AI rivalry and tail-risk contagion: Evidence from DeepSeek's creative destruction shock

The disruptive entry of DeepSeek convulsed the artificial intelligence (AI) market. This study integrates creative destruction theory with the CAViaR–TVP-VAR and VAR on VaR frameworks to examine whether DeepSeek's launch induced a persistent shift in 5% and 95% tail-risk connectedness among leading U.S. (Apple, Microsoft, Nvidia, Meta, Alphabet) and Chinese technology firms (Baidu, Alibaba, Tencent, SenseTime, iFlytek). The study used daily data before and after DeepSeek's launch on January 27, 2025, and revealed that while the pre-DeepSeek negative tail-risk pairwise connectedness (PCI) was weak, the post-DeepSeek negative tail-risk PCI was strong. On the contrary, the inter-country PCI for positive tail-risk decreased after the DeepSeek launch. The study found dominant negative tail-risk transmission from Meta and Apple to all Chinese firms, from Tencent to Alphabet and Nvidia, from SenseTime to Microsoft, from iFlytek and Baidu to Alphabet, and from Microsoft to Baidu. Positive tail-risk transmission is found prominently from Microsoft to Apple and Alphabet, and from Meta and Apple to Alibaba. The investigation also found that, in the post-DeepSeek era, Alibaba, SenseTime, Alphabet, and Apple's roles shifted from negative tail-risk transmission to receiving. Also, SenseTime was transitioned to a receiver, while Tencent, Meta, and Nvidia were transitioned to transmitters of positive tail risk in the post-DeepSeek era. The study noted an upsurge in the Total connectedness index at the onset of DeepSeek launch. These outcomes validate the creative destruction theory, implying that DeepSeek has catalyzed a structural realignment in the global tech ecosystem, redefining the pathways through which extreme financial shocks propagate across markets. It underscores the need for cross-border regulatory coordination, enhanced firm-level risk monitoring, and adaptive portfolio strategies in AI-driven financial integration.