China sovereign AI funding surge & talent war
Key Questions
What funding and valuation did DeepSeek recently achieve?
DeepSeek secured $7.4B in funding at a $50B valuation, accompanied by talent lock-in strategies. This reflects broader surges in China sovereign AI funding and efforts to retain top talent amid intensifying competition.
How does GLM-5.2 perform relative to Opus and what cost advantages does it offer?
GLM-5.2 achieves approximately 95% of Opus 4.8 performance at an 80% cost reduction. Combined with Huawei Ascend deployments at 90% lower cost, this underscores Chinese open-weight models delivering 4-10x cost advantages.
What does OpenRouter data reveal about US startup usage of Chinese LLMs?
US startups account for 61% of tokens routed to Chinese LLMs via OpenRouter. This highlights accelerating adoption driven by cost efficiencies and emerging data sourcing pipelines for model improvement.
What is Bindureddy's prediction regarding US and China open source AI leadership?
Bindureddy predicts the US will overtake China in open source AI within six months. This contrarian view challenges the prevailing narrative of Chinese dominance in the space.
Why should AI investors avoid choosing exclusively between Wall Street and Asia?
Regulatory divergence between the US and Asia is presented as a feature rather than a bug, supporting diversified capital allocation. This approach helps navigate decoupling signals like Alibaba's ban on Claude Code over security concerns.
DeepSeek $7.4B at $50B with talent lock-in. GLM-5.2 at ~95% of Opus 4.8 at 80% cost reduction. Manus buyback from Meta for $2B. OpenRouter: US startups 61% tokens to Chinese LLMs. Bloomberg 'China Shock' on Zhipu. Alibaba bans Claude Code over alleged backdoor risks. Bindureddy predicts US will overtake China in open source AI within 6 months. Regulatory divergence is a feature, not a bug.