The United States faces a growing threat to its dominance in artificial intelligence, with China rapidly gaining ground. This isn’t just a business concern; it’s an “existential” risk to democratic innovation, according to Andy Konwinski, co-founder of Databricks and the AI venture firm Laude. The core problem? A shift away from open-source collaboration in favor of proprietary AI development.
The Erosion of Open Collaboration
Konwinski points to a disturbing trend: top AI researchers now report encountering more groundbreaking ideas originating from Chinese companies than American ones. This isn’t due to a lack of talent in the US, but a systemic shift. Major US AI labs—OpenAI, Meta, Anthropic—are locking down innovations instead of sharing them freely. This stifles the organic growth that historically drove progress.
The incentive structure further exacerbates the issue. Multimillion-dollar salaries offered by private labs are poaching academic talent, draining universities of the experts needed to foster open research. The result? A “drying up” of the free exchange of ideas that once defined American AI leadership.
China’s Open-Source Advantage
In contrast, China actively encourages open-source AI development. Labs like DeepSeek and Alibaba’s Qwen release their innovations freely, allowing others to build upon them. This collaborative environment fuels rapid breakthroughs. The Chinese government’s support ensures that these advances aren’t locked behind corporate walls.
The critical lesson? The nation that makes the next “Transformer architectural level” breakthrough—the equivalent of the freely available research paper that birthed generative AI—will seize the advantage. China is positioning itself to be that nation.
The Business Threat to US Labs
This isn’t just a matter of national security. Konwinski warns that US AI labs are “eating their corn seeds” by prioritizing short-term proprietary gains over long-term open collaboration. In five years, these same labs will suffer as the fountain of innovation dries up.
The current trajectory threatens not only US leadership but also the business viability of the very companies that are driving the shift toward secrecy. The long-term cost of locking down AI research far outweighs any perceived short-term benefits.
The Path Forward
To regain its edge, the US must prioritize open-source collaboration. This means incentivizing researchers to share their work, supporting open-source projects, and fostering a culture of transparency. The alternative is a slow but inevitable decline as other nations—particularly China—leap ahead.
The stakes are high. The future of AI—and the global balance of power—depends on whether the US can rediscover its commitment to open innovation
