MaKe | DeepWise Completes Over 800 Million Yuan Series C Financing

Recently, MaHui member DeepWise announced the completion of its Series C financing round totaling over RMB 800 million. The round was jointly funded by Fortune Venture Capital, Jingguorui, Beijing Artificial Intelligence Industry Investment Fund, Beijing Medical and Health Industry Investment Fund, Lenovo Capital and Incubator Group, Yuanhe Puhua, and other institutions. Source Code Capital was a co-lead investor in the company's Series B round.

The proceeds will be primarily used to continue attracting and cultivating top industry talent, further evolving DeepWise's "Scientific Discovery Intelligence Engine," solidifying its full-stack capabilities from foundational technology innovation to intelligent research tools and industry solutions, and accelerating the market expansion and scaled application of intelligent scientific discovery products and services in basic research, life sciences, and materials science. The completion of this financing marks a solid step forward for DeepWise in its mission to build a next-generation scientific discovery intelligence engine.

Today, AI for Science has become a global consensus, with the fundamental goal of transforming how humanity explores the unknown and discovers new scientific knowledge and assets. In August 2025, China's State Council released an opinion on deepening the implementation of the "AI+" initiative, placing "AI + scientific research" at the top of its priorities, with particular emphasis on accelerating scientific discovery and driving innovation and efficiency gains in technology R&D models. Meanwhile, Europe's Horizon program has made AI-enabled scientific research a key focus; the United States has launched the Genesis Mission, centered on leveraging AI to accelerate scientific breakthroughs and improve the efficiency of AI-driven scientific discovery and industrial application — a program elevated to strategic parity with the Manhattan Project. Tech giants including Google DeepMind, NVIDIA, and Microsoft continue to invest and deploy resources in this space, while venture capital markets are actively backing innovative companies in the field. This global undertaking aimed at enhancing humanity's fundamental capacity for innovation and fully tapping the wealth of scientific knowledge has officially entered its Age of Exploration.

Behind this global consensus, four core missions of AI for Science are coming into focus. Scientific research is the process by which scientists use a set of scientific tools to explore the world. In the past, scientific productivity was severely constrained by outdated and inefficient traditional tools, as well as by the limited number of brilliant minds actually participating in research and the difficulty of effectively amplifying and combining human intelligence. Now, against the backdrop of AI deeply empowering the intelligent transformation of traditional research tools and AI agents creating new participants in scientific discovery, AI for Science is bringing systemic restructuring opportunities to the scientific discovery system: AI activating scientific data, AI reshaping scientific software, AI driving scientific instruments, and AI creating scientists.

As a global pioneer and leader in AI for Science, DeepWise has built a "Read, Compute, Act, Intelligence" capability system centered on the "Bohr Research Space Station," along with a Science as a Service intelligent research product and service matrix: Bohr Scientific Navigation, Bohr Lebesgue Intelligent Computing, and microscale R&D software including Hermite and Piloteye, Bohr Cyber Laboratory, SciMaster scientific intelligence agents, and "large-scale facilities" and R&D services for scientific discovery — providing deep yet flexible, combinable solutions for scientists and R&D organizations in basic research, life sciences, and materials science.

To date, DeepWise's Bohr Scientific Navigation has served over 3 million scientists from more than 1,000 universities and organizations worldwide, including nearly 100 Project 985 and Project 211 universities such as Peking University, Shanghai Jiao Tong University, and Wuhan University that have fully adopted the platform. It has supported over 1,000 research projects, answered an average of 12 million scientific questions annually, and saved scientists over 2 billion minutes of work time.

DeepWise's scientific intelligence products and solutions have helped over 150 advanced R&D enterprises upgrade their research intelligence, deeply empowering more than 100 R&D pipelines across over 70 life science companies including Fosun Pharma, Sinopharm Group, Hansoh Pharmaceutical, Huadong Medicine, East Sunshine Pharmaceutical, and Qilu Pharmaceutical, as well as Unilever and Yunnan Baiyao. In materials science, it serves clients including Suzhou National Laboratory, PetroChina, China Iron & Steel Research Institute, CATL, BYD, and GAC. These partnerships have helped create over 50 high-value scientific assets. According to partner measurements, intelligent upgrades to R&D systems represent an investment with extremely high returns: team literature research and organization efficiency improves 100x; the introduction of AI computing methods reduces wet lab demand and costs by 76%; intelligent laboratory adoption increases instrument utilization efficiency and experimental throughput by more than 3x; and "AI scientists" can reduce scientists' tedious, repetitive work by approximately 70%.

The underlying infrastructure supporting this rich product service system and research ecosystem is the "DeepWise · Yuzhi" foundation that the company has spent seven years continuously building. "DeepWise · Yuzhi" began as a pre-trained large model system for scientific domains and has now been fully upgraded to a Scientific Discovery Intelligence Engine: driven by scientific intelligence agents, it connects the "Read-Compute-Act" closed loop, constructing the shortest path to humanity's unknown knowledge for AI4S scenarios — reading to integrate existing knowledge, computing to explore and generate unknown spaces, acting to complete validation loops — thereby enabling "AI scientists" to become discovery entities capable of learning, thinking, executing, and feedback.

Around this Scientific Discovery Intelligence Engine, DeepWise has unified scientific data, computing, and experimental capabilities into callable R&D infrastructure: Bohr Scientific Navigation has integrated over 170 million high-quality English literature items, over 200 million patents, and 80 million Chinese literature knowledge items; vertical application models built on AI4S large models for atoms, molecules, genes, proteins, and other directions now exceed 1,000; through the Bohr UniLabOS intelligent laboratory operating system, over 100 frequently used experimental instruments have been integrated; through automated compilation and deployment capabilities, over 50,000 scientific tools have been supported for unified invocation in Agent-Ready form. The platform is serving over 3 million scientist users from 1,000+ universities and research institutions worldwide, continuously generating usage and feedback from real research tasks, gradually forming an ecological cycle where "research tools — research content — researchers" reinforce one another.

Building on the capabilities and ecosystem accumulated through the "DeepWise · Yuzhi" Scientific Discovery Intelligence Engine, DeepWise is driving "AI scientists" from point applications toward intelligent scientific production systems: collaborating with ecosystem partners to abstract research workflows and methodologies from different disciplines into combinable intelligence agent modules, enabling research intelligence agents to be rapidly built, continuously evolved, and constantly spawn new "Master Agents." As a representative achievement, ML-Master has achieved leading performance in top benchmarks such as HLE and MLE; MatMaster, jointly released with Suzhou National Laboratory, demonstrates the system's落地 and extension capabilities in specific disciplines. The resulting efficiency advantages stem from long-term accumulation of closed-loop infrastructure and real usage feedback, forming core barriers that cannot be easily replicated by single-point models or standalone tools.

(The DeepWise · Yuzhi large model system has been fully upgraded to the Scientific Discovery Intelligence Engine)

Just as search engines built the shortest path to humanity's known information, and large language models further built the shortest path to humanity's known knowledge, AI4S is now building the shortest path to humanity's unknown knowledge. DeepWise's mission and vision is to develop AI scientists capable of helping humanity discover entirely new scientific achievements, along with a series of intelligent systems capable of autonomous scientific discovery, making scientific discovery as simple as using a search engine. Based on this new paradigm, the company aims to empower the intelligent upgrade of global R&D systems, liberating scientists from tedious, repetitive labor to focus on creative inspiration, thereby systematically accelerating the process of scientific discovery and application, and enabling the world's annual $2.8 trillion, nearly 100 million full-time-equivalent investment in research and development to generate stronger innovation performance.

Linfeng Zhang, Founder and Chief Scientist of DeepWise, stated: "At DeepWise, we see AI for Science not merely as an emerging track, but as an infrastructure construction project for scientific discovery over the coming decades. From the DeepWise · Yuzhi® large model system to research platforms and solutions for different domain scenarios, to domain-specific AI scientists such as PharMaster and MatMaster, DeepWise has consistently made long-term investments around 'real scientific problems' and 'real industrial needs.' What DeepWise hopes to achieve is a self-consistent, operational Scientific Discovery Intelligence Engine, where AI is not merely an accelerator for one particular环节. Through this financing round, DeepWise will further focus on enhancing this engine's usability and evolutionary capability in real research scenarios, on one hand driving the continuous emergence of cutting-edge scientific capabilities, and on the other hand accelerating its落地 application in critical domains such as life sciences and materials science, promoting the continuous emergence of 'AI scientists' for various fields."

Weijie Sun, Founder and CEO of DeepWise, stated: "This financing round comes at a critical stage when the nation is deeply implementing the 'AI+' initiative and AI for Science has been elevated to the forefront of global technology competition. It represents both recognition of DeepWise's阶段性 progress and an entrustment of our next-phase mission. Going forward, DeepWise will continue to uphold the philosophy of 'accelerating scientific discovery, unleashing scientific value,' with the long-term goal of building 'AI scientists' and 'scientific discovery intelligent systems,' further accelerating value creation from AI for Science in basic research and industrial R&D. We will strive to grow into a technology company originating from China and leading the world, enabling the new generation of intelligent scientific infrastructure to generate greater international influence."