Source Code Capital | Physical AI Firm DeepWise Closes Hundreds of Millions of RMB in Series B Funding, Existing Shareholder Source Code Capital Increases Stake

Recently, MaHui member and physical AI company DeepWise (DeepCtrls) closed a Series B round of several hundred million RMB. The round was led by strategic investor JinkoSolar, with co-leadership from SDIC Venture Capital and CMB International. Existing shareholders Source Code Capital, HSG, Forebright Capital, and China Merchants Capital continued to increase their stakes, creating a strategic capital structure backed by a new-energy leader, a national-level investment platform, top-tier institutions, and a tech giant — all accelerating the global deployment and reconstruction of physical AI across computing and energy infrastructure.

Source Code Capital led DeepWise's Pre-A round in 2022 and has continued to make substantial follow-on investments across multiple subsequent rounds, providing long-term support for the company's growth. Lei Xin, Investment Vice President at Source Code Capital, commented: "We are bullish on the company's technical accumulation in the physical AI foundation, and on the moat it continues to build around data, models, and closed-loop control. Over the past few years, its products have been deployed at leading CSPs, FABs, and various high-end industrial scenarios, forming a replicable and scalable commercial闭环. We remain optimistic about DeepWise team's growth potential and execution capabilities, and look forward to continuing our journey alongside the company."

Dr. Hui Li, Founder and CEO of DeepWise, stated: "The global AI industry is crossing from the digital world into the physical world. Generative AI represented by DeepSeek has significantly enhanced AI's ability to think and reason in the digital realm; meanwhile, the physical AI that DeepWise DeepCtrls has been deeply cultivating enables AI to autonomously optimize and exercise closed-loop control in the real physical world — the ultimate form of AI industrial deployment. The endgame of AI is energy. DeepWise will continue to deepen its core PhyAI physical AI technology, driving global computing and energy infrastructure toward an L5 autonomous intelligence era, and building a sustainable future for AI computing power and energy."

DeepWise is a pioneer and leader in the global deployment of physical AI in the energy sector. As early as 2018, before the concept of physical AI had taken shape worldwide, the company was the first to propose and commit to a "mechanism framework + AI" technical approach. It developed its core PhyAI physical AI engine, deeply coupling physical mechanisms with AI to enable AI to truly "understand" the physical world. This fundamentally addresses three major challenges of pure AI: "untrustworthy black boxes, weak generalization, and difficult cross-scenario transfer." Every control decision becomes interpretable, predictable, and highly safe and reliable — solving the core pain points of industrial AI that is "unusable, unusable with confidence, and even less controllable with confidence."

Built on its self-developed PhyAI physical AI engine foundation, the company launched the industry's first L4-L5 energy intelligent agent product matrix: from large system-level DeepSYS agent (L4), standardized product DeepBot agent (L4.5), to device-level DeepChip control module (L5), and campus-level DeepOS computing-electricity coordination and source-grid-load-storage operating system — forming a complete technology stack from device terminal control to system dispatch, from single scenarios to full-domain coordination.

The company's products have already been deployed at TSMC, CATL, Foxconn Industrial Internet, Tencent, ByteDance, and hundreds of other global leading enterprises, covering core tracks including intelligent computing centers, semiconductors, and new energy. The company's globalization is advancing rapidly, with product sales already covering more than ten countries across Southeast Asia, the Middle East, and North America, contributing Chinese wisdom to the upgrade of global computing and energy infrastructure through physical AI technology.

With this funding round, the company will continue to solidify its PhyAI physical AI core technology foundation and broaden the application boundaries of global computing and energy infrastructure: from energy-efficiency AI, to source-grid-load-storage integration, to computing-electricity coordination — continuously refining and building the intelligent hub for global computing and energy infrastructure in the AI era.