China’s AI blitz creates ‘death zone’ for U.S. rivals

A wave of Chinese artificial intelligence model launches over the past eight weeks is reshaping the competitive landscape between Beijing and Silicon Valley, creating what industry analysts describe as a "death zone" for American AI companies unable to match either the performance or the pricing of their Chinese rivals.

Five Models in Eight Weeks

Alibaba Group Holding ↗ 3.04% became the latest Chinese company to debut a model at the top of global benchmarks with its Qwen3.8-Max, which appeared to match or exceed Anthropic's flagship Fable 5, according to Bloomberg. The release follows Moonshot AI's Kimi K3, which demonstrated performance comparable to the most expensive U.S. models on a far smaller budget, and DeepSeek's V4 Flash, which undercut American pricing by orders of magnitude.japantimes

ByteDance's Seedance 2.5 has taken the lead in AI video generation, while Z.ai's GLM-5.2 became the highest-ranked open-source model globally upon its June release. Together, the five launches represent what Beijing-based analyst Poe Zhao called "a repeatable system for producing models close to the global frontier," rather than the isolated breakthrough that DeepSeek represented in early 2025.japantimes

The Price War Intensifies

The cost disparity is stark. Independent evaluator Artificial Analysis found that executing a complex real-world workload costs $0.03 with DeepSeek V4 Flash compared to $3.15 with Claude Fable 5. On a widely shared benchmark chart, a "DeepSeek death zone" has emerged — any model that charges more for equivalent performance or delivers less capability at the same price faces irrelevance.japantimes

U.S. companies have responded aggressively. OpenAI cut the price of its GPT-5.6 Luna by 80% per million tokens, bringing it into the realm of DeepSeek V4 pricing, while Google Alphabet Inc. ↗ 0.67% introduced more affordable Gemini 3.6 Flash and 3.5 Flash-Lite models, according to Tom's Hardware. OpenAI is now selling intelligence equivalent to its March flagship at roughly one-thirteenth the token price.tomshardware

Strategic Implications

The rapid advances have deepened U.S.-China tech friction. President Trump told reporters last Wednesday that his administration is weighing China's threat against the need for safety controls: "We don't want to restrict them where all of a sudden, we come in second to China".japantimes

Both OpenAI and Anthropic are planning initial public offerings with aspirations for valuations of at least a trillion dollars — targets that hinge on high-margin business models now under direct threat. "If there weren't these Chinese open-source models, OpenAI and Anthropic would be laughing all the way to the bank," said AI pioneer Kai-Fu Lee. "Now there's an alternative, and it's cheaper".japantimes

OpenAI has committed to some $600 billion in compute spending by 2030, yet its most popular models are getting cheaper, not more profitable. The bet across the industry is that dramatically lower prices will drive enough volume to compensate — a wager on Jevons Paradox that remains unproven as bills come due.tomshardware