In recent years, AI drug discovery has attracted considerable attention.

 

From identifying disease targets and designing drug molecules, to virtual screening and lead compound optimization, artificial intelligence is steadily permeating the drug R&D pipeline, giving rise to a number of high-profile companies such as Insilico Medicine and XtalPi. The focus of capital has long revolved around drug discovery.

 

Now, a new direction is beginning to enter the capital's field of vision.

 

AI Virtual Cell (AIVC) company Huayuan Zhiyin recently announced the completion of a multi-tens-of-millions-yuan seed funding round, led by Shuimu Ventures. Last month, Baiyao Technology also announced the completion of a multi-tens-of-millions-yuan new financing round, led by China Reform Venture Capital Fund, with follow-on investments from Daotong Capital and Yunqi Capital, and continued participation from Frees Fund and BV Baidu Venture Capital. Yaosu Technology, another virtual cell player, has been even more aggressive, raising a cumulative RMB 400 million this year alone.

 

Within a short period, several startups focused on AI virtual cells have secured financing in succession, drawing attention to this emerging direction.

 

Currently, virtual cells are still in the early stages of development. However, capital is already placing its bets ahead of the curve. So, has the investment wind shifted in AI drug discovery? Why is the virtual cell starting to gain traction?

 

1: AI Drug Discovery — Has the Investment Wind Shifted?

 

Looking back at the development of AI drug discovery over the past decade, the investment direction of capital has been remarkably concentrated.

 

In the early stages, AI was primarily used in drug discovery stages such as target discovery, protein structure prediction, molecular design, virtual screening, and lead compound optimization — with the hope of shortening R&D cycles and reducing costs through algorithms. In recent years, the continuous iteration of large models and generative AI technologies has further propelled AI from assisted analysis to active molecule generation.

 

This has given rise to a number of representative companies. Insilico Medicine has built a platform around AI-driven drug discovery and advanced multiple proprietary pipelines into clinical trials; XtalPi combines AI with robotic laboratories to build an automated R&D platform; and overseas companies such as Exscientia and Recursion have established their own technological advantages in AI molecular design and cell phenotypic analysis, respectively.

 

After years of development, AI drug discovery has gradually formed a relatively mature development path: AI platforms enhance drug discovery efficiency, ultimately realizing value and commercialization through proprietary pipelines, joint R&D, BD licensing, and other means. At the same time, as more companies concentrate on drug discovery, the technological routes and business models of AI drug discovery have gradually matured.

 

However, as the industry matures, a question has begun to emerge: Has AI drug discovery run out of new stories to tell? Is capital experiencing "aesthetic fatigue" with this track?

 

"The development of the AI therapeutics field up to now has mainly been about technological iteration, but it has all remained concentrated at the molecular design stage. At this stage, there is no real impact on the business model of AI drug discovery, so ultimately it still comes down to the value of the molecule itself. When an early-stage AI drug discovery platform is first established, it can rely on the platform's imagination and valuation logic to secure its first two rounds of financing. But ultimately, especially when the drug enters the clinical stage, it must return to the molecule itself. So the value of a project should be judged from the perspective of the molecule's value; the AI drug discovery platform may command a slight premium, but it cannot sustain the valuation entirely on its own." an investor at a well-known biotech investment firm told TXY.

 

This shift in capital focus does not mean that AI drug discovery has lost its appeal. Rather, as drug discovery matures, investment institutions are beginning to seek the next wave of technological innovation, pushing AI to extend beyond drug discovery into other stages of R&D.

 

Huayuan Zhiyin and Baiyao Technology, which recently secured financing, are not focusing on finding new drug molecules. Instead, they are attempting to use AI to simulate cellular processes, aiming to build digital models capable of predicting cell states and their responses to drugs and genetic perturbations.

2: Why Virtual Cells?

 

Virtual cells refer to the use of AI to integrate massive biological data — including single-cell sequencing, multi-omics, and spatial omics — to build digital models in silico that can simulate cell states and predict cellular responses to various perturbations such as drugs and gene editing.

 

In the past, AI primarily helped researchers answer the question "what molecules might be effective." Virtual cells aim to go further, addressing questions such as "what happens after a drug enters a cell," "why do different patients respond differently," and "which side effects can be predicted in advance."

 

What it seeks to simulate is not just a drug molecule, but the dynamic operational process of an entire cell — or even an entire biological system.

 

With the continuous maturation of single-cell sequencing, multi-omics, and AI foundational models, a growing number of researchers believe that in the future, a significant portion of experiments may first be validated in "digital cells" before entering physical experimentation — further reducing R&D costs, improving drug development efficiency, and expanding into application scenarios such as precision medicine, disease research, and cell therapy.

 

This is precisely why capital is beginning to position itself early. In recent years, overseas financing for virtual cells, AI cell models, and computational biology platforms has warmed up significantly. The most notable case is Xaira Therapeutics, which raised a $1 billion Series A round led by top life science investors including ARCH Venture Partners and Foresite Capital, with participation from tech investors including former Google CEO Eric Schmidt — making it one of the largest startup financings in AI biology history.

 

Earlier this year, Xaira Therapeutics announced the launch of its first virtual cell model, X-Cell, further driving market attention to the virtual cell track.

 

Meanwhile, in recent years, companies such as Turbine, Tahoe Therapeutics, Shift Bioscience, and Cellular Intelligence have also secured tens of millions of dollars in financing, with backing from internationally renowned venture capital firms such as Accel, General Catalyst, BGF, and F-Prime Capital, as well as corporate venture arms like Databricks Ventures — indicating that this track is attracting interest from both tech and life sciences capital.

 

The domestic market is also in the early stages of deployment, with financing activity gradually picking up in recent years. Since the beginning of this year, companies including Huayuan Zhiyin, Baiyao Technology, Yaosu Technology, and Wujie Jinhua have completed financing rounds. Notably, in addition to investment institutions with long-standing focus on healthcare such as Daotong Capital and Yunqi Capital, state-owned funds and tech investment institutions such as China Reform Venture Capital Fund, China Life Equity, BV Baidu Venture Capital, and Shuimu Ventures have also begun to make concentrated moves.

 

Behind this influx of capital, virtual cells are transitioning from a cutting-edge research concept to the stage of industrialization exploration.

 

3: The Next Investment Hotspot in Biopharma?

 

Although still in its early stages, many investors see significant promise in virtual cells.

 

This is because the true value of virtual cells lies not merely in helping develop a single drug, but in its potential to become an important infrastructure for the future era of AI biopharma.

 

The investor previously quoted told TXY: "The appeal of virtual cells lies in their potential to become a key component of biopharma infrastructure, upon which a series of other AI applications could be built."

 

If this vision is ultimately realized, then future drug development, toxicity prediction, disease mechanism research, cell therapy, and even precision medicine could all be built upon this foundational capability of virtual cells.

 

For capital, this means its potential value is not limited to a single drug, but rather has the opportunity to become a general-purpose capability — like large models and cloud computing platforms — that underpins industry-wide innovation. This is why many investment institutions are choosing to position themselves early, before the technology has fully matured, hoping to secure a gateway to the next-generation AI biology platform.

 

This trend is also evident in overseas financing cases. Whether it is Xaira Therapeutics receiving backing from top life science funds such as ARCH Venture Partners and Foresite Capital, or tech-investor-backed firms such as Accel, General Catalyst, and Databricks Ventures investing in related companies, the focus of capital is not merely on whether a company has drug pipelines, but on whether it can build AI models, data systems, and computational platforms with long-term competitive advantage. The investment logic is shifting from "investing in a company" to "investing in a foundational capability."

 

However, the industry still has a long way to go before reaching true maturity.

 

"The entire track is still at a very early stage. Many people are watching this direction, but there are very few companies, so valuations are rising rapidly. However, none of the companies have yet explored a mature commercialization path," the investor candidly told TXY.

 

In fact, this is a common challenge facing the global virtual cell track. Unlike AI drug discovery, which has already established relatively mature business models such as platform licensing, proprietary pipelines, and pharma partnerships, virtual cells are still primarily focused on research exploration and early-stage entrepreneurship. How to continuously accumulate high-quality, multimodal biological data, build sufficiently credible cell models, validate model predictions through extensive experimentation, and ultimately develop commercial products that can be widely adopted by pharmaceutical companies and research institutions — this will still require considerable time for technological accumulation and industrial validation.

 

Dr. Dan Liu, Managing Partner of Pivotal bioVenture China, also told TXY: "Virtual cells are actually even more 'virtual' — they don't have pipelines yet. But it is generally agreed that they should not be viewed through the lens of pharmaceutical logic, but rather from the perspective of a tech company — looking at what model problems it can solve."

 

In the previous wave of AI drug discovery, capital was betting on "finding a molecule faster." Today, with virtual cells, capital is betting on "understanding a cell more authentically."

 

From predicting molecules to simulating biological systems, AI is continuously expanding the boundaries of biopharmaceutical R&D capabilities. Whether virtual cells will ultimately grow into the infrastructure of the next generation of AI biopharma remains to be seen.

 

But one thing is certain: as technologies such as AI continue to evolve, transformation in the biopharma industry is on its way.