AI4S: From Tool to Partner | Decoding the New Industrial Implementation Path of Life Sciences × AI

When artificial intelligence is no longer just a single computing tool in the laboratory, but becomes a core partner deeply integrated into the entire process of life science research and development and production, the industrial transformation of biomedicine and biomanufacturing has reached a new tipping point.

The "AI4S: From Tool to Partner" themed salon, jointly hosted by Nest.Bio and Google Cloud, recently concluded successfully.

The entire event revolved around in-depth discussions on the technological breakthroughs, industrial applications, and industry pain points of AI in the life sciences field. It not only fully demonstrated the industry's high attention to AI large models, Physical AI, and automation compliance, but also provided valuable first-hand insights for the next stage of industry development and market expansion.

This salon invited six distinguished guests from top global institutions and local innovative enterprises to share in-depth insights and first-hand practical experience across all dimensions, from cutting-edge basic research, industrial biomanufacturing, new drug development, automation implementation, compliance empowerment to underlying computing power support.

The presentation began with a speech by Joseph Ledsam, Google Health for JAPAC Leader. Starting with the development of computational biology over the past decade, Joseph Ledsam systematically reviewed DeepMind's landmark breakthroughs in protein structure prediction.

From AlphaFold 1, which validated the potential of deep learning, to AlphaFold 2, which solved a problem that had plagued the academic community for 50 years, and then to the latest iteration AlphaFold 3—by introducing diffusion networks, it has achieved high-precision prediction of the structure and interactions of all living molecules, with a leapfrog improvement in accuracy compared to existing methods.

He also shared AlphaMissense's practical application in predicting pathogenic missense mutations, as well as Med-Gemini's outstanding performance in medical multimodal data processing and long text reasoning scenarios, setting new technological boundaries for AI to empower basic research in life sciences.

Next, Liu Dapeng, founder and CEO of AlxBio, discussed with everyone how generative AI can reshape the R&D efficiency of industrial biomanufacturing.

He pointed out that the technical logic of AI large language models is also applicable to interpreting genetic instructions and sequence information within organisms.

Based on a dedicated AI model trained on high-quality private data, AlxBio can complete the construction of transcription maps in seconds, which would take months using traditional methods. It can also accurately simulate gene intervention results in the "dry experiment" stage.

This "dry-wet closed-loop" R&D model significantly shortens the development cycle of industrial strains, while also significantly improving the production efficiency and R&D success rate of the fermentation industry.

Wang Yinan, founder and CEO of Noah.AI, presented a brand-new concept: an "AI-native virtual biotechnology company."

He proposed that the ultimate role of AI should be that of a co-scientist for researchers.

He shared that by building an intelligent system that includes multiple types of AI agents, it is possible to achieve autonomous reading of literature, automated design of experiments, efficient screening of targets, and even autonomous generation of research hypotheses.

Among them, multimodal models that combine text, visual and genomic data have shown great potential in the early stages of drug discovery, while intelligent agent networks based on large models are expected to completely restructure the cost structure and time cycle of new drug development in the future.

Li Zhengyuan, head of Bota AI products at Enhe Technology, asked: What kind of energy will be unleashed when AI steps out of the digital world and deeply integrates with physical laboratories?

Li Zhengyuan focuses on the engineering application of "Physical AI" in life science laboratories, directly addressing the long-standing data gaps and operational bottlenecks in the industry.

He detailed the “AI-driven biological execution system” that integrates large-scale model technology with high-throughput automated laboratories: the system can directly understand natural language experimental instructions and automatically convert them into operation codes that can be executed by laboratory robots, realizing a fully automated closed loop from “experimental design” to “sample testing”. At the same time, it can collect high-quality experimental data in real time to feed back AI model iteration, truly realizing full-process intelligentization of biological experiments.

Song Naibing, a pre-sales architect at NEBULA DATA, discussed with everyone how AI can reduce costs and increase efficiency in the compliance management of pharmaceutical companies.

Song Naibing presented a GMP intelligent suite built on Google Cloud and AI technology, which directly addresses the core pain points of compliance and QA review in the pharmaceutical industry.

He shared three core application scenarios on site:

First, we leverage Gemini and Document AI to achieve full automation of GMP document review;

Secondly, by constructing an audit inspection agent through a large language model, it can cross-reference batch production data and CAPA documents to automatically generate standardized audit responses;

Third, we will build an intelligent question-and-answer robot based on industry regulations and enterprise SOPs to provide real-time compliance guidance for the production workshop.

The entire solution not only improves compliance efficiency through AI, but also comprehensively protects the data compliance and privacy security of enterprises (in accordance with international standards such as GDPR).

Pranav Mehrotra and Edward Li of Google Cloud APAC stated that Google Cloud empowers practical applications of AI in life sciences. The successful implementation of all AI applications relies on the solid support of underlying computing power and the platform.

The two guests systematically introduced the dedicated computing power and AI platform capabilities that Google Cloud has built for life science companies, focusing on the core value of dedicated computing infrastructure (such as TPU) and the Vertex AI platform.

Through Vertex AI Model Garden, enterprises can deploy and fine-tune world-leading AI models, including Gemini and AlphaFold, with a single click. At the same time, they can quickly complete the development and implementation of enterprise-level AI applications while meeting stringent standards for data privacy (CMEK encryption) and industry compliance (HIPAA/HITRUST), providing full-stack technical support for the AI ​​transformation of life science companies.

The final segment was a high-level roundtable discussion, which directly addressed industry pain points and explored the deeper aspects of AI implementation.

The roundtable discussion was moderated by Mr. Li Ming, and featured dialogues from several industry pioneers, including Liu Yuyang, Wang Yinan, Warren Li, and Li Zhengyuan. The interaction was lively, with enthusiastic questions from the audience, and the discussion ultimately yielded many insights and consensuses that directly addressed the essence of the industry.

The application of AI in the life sciences has completely moved beyond the early "proof of concept" stage and has officially entered the deep waters of "engineering and industrialization".

However, in the process of scaling up industrial applications, the industry still faces three core pain points and challenges:

Despite the rapid iteration of computing power and algorithms, high-quality, standardized closed-loop data from "dry and wet experiments" remains the biggest bottleneck restricting the prediction accuracy and practical application of AI models.

The industry has an unprecedented and urgent need for interdisciplinary talents. Those who understand the underlying logic of AI algorithms and are also well-versed in the mechanisms of biological and pharmaceutical research and development are still the core and scarce resources for industrial development.

Pharmaceutical companies have a clear need to introduce AI technology, but there are still many practical problems that need to be solved in key areas such as cross-border data compliance, patient privacy protection, and the verification and auditing of AI systems under GMP conditions.

AI in life sciences has entered the deep waters of industrialization. Data, talent, and compliance are the three core challenges. Only through industrial collaboration can the transformative value of the technology be unleashed.

This "AI4S: From Tool to Partner" themed salon not only built a high-quality platform for practitioners in the life sciences and AI fields to exchange cutting-edge technologies and connect with industry resources, but also pointed out the direction for the next stage of industry development through first-hand practical sharing and in-depth industry discussions.

In the future, Nest.Bio and Google Cloud will continue to delve into the intersection of life sciences and AI, bringing more high-quality industry exchange activities and witnessing a new future where AI empowers life sciences together with all industry colleagues.

Nebula Data, headquartered in Singapore, has branches in Jakarta, Guangzhou, Shanghai, and Hong Kong. The company independently developed Nebula Lab, a one-stop AI content generation and model aggregation platform, equipped with an enterprise-grade AI Agent, aggregating globally applicable large-scale models and industry-specific vertical models. Simultaneously, it launched the Nebula AIoT hardware ecosystem (including smart interactive terminals, IoT gateways, and other products), forming a full-link intelligent solution from cloud to edge to device. This provides integrated services to customers in e-commerce, manufacturing, retail, and other fields, from cloud computing power support and AI intelligent decision-making to terminal scenario implementation. Furthermore, it offers global AIDC (AI Intelligent Computing Center) + low-latency network services, empowering enterprises to embrace AI, connect to the physical world, and expand their global business through its technological foundation.