Chinese Scientists Take Center Stage in Silicon Valley

Robotics startup Dyna Robotics recently announced the completion of a $120 million Series A funding round (approximately 850 million RMB), bringing the company's valuation above $600 million (approximately 4.27 billion RMB).

Notably, industry giants including NVIDIA, Amazon, Samsung, and LG participated in this funding round. The specific investors include robotics-focused investment fund Robostrategy, with CRV and First Round Capital leading the round, and Salesforce Ventures, NVIDIA, Amazon, Samsung, and LG Technology Ventures participating.

Founded in September 2024 and headquartered in Houston, Texas, Dyna Robotics maintains a hardware R&D center in Shanghai. In March of this year, the company completed a seed funding round of over $20 million, led by CRV and First Round Capital, with ZhenFund participating. At that time, the company was valued at approximately $100 million.

This means that in just one year since its founding, the company's valuation is poised to increase fivefold.

A Chinese Team

Dyna Robotics was founded in September 2024, with headquarters in Houston, Texas, and a hardware R&D center in Shanghai, China. The company was co-founded by three partners: Lindon Gao, York Yang, and Jason Ma. The backgrounds of these three founders represent a perfect combination of entrepreneurial experience, technical expertise, and academic strength.

Lindon Gao and York Yang are long-term business partners who previously co-founded and successfully sold smart shopping cart company Caper AI for $350 million. Lindon Gao immigrated to the United States from China with his parents as a child and possesses rich business vision and cross-cultural management experience. York Yang was born in Hangzhou, China, graduated from Zhejiang University's Electronic Engineering program in 2010, and later pursued advanced studies in the US, earning a Master's degree in Computer Science from UCLA.

The third co-founder, Jason Ma, is an academic authority in the robotics field. He earned his PhD from the GRASP Laboratory at the University of Pennsylvania and has worked at NVIDIA AI, Meta AI, and Google DeepMind, focusing on building robotic foundation models and leading the development of multiple breakthrough algorithms. Other core team members come from prestigious institutions like Harvard, MIT, and Berkeley, as well as leading tech companies including Google, NVIDIA, Aurora, and Cruise, forming a top-tier team that combines research expertise with engineering capabilities.

In terms of company development, Dyna Robotics has demonstrated remarkable growth speed. In March 2025, the company completed its seed funding round of over $20 million, led by CRV and First Round Capital with ZhenFund participating, when the company was valued at approximately $100 million. Just six months later, the company's valuation has soared to $600 million, attracting more strategic investors.

Regarding technical capabilities, Dyna Robotics currently employs approximately 30 people and is focused on developing AI models that help robots learn and improve their capabilities in real-world scenarios. At the end of April this year, the company officially launched its first all-weather, efficient, and stable autonomous dexterous manipulation model DYNA-1 (Dynamism v1), which is the world's first dexterous manipulation foundation model that can be deployed in commercial scenarios, demonstrating excellent performance in quality, speed, and stability.

The company's robotics philosophy aligns closely with NVIDIA CEO Jensen Huang's "Physical AI" concept. CEO Lindon Gao states that Dyna doesn't write task instructions for robots, but instead lets them gradually learn through data input obtained from their environment. By integrating robots into the real world, Dyna's models can become faster and smarter. "Our ultimate goal is to unlock physical AGI."

However, unlike some humanoid robot companies, Dyna Robotics' hardware focus is not on humanoid products. Currently, the company's fixed robotic arms are already being applied across multiple industries, including folding napkins for restaurants, organizing towels for fitness centers, and handling clothing for self-service laundromats. These seemingly simple tasks actually require high precision and adaptability, representing an important step for robots operating in the real world. Lindon Gao states that in the long term, Dyna will explore robots that are "closer to human form," but currently focuses more on solving specific problems in actual commercial scenarios.

Dyna Robotics' successful funding also reflects the intense heat in the entire robotics industry. According to PitchBook data, the robotics industry has attracted $12.1 billion in investment in the first half of this year. From the market's perspective, the robotics industry is in the early stages of trend investment, where initial order quantities don't constitute key signals. The core lies in whether it can solve two key bottlenecks for humanoid robots: high hardware costs with complex and undefined structures, and insufficient intelligence in the "brain."

Regarding the use of this funding, the company plans to further optimize its AI models and deploy more robots. With the injection of capital, Dyna Robotics is expanding its world-class research and engineering team, accelerating the development of next-generation foundation models, committed to delivering high-performance general-purpose robots in commercial environments.

From the banks of the Fuchun River to the heights of Silicon Valley, Dyna Robotics' Chinese-background founding team is writing new chapters in the frontier field of embodied intelligence with their unique technical philosophy and business approach.

Their story is not just about technological and commercial success, but also about the fusion of Eastern and Western wisdom and thinking about how to make robots truly valuable in the real world. With the backing of giants like NVIDIA, this startup is becoming an important builder of the bridge connecting robotics research and widespread applications.

Jensen Huang's Frequent Investments

Speaking of which, Jensen Huang's investment activities have been quite frequent in recent years. In 2024 alone, NVIDIA made approximately 45 investments, with its investment portfolio already containing about 40 unicorn or near-unicorn companies.

Huang's investment logic is clear and firm—he doesn't simply chase trends, but focuses on companies that can strengthen NVIDIA's ecosystem, prove the value of its technology platform, and extend AI capabilities from the digital world to the physical world. His heavy investment in embodied intelligence and robotics, especially recent key investments, deeply reflects this strategic intent.

NVIDIA's investment in emerging robotics company Dyna Robotics is a typical example of its layout in embodied intelligence.

This company, co-founded by a Zhejiang University alumnus, saw its valuation soar 5-6 times in just six months, securing $120 million in its latest funding round with a valuation exceeding $600 million. Dyna Robotics' core attraction lies in its development of the world's first commercially deployable dexterous manipulation foundation model DYNA-1 (Dynamism v1), which enables robots to smoothly complete multiple complex end-to-end tasks, such as dual robotic arm collaboration for organizing items, folding towels, assembling sandwiches, and quickly responding to human voice commands.

Its uniqueness lies in not relying on traditional coding instructions, but letting robots gradually learn through environmental data input, which highly aligns with Huang's "Physical AI" concept. What NVIDIA values is precisely Dyna Robotics' potential to combine AI models with physical world operational capabilities and rapidly deploy in diverse scenarios like dining, fitness, and laundry, helping NVIDIA's hardware and simulation platforms occupy core positions in the robotics industrialization process.

Beyond direct investment, NVIDIA also broadly deploys in the embodied intelligence track through ecosystem building.

Huang is making distributed bets on multiple future embodied intelligence giants in the Chinese market, similar to how he bet on OpenAI. At the 2025 World Robot Conference, NVIDIA appeared alongside many Chinese robotics ecosystem partners, including Unitree Technology representing hardware bodies and Galbot representing software brains.

Unitree Technology has deployed NVIDIA's full-stack robotics technology on its new humanoid robot R1, utilizing NVIDIA's Isaac Sim platform to train robot movements, with robots already applied in evening performances and commercial shows. Galbot has self-developed humanoid robot Galbot with built-in NVIDIA Thor chips, capable of quickly sliding to shelves and autonomously completing loading and unloading tasks in transport operations, already deployed in smart pharmacies and smart retail stores with contracts signed for over a hundred pharmacies.

NVIDIA's robotics ecosystem is not a simple procurement relationship, but a "computing power-simulation-data" three-ring interlocked infrastructure that weaves scattered hardware bodies, vertical scenarios, and algorithm teams into a network capable of rapid commercial deployment. Its purpose is to obtain rich scenarios and data through extensive partners, feeding back into the iteration and improvement of its technology platform.

Huang's investment in AI infrastructure is equally relentless. On the domestic US front, NVIDIA announced a $5 billion investment in troubled chip giant Intel.

Although this is not a direct investment in startups, this move deeply reflects Huang's determination to consolidate and expand AI computing infrastructure. Through cooperation with Intel, NVIDIA aims to more closely integrate Intel's x86 CPUs with NVIDIA's AI acceleration GPUs, developing new chips for personal computers and data centers.

Huang pointed out that this cooperation reveals that "the era of accelerated computing and AI computing has arrived," and is expected to open a new field with annual market opportunities of $25 billion to $50 billion for NVIDIA. This not only allows NVIDIA to penetrate deeper into the PC market and pivot toward "edge computing," diversifying business risks, but also strategically counterbalances AMD, which possesses both CPU and GPU capabilities, ensuring NVIDIA's dominant position in the key underlying architecture of AI computing power.

Analyzing Huang's reasons for investing in AI and embodied intelligence tracks, the primary factor is his strategic need to build and consolidate NVIDIA's ecosystem moat. By investing in companies with high technological synergy with its own, NVIDIA ensures that its core technologies including GPUs, CUDA, Isaac Sim simulation platform, and latest Thor chips can continuously receive feedback and drive from cutting-edge application scenarios.

Secondly, Huang keenly perceives that AI is moving from processing information to controlling the physical world, which he calls tapping the potential of the "$100 trillion physical world market." Investing in Dyna Robotics, Wayve, and Chinese robotics companies is precisely to seize the entry point of physical AI and dominate the paradigm shift of the next computing era.

Huang's investment logic is essentially a strategic layout centered on technological ecosystems, crossing digital and physical worlds. He is not simply pursuing financial returns, but through capital ties, deeply binding innovative forces that can validate his computing platform, expand his technological boundaries, and ultimately embed his chips and software into the future intelligent world.

From Silicon Valley robotics startups to British autonomous driving companies, to his emphasis on China's embodied intelligence potential, Huang is using his unique investment philosophy to build an unshakeable imperial foundation for NVIDIA in the AI era, continuously driving the revolutionary process of artificial intelligence from virtual code to physical reality.

Chinese Scientists at Silicon Valley's Tech Center

Speaking of which, the most eye-catching aspect of this company is that its founding team is entirely composed of Chinese members.

In fact, since ChatGPT's stunning debut at the end of 2022, Chinese presence in the global artificial intelligence field has become increasingly strong. Currently, the international tech community has formed a general consensus: research talent from China is becoming a key intellectual force supporting America's top AI institutions.

On July 9th, Musk officially launched xAI's fourth-generation product. However, the focus at the launch event was not him, but two Chinese scientists—Wu Yuhuai and Jimmy Ba. After the event, a group photo of the xAI team went viral on social media—the proportion of Chinese members was close to 80%, particularly striking.

Specifically, among xAI's 12 founding members, Chinese occupy five positions. Besides the aforementioned two, they include Yang Ge from Harvard's mathematics department, Dai Zihang who graduated from Tsinghua's computer science department, and Zhang Guodong from Zhejiang University.

It's not just Musk who favors Chinese scientists. Another Silicon Valley giant—Meta CEO Zuckerberg—is also actively recruiting Chinese technical talent.

He is launching Silicon Valley's most intense talent war in nearly a decade, with Chinese engineers as the primary target.

Among these, the most notable recruitment is the addition of Pang Ruoming, former head of Apple's foundation model team. According to Bloomberg, his contract with Meta is worth up to $200 million. Soon, his deputy Tom Gunter and several other core members also successively joined Meta.

Among Meta's currently disclosed 14-person recruitment list, Chinese-American engineers account for 8 people, most coming from OpenAI and Google DeepMind.

The first batch confirmed to join includes Tsinghua alumnus Zhao Shengjia, Yu Jiahui from USTC's gifted youth program, Zhejiang University graduate Bi Shuchao, and Peking University graduate Ren Hongyu. These individuals are all key R&D forces supporting OpenAI's achievements like GPT-4.1.

It's not just Meta; Google, Microsoft, and other tech giants have also joined the talent grab, listing Chinese scientists as core targets. For example, Google successfully recruited He Kaiming, a visual recognition expert who graduated from Tsinghua's pilot program and already holds tenure at MIT; NVIDIA CEO Huang recruited two Chinese students and promoted one to Vice President and Chief Scientist.

Particularly noteworthy is that while "Chinese scientists continue to influence Silicon Valley," China's domestic artificial intelligence capabilities are also rapidly improving, even achieving catch-up and leadership in some areas.

According to a report jointly released by the United Nations Industrial Development Organization China Investment Promotion Office and Shenzhen Dongbi Data, among the world's top 100 AI scientists, 65 have Chinese backgrounds. Of these, 50 currently work at Chinese institutions, while the other 15 work at US research centers.

Correspondingly, Chinese scholars have also established leading positions in AI core research output.

Stanford University's "2025 AI Index Report" shows that from 2010 to 2023, AI-related patents increased dramatically from 3,833 to 122,511, rising 29.6% in the past year alone. As of 2023, China ranks first in total AI patents, accounting for 69.7% of global granted patents.

The reason Chinese scientists can stand out in the AI era is largely due to the transformation of large model technology paradigms—this shift precisely highlights Chinese engineers' comprehensive advantages in efficiency, robustness, and innovation.

A Chinese engineer who has worked at both Meta and OpenAI shared: "They don't particularly favor Chinese people, but have recognized a fact: Chinese engineers are the best guarantee for achieving the triple goals of efficiency, stability, and innovation."