The US is losing its lead in AI

July 20, 2026

AI Summary

5 min read

In early 2025, DeepSeek, an obscure Chinese AI lab run inside a hedge fund, released a model that was vastly cheaper and more efficient than anything from Silicon Valley — and almost as good. Six months later, Moonshot and Alibaba each claimed frontier models competitive with anything from Google, Anthropic, or OpenAI. The old truism that China is six months behind the US in AI “is basically the best case scenario,” says Verge senior AI reporter Hayden Field. “It may be that they're already caught up.”

The distillation shortcut

The central mechanism driving China’s catch-up is a technique called distillation. Distillation lets a smaller, newer AI model learn from a larger, better-trained one by generating millions of exchanges with it — essentially asking the frontier model questions, collecting the answers, and using them to train a rival. Anthropic recently told regulators it believes three Chinese companies — DeepSeek, Moonshot AI, and Minimax — collectively generated roughly 16 million exchanges with Claude from tens of thousands of fraudulently created accounts.

Continue reading the full summary in the app — free to try.

Read Full Summary →

Free • No credit card required

Listen to Audio Summary Open in App

Never miss an episode of The Vergecast

Get every new episode summarized in your inbox — free, ~5 minutes to read.

No spam. Unsubscribe anytime.

What you'll learn

1 (00:02) **Intro: The US-China AI Race** - David Pierce frames the episode: Silicon Valley has led AI for years, but China is rapidly catching up, with recent models from Moonshot and Alibaba challenging frontier labs.
2 (00:22) **Segment: The "Six Months Behind" Myth** - Hayden Field explains the outdated consensus that China is six months behind, noting it may now be less or even zero for everyday models.
3 (03:50) **Guest Introductions: Hayden Field & Lauren Feiner** - Pierce brings in The Verge's senior AI and policy reporters to contextualize the geopolitical and technical dimensions.
4 (04:39) **Is China Already Caught Up?** - Field says the six-month gap is a best-case scenario; Chinese models are now equal or nearly equal, especially in consumer AI.
5 (05:28) **The Politics of AI Supremacy** - Lauren Feiner notes the US-China AI race is a perennial Washington topic, though the approach (e.g., export controls) has shifted under different administrations.
6 (06:57) **Main Mechanism: Distillation Explained** - Feiner defines distillation as smaller models learning from larger ones (e.g., via millions of exchanges), a shortcut for Chinese firms like DeepSeek, Moonshot, and Minimax to catch up quickly.
7 (08:54) **Distillation as a "Get Rich Quick" Scheme** - Pierce summarizes distillation as reverse-engineering a model by asking it questions; Field notes Anthropic and OpenAI have accused Chinese companies of this practice.

+ Full timestamped outline available in the app

Show Notes

Since the beginning of the AI revolution, a few companies have claimed to be building the best and most powerful AI models. All of them were American. More recently, a few Chinese companies seem to have caught up, shipping models that are much cheaper and maybe also just as good. The Verge's Hayden Field and Lauren Feiner join the show to explain how China made so much progress, the threat that progress represents to US companies and the US government, and what might happen next.