"Deep Principle" Closes Nearly RMB 1 Billion in Total Series A Funding, Linear Capital Continues to Increase Stake | Linear Portfolio

y'x

Today, Deep Principle, a global leader in AI for Materials, announced the completion of its Series A fundraising, bringing its total raised to nearly one billion RMB.

As the lead investor in Deep Principle's seed round, Linear Capital continues to hold strong conviction in the company's technical capabilities and commercial progress, and participated again in this round.

On the morning of July 19, Neil Zeng, Partner at Linear Capital, will attend the AI-Driven Materials Industry Innovation Ecosystem Forum hosted by Deep Principle at the WAIC Expo Center as a representative of the investment community. All are welcome to follow along.

Recently, Deep Principle, a global leader in AI for Materials, completed its Series A fundraising, bringing its total raised to nearly one billion RMB.

The latest round, Series A3, was led by Fortera Capital, with participation from Huakong Fund, Bayland Capital, and others. More than half of existing shareholders — including Shunxi Fund, Hillhouse, Vertex Ventures China, Gobi Partners, Linear Capital, Legend Star, Baidu Venture, and Qigao Capital — significantly oversubscribed.

The completion of the Series A round further validates the market's strong recognition of Deep Principle's technical strength and business progress.

Deep Principle has continued to achieve technical breakthroughs, releasing multiple models and reaching comprehensive SOTA performance.

In property prediction, Deep Principle launched the MPA model, achieving SOTA results on nearly 40 experimental property prediction tasks;

In materials generation, it developed SAGA, enabling intelligent agents to autonomously adjust in complex materials generation tasks, demonstrating stronger multi-objective optimization capabilities than models developed by major global tech companies;

In chemical reaction generation, Deep Principle launched the OA-ReactDiff, React-OT, and other model series, opening the black box of static prediction and reducing transition-state generation from traditional timelines of days or months to 0.4 seconds, making AI truly usable in industrial practice.

Building on its full-stack SOTA models, Deep Principle developed the AI Scientist platform Mira: Mira can autonomously complete the design, execution, review, and dynamic optimization of long-chain scientific research workflows, continuously accumulating scientific knowledge and achieving closed-loop autonomous iteration, driving the construction of the AI era.

Deep Principle adheres to the core philosophy of "AI for Science, Science for AI," committed to unlocking new world-changing materials through AI. Deep Principle has consistently focused on bridging the "last mile" between AI prediction and experimental validation, enabling AI to truly impact physical-world experimentation.

Therefore, Deep Principle established the L4-level high-throughput laboratory AI Materials Factory, achieving automated experimental workflow design, equipment control, and data collection by Mira. With the AI Materials Factory now operational, it has become the first in the industry to achieve a complete closed-loop R&D pipeline: AI prediction — high-throughput experimental validation — knowledge accumulation — AI capability iteration.

It focuses on core sectors including computing infrastructure materials, new energy, superconductors, and carbon neutrality, tackling key materials challenges for next-generation AI technology deployment; this year it will continue to build out multiple R&D pipelines, accelerating the industrialization of frontier materials technologies.

With its full-stack closed-loop AI capabilities, Deep Principle's commercialization is continuously accelerating.

On one hand, it has already accumulated multiple top-tier international and domestic clients, including multinational giants, and has secured recurring business from them, fully demonstrating industry recognition;

On the other hand, its client partnerships have deepened, evolving to Agent-based subscriptions and pioneering co-development of materials pipelines in the industry.

Deep Principle's business scope continues to expand, rapidly covering fine chemicals, nutrition and personal care, new energy, electronic materials, and other important new materials sectors.

Deep Principle is now landing multiple single-client orders in the tens of millions of RMB, further validating and unleashing its potential for industrial-scale deployment. Additionally, Mira has cumulatively served over 10,000 professional individual users — materials researchers and R&D engineers at home and abroad: through long-term interaction with these professional users, it continuously optimizes its models and products, while leveraging this community as natural touchpoints for dissemination, driving Mira's adoption by their research teams, R&D departments, institutes, and industrial companies.

Deep Principle ranks among the global industry leaders in model technology, pipeline R&D, and commercialization. With the Series A funding complete, Deep Principle will further increase resource investment in these areas, accelerating the reshaping of the AI-driven materials R&D full pipeline and continuously expanding its international leading position in AI for Materials.

Event Preview

On the morning of July 19, Deep Principle will host the "From Token to Formula: AI-Driven Materials Industry Innovation Ecosystem Forum" at WAIC, officially launching the AI Scientist platform Mira, and joining academicians and industry ecosystem partners to explore the scalable path of AI-driven materials discovery. Neil Zeng, Partner at Linear Capital, will attend the roundtable forum as a representative of the investment community. All are welcome to follow along.