How China Wins The AI War

Alice Hahn, director at Greenmantle and co-host of the China Decode podcast, joins Prof G Markets to analyze Chinese AI startup DeepSeek's push for a $50 billion valuation amid the US-China AI competition. DeepSeek's recent V4 model rivals top US large language models on benchmarks and supports a one-million-token context window, comparable to offerings from Gemini, Anthropic, and OpenAI. The discussion unpacks why this valuation seems modest, hardware constraints, and the multi-layered dynamics of the race.

DeepSeek's Valuation in Context

DeepSeek's $50 billion target valuation, backed primarily by China's state-led "Big Fund," is far below US peers like Anthropic or OpenAI, which trade at much higher multiples on secondary markets. Hahn attributes this to underdeveloped Chinese capital markets and lower price-to-earnings ratios for Chinese tech firms compared to US Magnificent Seven stocks. Unlike US fundraisings driven by private venture capital, DeepSeek's round emphasizes state support to build a full-stack AI ecosystem rivaling America. Despite the lower figure, DeepSeek leads in China and gained notice last year with its cost-efficient R1 model.