China's AI Is 20x Cheaper - And Catching Up - Prof G Markets Recap

Podcast: Prof G Markets

Published: 2026-02-19

Duration: 31 min

Guests: Alice Hahn, Robin Brooks

Summary

The episode explores the rapid advancements and competitive pricing of Chinese AI models compared to their US counterparts, highlighting a potential shift in the global AI landscape.

What Happened

The episode opens with Ed Elson noting the release of the Federal Reserve minutes, which revealed concerns about inflation and potential interest rate hikes, impacting Treasury yields and Bitcoin prices. The discussion shifts to China's recent AI model releases during the Lunar New Year, with companies like Alibaba and ByteDance unveiling powerful new tools that reportedly outperform US models in key benchmarks.

Alice Hahn, co-host of the China Decode podcast, joins to discuss these launches, emphasizing that Chinese AI models are not only improving in speed and performance but are also 10 to 20 times cheaper than US models. She notes that these models are often downloaded and fine-tuned locally, offering improved privacy and cost-effectiveness.

Hahn argues that the comparisons between Chinese and American AI models can be misleading, as each is used for different purposes. While Chinese models are cheaper and faster, American models focus on reasoning and precision, aiming for artificial general intelligence (AGI).

The conversation also touches on the competitive AI landscape in China, with numerous companies releasing models that excel in various areas, such as video generation and multi-language capabilities. Hahn highlights the potential for these models to gain popularity outside China due to their performance and cost advantages.

The episode also covers Sweden's consideration of adopting the Euro amid global uncertainty and trade tensions, with Robin Brooks from the Brookings Institution providing insights on the implications for Sweden's economy and the broader trend of currency shifts.

Finally, the episode delves into the tensions between Anthropic and the Pentagon, focusing on Anthropic's refusal to allow their AI technology to be used for autonomous lethal weapons and mass surveillance, leading to a strained relationship with the Defense Department.

Key Insights

• Chinese AI models are 10 to 20 times cheaper than their American counterparts and are often downloaded and fine-tuned locally, enhancing privacy and cost-effectiveness.
• While Chinese AI models excel in speed and cost, American models prioritize reasoning and precision, focusing on the development of artificial general intelligence (AGI).
• The competitive AI landscape in China is marked by numerous companies releasing models with advanced capabilities in video generation and multi-language processing, potentially increasing their global appeal.
• Anthropic's refusal to allow its AI technology to be used for autonomous lethal weapons and mass surveillance has led to tensions with the Pentagon, affecting its relationship with the Defense Department.