Shanghai State Capital Bets Nearly 50 Million Yuan on AI Engineering Foundation: UniForce AI Enters the Race
Add to Google Preferred Sources
UniForce AI (Shanghai) Intelligent Technology Co., Ltd. announced the completion of an angel+ round of nearly 50 million yuan (approximately $7.4 million), co-led by Shanghai Future Industry Fund and Shanghai Sci-Tech Innovation Group, with Fudan Sci-Tech Innovation adding a lead investment and Loongson Venture Capital and Chengwei Capital participating as co-investors. Founded in January 2026, the company targets the AI for Engineering sector, using controlled nuclear fusion as its first high-difficulty validation scenario to build an AI-native general-purpose foundation covering simulation, control, diagnostics, and design. Founder Wu Tailin holds academic credentials from Peking University, MIT, and Stanford, and currently serves as a distinguished research fellow at Westlake University. His algorithms have achieved industrial deployment of million-grid reservoir simulation at Saudi Aramco. The concentrated investment from Shanghai state capital signals that AI4Engineering is moving from a non-consensus phase into a critical period of technical scenario validation.
Key Elements
UniForce AI (Shanghai) Intelligent Technology Co., Ltd., an AI engineering company founded just seven months ago, announced the completion of an angel+ round of nearly 50 million yuan (approximately $7.4 million), co-led by Shanghai Future Industry Fund under Shanghai Guotou and Shanghai Sci-Tech Innovation Group, with Fudan Sci-Tech Innovation adding a lead investment and Loongson Venture Capital and Chengwei Capital participating as co-investors. Lighthouse Capital served as the exclusive financial advisor.
This represents a concentrated deployment by Shanghai’s state capital system in the AI for Engineering (AI4Engineering) sector. UniForce AI was founded in January 2026 and officially established its presence in Shanghai’s Zhangjiang area in early August. The company targets AI-native general-purpose foundation models for industrial applications, covering four core needs: real-time simulation, operational control, diagnostics, and design. Its application scenarios use controlled nuclear fusion as the benchmark, extending to energy, high-end manufacturing, aerospace, and other engineering fields.
The funds from this round will primarily be used for continued R&D of generative simulation and intelligent control engines, device-level validation of benchmark scenarios such as fusion, expansion into energy and high-end manufacturing industry clients, and building a world-class interdisciplinary talent team.
A Physicist’s Cross-Disciplinary Venture
UniForce AI founder Wu Tailin currently serves as a distinguished research fellow and assistant professor at Westlake University’s School of Engineering, while also heading the university’s AI and Scientific Simulation Discovery Laboratory. He earned his bachelor’s degree from Peking University’s School of Physics, received his Ph.D. in physics from MIT, and conducted postdoctoral research at Stanford University’s Computer Science Department under Professor Jure Leskovec.
His academic trajectory took a pivotal turn during the second half of his third year as a Ph.D. student. When laboratory funding was abruptly cut off, Wu spent a month deliberating between changing research directions or switching labs, ultimately shifting his research focus from quantum computing to artificial intelligence. He did not abandon his physics background but chose to integrate AI with physics.
“It was essentially a do-or-die move,” he said of that pivot. To accelerate his entry into the AI field, he moved to Los Angeles six months before his Google internship, engaging in deep daily discussions with his advisor. Within a year, his research began gaining momentum, with results surging by 2019.
Wu co-proposed the “theory learning” paradigm and a series of methods centered on the “AI Physicist” concept with his Ph.D. advisor Professor Max Tegmark, producing multiple pioneering results in generative AI-enabled simulation, control, and design of complex physical systems. These algorithms have moved beyond academic papers—at Saudi Aramco, the world’s largest oil company, his algorithms have achieved million-grid reservoir simulation and completed industrial deployment.
Nuclear Fusion as the Highest-Difficulty Validation Scenario
UniForce AI’s choice of controlled nuclear fusion as its first validation scenario is grounded in technical logic. Magnetic confinement fusion involves high-fidelity simulation, state estimation, and closed-loop control of complex plasma systems under extreme conditions, spanning macroscopic magnetohydrodynamics and microscopic turbulence scales, with nearly exacting requirements for model accuracy, real-time performance, and safety.
“The fusion problem is too important, and it’s now on the verge of a breakthrough. I expect demonstration commercialization could arrive by 2040, at which point it could provide humanity, including AI, with nearly limitless and low-cost energy,” Wu said.
The company’s technical approach differs from traditional numerical solvers. It does not aim to replace conventional CAE software. Instead, it uses numerical simulations and experimental data as a high-fidelity training foundation, incorporates physical priors, and employs methods such as neural operators and diffusion models to learn the evolution laws and conditional distributions of physical fields, enabling rapid prediction and coupled generation directly at the inference stage.
This technology has achieved tens to hundreds of times acceleration on complex physical simulation tasks including fluids, reservoirs, and plasmas. Some computational tasks that previously required hours or even days can now be compressed to seconds, with models continuously calibrated using data from real devices and industrial sites.
Currently, UniForce AI is collaborating with Chinese fusion research institutions and leading commercial fusion companies to advance plasma simulation, soft-landing control, and configuration control. Nuclear power multi-physics simulation, 3C thermal-mechanical simulation, and aerospace are all part of the company’s expansion plans.
Shanghai State Capital’s Forward-Looking Deployment
In this funding round, the presence of Shanghai-based state capital institutions is particularly prominent. Shanghai Guotou has consistently maintained a strategic capital positioning, focusing on hard-tech startups in the super-intelligence convergence space.
As early as May 2024, when China’s fusion industry had not yet fully heated up, Shanghai Guotou had already initiated forward-looking research, identifying 30 scientist-led entrepreneurial teams and positioning itself early in this long-cycle sector. A year later, Shanghai officially issued “Several Measures on Accelerating Frontier Technology Innovation and Future Industry Cultivation,” and the Shanghai Future Industry Fund—with a total scale of 10 billion yuan (approximately $1.5 billion), later expanded to 15 billion yuan (approximately $2.2 billion)—was formally established.
The three lead investors—Shanghai Future Industry Fund, Shanghai Sci-Tech Innovation Group, and Fudan Sci-Tech Innovation—carry strong Shanghai state capital and sci-tech innovation credentials. Given that UniForce AI’s first validation scenario is nuclear fusion, and Shanghai happens to be a highland for China’s controlled nuclear fusion industry, the intent behind this investment is not difficult to understand.
Within the broader AI for Science (AI4S) concept, AI4Engineering is an important branch oriented toward engineering physical systems and industrial implementation. China has seen the emergence of many AI engineering companies focused on single vertical domains, but general-purpose infrastructure players serving entire industries remain relatively scarce. UniForce AI’s positioning is precisely to build an AI-native foundation covering simulation, control, diagnostics, and design, enabling cross-industry reuse.
“Many problems across different fields are similar—such as multi-scale, multi-physics simulation and high-performance control. A unified underlying system can serve multiple fields simultaneously, allowing cross-domain learning and even dimensional reduction advantages, with greater robustness,” Wu explained.
This type of “foundation platform” company is rare in the market. Wu’s team’s previous industrial deployment of million-grid reservoir simulation at Saudi Aramco has already demonstrated the feasibility of moving technology from laboratory to engineering sites. And the directions Shanghai is actively pursuing—fusion, high-end manufacturing, aerospace—all require exactly this type of AI-driven simulation and control capability.
The entry of state capital often signals that a direction has moved from “whether anyone is paying attention” to “whether it is worth long-term investment.” For a company less than a year old, securing simultaneous investment from Shanghai Guotou, Shanghai Sci-Tech Innovation, Fudan Sci-Tech Innovation, Loongson Venture Capital, and Chengwei Capital indicates that market consensus on AI4Engineering is forming.
However, Wu also acknowledged that the sector remains in a non-consensus phase. “Many investors are still on the sidelines. Those who believe in us are very committed. Some funds that missed out due to timing may join in the next round.”
From a longer-term perspective, his vision extends beyond solving engineering simulation problems. In Wu’s vision, if fusion energy achieves commercialization, fusion-powered engines could potentially increase rocket speeds from chemical propulsion’s roughly ten-plus kilometers per second to a few percent of light speed—reaching Mars in hours and Pluto in days. “Humanity will gradually become a solar system civilization, bringing about an increase of hundreds of millions of times in the future.”
Once added, BigGo Finance appears first in Google Search Top Stories, so you get the broadest, most up-to-the-minute, and most comprehensive global financial news first.
