# Trump on US vs China race to AGI | Lex Fridman Podcast

Source: https://www.youtube.com/watch?v=0ByJ0zyE2LM
Recap page: https://rapidrecap.app/video/0ByJ0zyE2LM
Generated: 2026-02-07T22:01:34.654+00:00

---
## Quick Overview

The discussion centers on the geopolitical race in AI development, specifically highlighting the need for the US to invest heavily in building leading, genuinely open-weight AI models through initiatives like The ATOM Project to compete with China's rapidly advancing open-source ecosystem and avoid military/security risks associated with relying solely on closed, proprietary models.

**Key Points:**
- The speaker argues that one of the biggest stories of 2020 was the feeling of the US falling behind China in the open-weight AI model race, consuming significant energy trying to influence US policy.
- The ATOM Project is a US-based initiative focused on building and hosting high-quality, genuinely open-weight AI models to compete with China's advancements in open-source AI.
- The ATOM Project received the largest NSF grant ever awarded for AI2, totaling $100 million over four years, aimed at fostering US leadership in open models.
- The speaker suggests that open models will be cheaper to run due to optimization, making them a more viable option than proprietary, closed models in the long run.
- There is a concern that if China dominates the open-source model space, it could lead to national security risks and an authoritarian AI system dominating globally.
- The Biden and Trump administrations have both shown support for promoting open-source and open-weight AI through policy actions, such as the White House AI Action Plan.

![Screenshot at 00:18: The guest reacts to the idea that US policy efforts in 2020 were focused on catching up to Chinese open-weight models, leading to the creation of initiatives like The ATOM Project.](https://ss.rapidrecap.app/screens/0ByJ0zyE2LM/00-00-18.jpg)

**Context:** This segment features a discussion between Lex Fridman and a guest (likely Nathan Lambert, founder of The ATOM Project, given the on-screen text reference) focusing on the geopolitical competition between the US and China in the field of Artificial Intelligence, particularly concerning the development and release of open-source large language models (LLMs). The context is framed around the strategic importance of open-weight models for future innovation, safety, and national security.

## Detailed Analysis

The conversation emphasizes the urgency for the United States to compete effectively with China in the AI race by championing open-weight models, citing the growing momentum of Chinese open models in usage and downloads. The speaker details the creation of The ATOM Project, an initiative explicitly aimed at building and hosting high-quality, genuinely open-weight AI models to counter China's lead and secure US leadership in the ecosystem. A significant point is the $100 million NSF grant awarded to AI2 for this purpose, signaling serious government backing. The speaker argues that open models are not only crucial for fostering innovation and startups but are also more cost-effective to run in the long term compared to closed systems. There is an underlying fear that if the US fails to develop its own leading open models, the world might default to authoritarian AI systems, making the development of US-centric open models a matter of geopolitical and military strategy. The discussion touches upon the fact that both the Biden and Trump administrations have shown support for open-source AI policies, suggesting a bipartisan consensus on the strategic importance of this area.

### The ATOM Project & US Strategy

- The ATOM Project is a US-based initiative to build and host genuinely open-weight AI models
- It explicitly aims at competing with and catching up to China's rapidly advancing open-source AI ecosystem
- The project's core goal is to establish a US lab dedicated to training models at the 10,000+ NVIDIA H100 GPU scale.

### Geopolitical Stakes

- China's open models are set to overtake the US in usage, and the 'flip' is happening now
- The speaker believes that open models becoming global standards is vital for geopolitical value and US economic standing
- Allowing China to dominate open-source AI poses national security risks.

### Funding and Effort

- The ATOM Project received a $100 million NSF grant over four years, the largest ever awarded for AI2
- The speaker spent significant energy in 2020 trying to influence US policy regarding open models.

### Open vs. Closed Models

- Open models are seen as the foundation of AI research and the most important tool for securing leadership
- Closed models risk concentrating power and knowledge within a few companies, potentially leading to an authoritarian outcome.

![Screenshot at 00:18: The guest discusses the narrative gap between US and Chinese open-source AI efforts, which drove the creation of The ATOM Project.](https://ss.rapidrecap.app/screens/0ByJ0zyE2LM/00-00-18.jpg)
![Screenshot at 00:42: A screenshot of The ATOM Project website detailing its mission: 'Reinvigorating AI research in the U.S. by building leading, open models at home.'](https://ss.rapidrecap.app/screens/0ByJ0zyE2LM/00-00-42.jpg)
![Screenshot at 01:02: The guest explains the two core propositions: open models as an engine for AI research and the necessity for the US to own the best models.](https://ss.rapidrecap.app/screens/0ByJ0zyE2LM/00-01-02.jpg)
![Screenshot at 02:09: The guest gestures while explaining that the story of Chinese AI degrading the environment is the way to make the ATOM Project's message go viral.](https://ss.rapidrecap.app/screens/0ByJ0zyE2LM/00-02-09.jpg)
![Screenshot at 05:25: The co-host discusses the importance of talent identification and education for the future of AI development, suggesting open source is key.](https://ss.rapidrecap.app/screens/0ByJ0zyE2LM/00-05-25.jpg)