(Yicai) July 19 –(Yicai) July 19 — China's red-hot artificial intelligence ecosystem is drawing intense interest from Silicon Valley venture capital heavyweights across the Pacific.

"I think the idea of the conference is great. The idea is that we should get people from all over the world to come together to talk about AI," Bill Reichert, a veteran Silicon Valley venture capitalist and partner at Pegasus Tech Ventures, told Yicai in an interview at the 2026 World Artificial Intelligence Conference (WAIC), which opened in Shanghai on July 17. 

"Most AI conferences that I've been to so far are either US or Western, and there are very few in terms of the developing world. And the barrier between the US and China limits our ability to have that discussion as well," he added.

"So I think the spirit of the conference is to try to be open and collaborative, and to enable people around the world to hear what different people are thinking. So, I think that's a great aspect of this conference." Cooperation in the AI landscape, he agreed, matters now more than ever.

This marks Reichert's debut attendance at WAIC. "I haven't had a chance to see much of the conference yet. I've seen a few companies, heard a couple of speeches, and I've read a couple of speeches," he noted, adding that he planned to head to the main exhibition floor shortly to review the exhibitors.

Risk Assessment for Early-Stage AI Startups Has Fundamentally Shifted

Reichert has been a venture capital investor for nearly three decades, originally serving as co-founder and managing director of Garage Technology Ventures, a seed and early-stage firm later acquired by Pegasus. Headquartered in Silicon Valley, Pegasus Tech Ventures has invested in more than 300 startups globally—including SpaceX, OpenAI, Anthropic, xAI, Airbnb, SoFi, and Coinbase—and manages over USD 2 billion in assets. As a partner at the firm, Reichert focuses on deep technology investments spanning AI, quantum computing, next-generation chips, materials science, and medical technology.

"AI has become a dramatically more important aspect of our lives, of our world, in just the last couple of years," Reichert said. "While AI has a 70-year history, it's just in these last few years that we're seeing that the impact of AI can be dramatic and not always positive. And so getting more focused attention on the potential impact, the potential risks, as well as the potential benefits of AI, for both the developing world and the developed world, I think, is critically valuable."

Reichert stated that his risk assessment of early-stage AI startups has fundamentally changed since the emergence of foundation models.

"It has become unbelievably simple to create new apps. And so, we're seeing an explosion in the number of entrepreneurs that are developing applications using these large language models to address specific real-world issues—whether it's healthcare, legal services, government administration, logistics, supply chains, you name it," he told Yicai.

"So, we're seeing this explosion of applications, which we refer to as agentic AI applications. And so it makes it really hard for investors to figure out who's going to win because next week there could be another competitor with an equally good or better model," he explained. "And so, it does make it very hard for us."

For venture capitalists, he observed, there is a clear dividing line between software-driven agentic AI and large language models on one side, and physical AI and robotics on the other. "To win in robotics requires a combination of skill sets and technology that is not easily duplicated," he said. Companies that establish leadership in robotics can sustain that leadership and competitive advantage in ways that software-based agentic AI companies will find much harder to match. "And so that's our core challenge as investors in this space," he remarked.

Physical AI Offers Unique Investment Upside

When asked which AI vertical will deliver the most sustainable investment returns over the next five years, Reichert told Yicai that physical AI will have an accelerating impact on society, business, and the wider world.

"We now have a good feel for the impact of large language models and generative AI. And so, I think we understand how that is evolving and the momentum there," he said. "I think with physical AI, we're still in the process of learning where it's going to actually influence our countries, our world, our business, our lives. And so, I think that's an area of particular interest and, therefore, also of opportunity as an investor."

'Build, Validate, and Then Come to Us for Money and Scale'

Asked what advice he would offer non-US AI founders searching for capital, Reichert first gave a tongue-in-cheek answer: his primary advice to any non-US founder who wants to build an AI-based company is simply to move to the United States.

"Now, having said that," he noted that he also serves as chief evangelist for the Startup World Cup—a global competition founded and managed by Pegasus. He described it as the largest startup competition on the planet, featuring regional qualifying rounds held in more than 100 countries and regions annually. The role regularly takes him to innovation ecosystems around the globe; he had just returned from Tbilisi, Georgia, where the winner of the local regional round is developing highly advanced AI technology.

"When it comes to AI knowledge, the world is actually very flat. You don't have to be in Silicon Valley in order to build a very powerful AI model," he said, pointing to open access to technology, open-source AI frameworks, data, and global talent. "Access to capital, however, is different."

"The good news for non-US AI founders and entrepreneurs is you can build and validate an AI business wherever you are in the world. But to scale it, you need resources; you need some connection to a mechanism to get distribution. And generally, it's going to mean you also need money," he concluded. "So build, validate, and then come to us for money and scale."