial drugs, cell drugs, and engineered microbial and cell drugs. The human body is a complex system; our drugs will inevitably move toward complex systems. Drugs will become more diverse and increasingly complex. Additionally, live therapeutics will increasingly move toward personalization and precision.

Yuchong Wang: I thought about the healthcare industry ten years ago and feel that based on judgments then, it would have been hard to imagine reaching today's heights. The entire industry is in continuous accelerating development. When Cellnovo Biotech was founded we set two ten-year visions, which both felt very ambitious at the time: first, to participate in the industry and collectively witness cancer becoming a controllable chronic disease. Second, to bring order-of-magnitude efficiency improvements to innovative medicine through informatization.

Xuwo Ji: 5-10 years may be hard to say, but if we extend the time scale slightly — looking at 10 or 20 years — I believe "massive omics data-driven new drug R&D" will definitely accelerate beyond today's pace. We're already seeing acceleration in the past 3-5 years: on one hand, rapid accumulation of omics data; on the other, rapid development of deep learning technology, with progress in the recent 3-5 years exceeding the previous 15 years. Going forward, I think it will be very clear: data carries everything and drives everything.

Weiwen Qian: The pharmaceutical industry itself carries humanity's vision and expectations for a better life. The industry's development has entered an acceleration over the past decade; perhaps in the future we can have an even better vision where most humans live longer than before, with much higher quality of life.

Wen Wen: I quite agree with what several have said: one is precision, the other is digitization. If we imagine life sciences ten years ago, the Human Genome Project was such a massive undertaking at the time — to sequence one person's genome required hundreds of scientists worldwide spending a very long time. Today, it costs a little over ten dollars to get your genes sequenced, which was unimaginable ten years ago.

So in the next ten years, as IT technology, AI algorithms, detection technology, and engineering technology develop, human body digitization will go deeper. Every omics layer can be obtained at very low cost, models can be built, and truly precise individualized medicine can be achieved. This is what we're working toward together.

5Y Capital (formerly Morningside Venture Capital) currently manages approximately five billion USD across USD and RMB dual-currency funds. We believe that if the crazy you in others' eyes begins to be believed in, the world will be a better place.

BEIJING · SHANGHAI · SHENZHEN · HONG KONG

WWW.5YCAP.COM