AI News AI资讯 3h ago Updated 1h ago 更新于 1小时前 49

National Team Enters AI Virtual Cells, 'Baiyao Technology' Completes New Round of Financing Tens of Millions of Yuan | 36Kr Exclusive 国家队下场做AI虚拟细胞,「百曜科技」完成数千万元新一轮融资|36氪首发

Baiyao Technology's securing of tens of millions in financing led by China Reform Holdings may appear on the surface as an inevitable outcome of a booming sector and a strong team, but at its core, it resembles a precisely timed signal flare: the seemingly ultra-sci-fi competition of AI Virtual Cells (AIVC) has officially transitioned from "scientists' paper concepts" into the "range of state-backed capital" on the Chinese track. 百曜科技拿到国新创投领投的数千万元融资,这事儿表面看是赛道火热、团队硬核的必然结果,但内核里,它更像一记精准的信号弹:AI虚拟细胞(AIVC)这场听起来极其科幻的竞赛,在中国赛道上,正式从“科学家的论文概念”进入了“国家队资本的射程范围”。

75
Hot 热度
70
Quality 质量
65
Impact 影响力

Analysis 深度分析

Where the money comes from often reveals more than where it goes. China Reform Holdings is a national-level state-owned capital operating platform. Its role as the lead investor elevates this beyond a typical VC bet on a trend—it constitutes a "stamped certification" with strategic undertones. What it's investing in is not merely Baiyao Technology as a specific company, but a crucial layer on the grand map of "AI + Life Sciences": cell-level simulation. This marks a fundamental divergence from the previous logic of flocking to AI-driven pharmaceutical development and molecular screening. The latter is akin to optimizing paths on an existing map, whereas AIVC attempts to redraw the underlying map of life systems itself. China Reform's entry effectively signals national-level recognition of the strategic reserve value of this "map-redrawing" capability.

However, amidst the excitement, a cold shower of reality is necessary: the gap between a "model with second-best global predictions" in the lab and "precise predictions of drug toxicity" on the production line is not a small ditch, but a canyon filled with data black holes, validation challenges, and an industry cognition chasm. Baiyao team's academic credentials are impeccable, from the Institute of Zoology of the Chinese Academy of Sciences to cover articles in Cell Research, and their technical narrative is polished. But scientific "polish" and commercial "usability" are two different things. The core promise of AIVC is "model-driven R&D," replacing the traditional "experiment-driven" approach. This sounds revolutionary, but anyone with basic respect for biological experiments would ask: on what basis? Why believe that a digital cell trained on public omics data can truly simulate the complex, chaotic, and individually variable real cell systems within the human body? "Hallucinations" in models mean nonsense in the linguistic domain; in life sciences, they could mean tens of billions of dollars in trial-and-error.

This raises an even sharper question: how much of the current heat in the AIVC track stems from solid technological conviction, and how much is a concept premium manufactured by capital seeking the "next grand narrative" in the post-AI drug discovery era? With AI applications at the molecular level already crowded into a red ocean, investors urgently need a new, more disruptive-sounding story to maintain market excitement. "Cell-level simulation" perfectly fits the bill—it's grand, deep, distant, and inspires imagination. Consequently, we see the passionate vocabulary in investor viewpoints: "disruptive," "most imaginative space." In venture capital, these words are stimulants. But on the long march of technological implementation, they should more appropriately serve as a ruler to constantly measure the distance between reality and imagination.

Baiyao Technology mentioned building a "data closed loop," where models, self-generated data, and application scenarios mutually reinforce each other. This is the correct direction, but also the toughest nut to crack. High-quality, continuous, perturbation-responsive time-series cellular data is the oil for AIVC, and such data is currently extremely scarce and expensive. Their plan to launch a self-generated dataset platform is undoubtedly playing their hardest, heaviest move. Without this, all models are castles in the air. But building such a platform—its cost, complexity, and time required—may far exceed the surface figure of the financing. These tens of millions might be enough to drill a deep well, but the distance to building a continuously producing oil field is still vast.

Therefore, this financing is more a serious beginning than a victory fanfare. It marks AIVC's transition from an endeavor for geeks to an important piece on the chessboard. But whether that piece can become an ace depends on every subsequent step: whether the model withstands repeated crucifixion in wet-lab experiments, whether the data flywheel can truly spin, and whether the team can withstand the strategic demands and time pressure that may accompany the "national team" capital without being derailed by the market's inflated valuation expectations.

The real contest has only just begun. Baiyao Technology stands at an excellent starting position, but the track ahead is half imagination of a vast future and half the harsh reality of muddy, bumpy terrain. What it needs is not just algorithms and data, but a nearly obsessive engineering capability and industrial patience that can keep its feet firmly on the ground while gazing at the stars. After all, in the field of life sciences, the end of any shortcut is often a longer detour.

百曜科技拿到国新创投领投的数千万元融资,这事儿表面看是赛道火热、团队硬核的必然结果,但内核里,它更像一记精准的信号弹:AI虚拟细胞(AIVC)这场听起来极其科幻的竞赛,在中国赛道上,正式从“科学家的论文概念”进入了“国家队资本的射程范围”。

钱从哪来,往往比钱到哪去更能说明问题。国新创投是国家级国有资本运营平台,它当领投方,这不再是寻常的VC押注风口,而是一种带有战略色彩的“盖章认证”。它投的不只是百曜科技这家特定公司,而是“AI+生命科学”这张大图上的一个关键图层——细胞级别的模拟。这和之前扎堆投AI制药、做分子筛选的逻辑有了本质区别。后者更像在已有地图上优化路径,而AIVC则试图重新绘制生命系统的底层地图。国新的入场,等于说在国家层面上,认可了这种“重绘地图”能力的战略储备价值。

但热血沸腾的背后,必须泼一盆冷水:从实验室的“模型预测全球第二”到生产线上的“精准预测药物毒性”,中间隔着的不是一条小水沟,而是一条布满数据黑洞、验证难题和产业认知鸿沟的峡谷。百曜团队的学术背景无可挑剔,从中科院动研所到Cell Research封面文章,技术叙事很漂亮。但科研的“漂亮”和商业的“能用”是两码事。AIVC的核心承诺是“模型驱动研发”,取代传统的“实验驱动”。这听起来是革命,但任何对生物学实验有基本敬畏的人都会问:凭什么?凭什么相信一个基于公开组学数据训练出来的数字细胞,就能真正模拟出人体内那个复杂、混沌、充满个体差异的真实细胞系统?模型的“幻觉”在语言界是胡说八道,在生命科学界,可能意味着数十亿美元的试错。

这就引出了更尖锐的问题:目前AIVC赛道的热度,有多少是扎实的技术信仰,又有多少是后AI制药时代,资本为了寻找“下一个宏大叙事”而制造的概念溢价?当分子层面的AI应用已经卷到红海,投资人们迫切需要一个新的、听起来更具颠覆性的故事来维持市场的兴奋度。“细胞级模拟”这个故事无疑完美符合要求——够大、够深、够遥远、够令人遐想。于是,我们看到了投资人观点里那些充满激情的词汇:“颠覆性”、“最富想象空间”。这些词汇在风险投资领域是兴奋剂,但在技术落地的长征路上,更应该是一把尺子,时刻用来丈量现实与想象的距离。

百曜科技提到要构建“数据闭环”,即模型、自产数据和场景相互反馈。这是正确的方向,也是最硬的骨头。高质量、连续、扰动下的时间序列细胞数据,是AIVC的石油,而这种数据目前极度稀缺且昂贵。他们计划落地自产数据集平台,这无疑是在下最苦最重的一手棋。没有这个,所有模型都是空中楼阁。但建设这样一个平台,其成本、复杂性和所需时间,可能远超融资额的表面数字。这数千万元,或许够打一口深井,但离建成一个持续产出的油田,还有漫长距离。

所以,这笔融资更像是一个严肃的开端,而非胜利的号角。它标志着AIVC从极客的极客变成了棋盘上的重要棋子。但棋子能否变成王牌,取决于接下来的每一步:模型是否经得起湿实验的反复拷打,数据飞轮能否真正转起来,以及,团队能否顶住“国家队”资金背后可能伴随的战略要求与时间压力,同时不被市场过高的估值预期带偏节奏。

真正的较量现在才开始。百曜科技站在了一个极佳的起跑位置上,但前方的赛道,一半是星辰大海的想象,一半是泥泞坎坷的现实。它需要的不只是算法和数据,更是一种在仰望星空的同时,能死死盯住脚下每一步的、近乎偏执的工程化能力和产业耐心。毕竟,在生命科学领域,任何捷径的终点,往往都是更长的弯路。

Disclaimer: The above content is generated by AI and is for reference only. 免责声明:以上内容由 AI 生成,仅供参考。

融资 融资 医疗AI 医疗AI 科学研究 科学研究 大模型 大模型
Share: 分享到: