Back
The Rise of DeepSeek: What the Headlines Miss
webCredibility Rating
4/5
High(4)High quality. Established institution or organization with editorial oversight and accountability.
Rating inherited from publication venue: RAND Corporation
RAND commentary from January 2025 offering a policy-focused analysis of DeepSeek's geopolitical significance, relevant to debates about compute governance and AI export control effectiveness.
Metadata
Importance: 55/100opinion piececommentary
Summary
A RAND commentary analyzing the broader implications of DeepSeek's emergence beyond typical media coverage, examining what China's competitive AI development means for US-China technology competition, export controls, and AI governance. The piece argues that the significance of DeepSeek lies not just in its capabilities but in what it reveals about the limits of compute-focused containment strategies.
Key Points
- •DeepSeek demonstrates that AI capability development can advance significantly with fewer high-end chips, challenging the premise of US export control strategies.
- •The media narrative focused on cost and performance benchmarks misses deeper geopolitical and strategic implications for AI competition.
- •Restricting compute access may not be sufficient to prevent rival nations from developing frontier-level AI systems.
- •DeepSeek's rise raises questions about the effectiveness and long-term viability of hardware-based AI governance mechanisms.
- •The development highlights the need for more comprehensive AI governance strategies beyond compute controls alone.
Cited by 1 page
| Page | Type | Quality |
|---|---|---|
| US AI Chip Export Controls | Policy | 73.0 |
Cached Content Preview
HTTP 200Fetched Mar 20, 202617 KB
- [Skip to page content](https://www.rand.org/pubs/commentary/2025/01/the-rise-of-deepseek-what-the-headlines-miss.html#page-content)
# The Rise of DeepSeek: What the Headlines Miss
Commentary
Jan 28, 2025
Photo by Dado Ruvic/Reuters
By [Lennart Heim](https://www.rand.org/pubs/authors/h/heim_lennart.html)
Recent coverage of DeepSeek's AI models has focused heavily on their impressive benchmark performance and efficiency gains. While these achievements deserve recognition and carry policy implications (more below), the story of compute access, export controls, and AI development is more complex than many reports suggest. Here are some key points that deserve more attention:
1. **Real export restrictions on AI chips only started in October 2023, making claims about their ineffectiveness premature.** DeepSeek [trained on Nvidia H800s](https://arxiv.org/abs/2412.19437), chips designed specifically to circumvent the original October 2022 controls. For DeepSeek's workloads, these chips perform similarly to the H100s available in the United States. The now available H20, Nvidia's most recent AI chip which can be exported to China, is [less performant for training](https://epoch.ai/gradient-updates/us-export-controls-china-ai) (though it still offers significant deployment capabilities that should be addressed [\[1\]](https://www.rand.org/pubs/commentary/2025/01/the-rise-of-deepseek-what-the-headlines-miss.html#chip)).
Export controls will affect China's AI ecosystem through reduced deployment capabilities, limited company growth, and constraints on synthetic training and self-play capabilities.
2. **Export controls on hardware operate with a time lag and haven't had time to bite yet.** [\[2\]](https://www.rand.org/pubs/commentary/2025/01/the-rise-of-deepseek-what-the-headlines-miss.html#diffusion) China is still running pre-restriction data centers with tens of thousands of chips, while U.S. companies are constructing data centers with hundreds of thousands. The real test comes when these data centers need upgrading or expansion—a process that will be [easier for U.S. firms](https://x.com/morqon/status/1882794870114525498) but challenging for Chinese companies under U.S. export controls. If next-generation models require 100,000 chips for training, export controls will significantly impact Chinese frontier model development. However, [even without such scaling](https://blog.heim.xyz/crucial-considerations-for-compute-governance/), the controls will affect China's AI ecosystem through reduced deployment capabilities, limited company growth, and constraints on synthetic training and self-play capabilities.
![Graph showing revised October 2023 export controls with cross lines showing years from 2022
... (truncated, 17 KB total)Resource ID:
b193d0cfd04f2b86 | Stable ID: ZDNiMDQxYz