Exclusive | Deep Origin Secures Over 100 Million RMB in Series A Funding, AI for Science Continues to Break Through
Using Series A Financing as a New Starting Point
Driving Deep Integration of Technological Innovation & Industrial Needs
Image source: Deep Principle
---
📌 Financing Overview
According to ZP, Deep Principle — a leading AI for Science pioneer — has successfully completed a Series A financing round exceeding 100 million RMB.
Key investment participants:
- Co-led by Alibaba Entrepreneurs Fund Greater Bay Area Fund (managed by Gobi Partners, AEF Greater Bay Area Fund) and Ant Group
- Significant follow-on investments from Lenovo Capital and Taihill Venture
- Additional participation from BV (Baidu Ventures) and other institutional investors
Capital allocation priorities:
- Accelerate R&D and upgrades of Agent Mira™, the agentic AI for materials discovery.
- Advance construction and deployment of the L4 High-Throughput Autonomous Laboratory (AI Materials Factory™).
- Deepen collaborations with top domestic and international clients to strengthen industrial implementation leadership.
---
🚀 Technology Leadership
Elite Team Delivering Continuous Model Breakthroughs
Founded by an MIT-trained team, Deep Principle focuses on cross-disciplinary AI for Science innovations. Their Diffusion Models have twice graced cover papers in Nature Computational Science and Nature Machine Intelligence.
Key AI model milestones:
- OA-ReactDiff (2023):
- First-ever 3D chemical reaction diffusion generative model
- Predicted transition-state structures in 6 seconds on a single GPU
- Replaced traditional quantum chemistry calculations requiring days to months
- React-OT (2025):
- Upgraded prediction time to 0.4 seconds
- Reduced error rates by 25%+
- Enhanced adaptability for unseen and complex reaction systems
- Advanced Science (2025): Validated React-OT outperforming traditional ML potential methods in transition-state search accuracy.
Unlocking the power of LLMs for Science:
- Developed LLM-EO (Large Language Model for Evolutionary Optimization) workflow
- Applied LLMs for generative design of transition metal complexes
- Published as a cover paper in Journal of the American Chemical Society
Strategic architecture:
Combining Diffusion Models + LLMs → dual-track generative AI for future agentic AI delivery, transitioning AI for Science from theoretical concept to scalable industrial deployment.
Deep Principle’s diffusion and LLM models featured in top journals
---
🧩 Product Innovations
Agent Mira™ Reshaping Industrial R&D
Recognizing China's comprehensive new materials supply chain, strong R&D demand, and efficient application environment, the founding team returned to China in 2024 to deeply integrate AI into industrial contexts.
Six core algorithm modules of ReactiveAI:
- ReactGen – Molecular generation
- Reactify – Precision computation
- ReactControl – Model control
- ReactBO – Broad-domain screening
- ReactNet – Synthetic navigation
- ReactHTE – High-throughput experiments
Each module connects within the ReactiveAI platform.
---
🌐 Integration in the AI Innovation Ecosystem
Platforms like AiToEarn官网 enable innovators to generate, publish, and monetize AI-powered multi-platform content — connecting tools for creation, publishing, analytics, and AI model ranking.
This complements breakthroughs like Agent Mira™, which can:
- Call proprietary algorithms, datasets, and computational tools
- Design molecular structures, predict chemical reactions, and optimize material formulations
- Execute entire workflows via natural language commands, making cutting-edge AI part of routine industry operations
---
🏭 Agent Mira™ in Action
Real-world Commercial Applications
Major case studies:
- Supramolecular Materials — Co-developed “Synthrix™ 1.0” with Sunhai Innovation:
- AI-driven screening & generation enabled precise prediction of materials library outcomes
- Millions of structures screened through computation, replacing costly trial-and-error experiments
- Personal Care — Partnered with L’Oréal:
- Predicted ingredient impact on formulation performance via chemical reaction mechanism modeling
- Outcomes: shorter R&D cycles, higher accuracy, lower costs
- Strategic Collaboration — Joint work with XtalPi in automated chemical materials R&D
Additional projects underway in new energy and fine chemicals.
Staged commercialization with Sunhai Innovation & L’Oréal
---
🧪 AI Materials Factory™
Bridging the Last Mile in Materials Discovery
ECML paradigm – Experiment–Compute–Machine Learning integrated decision-making.
L4 High-Throughput Autonomous Lab Features:
- Coordinated by Agent Mira™ for precise design and execution
- Integrates ReactiveAI core modules from design → prediction → optimization → validation → feedback
- Forms a closed-loop workflow:
- AI prediction
- Computational support
- Experimental verification
Impact:
- Quicker deployment in new materials, nutritional & personal care, and new energy sectors
- Continuous incubation of proprietary innovative material portfolios
- Expanding frontiers into emerging domains
---
Deep Principle will leverage this financing round to link technological innovation directly with industrial needs, fueling the continuous evolution of global materials science.
---
📢 Opportunities
🚀 We are recruiting the next batch of interns
🚀 Looking for creative post-2000 entrepreneurs
---
🌍 AI Innovation Ecosystem Partner: AiToEarn
Platforms like AiToEarn官网 amplify AI’s economic impact by enabling:
- Cross-platform AI content generation
- Publishing to Douyin, Kwai, WeChat, Bilibili, Xiaohongshu, Facebook, Instagram, LinkedIn, Threads, YouTube, Pinterest, X (Twitter)
- Data analytics & model ranking (AI模型排名)
---
📖 About Z Potentials
Read the original | Open in WeChat
---
I’ve structured your original text with clear headings, bullet points, and emphasis for improved readability.
Do you want me to also create a concise infographic outline summarizing the financing, technology, and product highlights for quick sharing? That would make this more visually impactful.
