Andy Konwinski is sounding the alarm about America’s place in the global AI race. The Databricks co-founder and longtime researcher believes the United States is at risk of losing its edge to China, a shift he describes as an existential challenge for both the tech ecosystem and democratic values.
Konwinski, who now runs the AI investment firm Laude alongside NEA veteran Pete Sonsini and Antimatter CEO Andrew Krioukov, shared his concerns during a session at the Cerebral Valley AI Summit. He pointed to a growing trend inside top universities that says a lot about where innovation is coming from.
If you ask AI PhD students at institutions such as Stanford or UC Berkeley what breakthroughs have impressed them recently, many will tell you they are reading more groundbreaking ideas coming out of Chinese labs than American ones. It is a dynamic that Konwinski finds deeply troubling.
Alongside Laude’s venture arm, he also leads the Laude Institute, an accelerator that funds researchers working on foundational AI technologies. His mission is rooted in the belief that innovation thrives when ideas flow freely, something he says has historically set the United States apart. But he argues that this flow has slowed dramatically.
Major AI companies including OpenAI, Meta, and Anthropic continue to release impressive research, yet much of it remains locked behind proprietary systems. At the same time, these companies are pulling top academics away from universities by offering multimillion dollar salaries that academic institutions simply cannot match.
Konwinski believes that a strong open ecosystem serves as the real backbone of progress. He points to the birth of modern generative AI as a perfect example of why open access matters. The entire movement was shaped by the introduction of the Transformer architecture, originally published as a freely accessible research paper on arXiv. That one openly shared breakthrough became the basis for everything from ChatGPT to Google’s Gemini models.
He argues that the next major leap in AI will likely come from whoever delivers the next architecture level innovation. If the United States does not encourage more open collaboration, he says the advantage could shift abroad.
Konwinski also highlights China’s strategic approach. Government backed labs such as DeepSeek and Alibaba’s Qwen are encouraged to release their work as open source projects. This allows developers and researchers around the world to build on top of them, which accelerates momentum and increases the likelihood of new discoveries. Reports from outlets like MIT Technology Review and Reuters have repeatedly noted how China’s AI sector benefits from centralized support and aggressive investment in talent.
In his view, the US is moving in the opposite direction. Collaboration between academic researchers, labs, and startups is not as fluid as it once was. He believes that a culture of secrecy is limiting the very idea exchange that helped power America’s early dominance in computing and software.
Konwinski warns that this is not only a national security concern but also a long term business risk for the companies that currently dominate the space. He describes the current trajectory as a system eating its own seed corn. If foundational research dries up, he believes even the biggest US AI labs will eventually suffer.
He stresses that the country needs to protect its reputation as the global center of innovation by strengthening open research, encouraging academic freedom, and ensuring ideas circulate widely. Without this, he believes the United States could lose a position it has held for decades.
Topics such as AI research strategy, geopolitics, investment incentives, and the open source movement continue to shape the global AI race. For founders, investors, and policymakers, the message from Konwinski is clear: the next breakthrough may decide the future of technological leadership and the US must make open innovation a priority.
