US vs China AI race: What the Chinese stock crash really changes
Follow Us
But a $480 billion collapse in the stock prices of Chinese AI and chip stocks marks the end of China’s threat to America. It does not. The algorithms continue to come, the top laboratories remain funded, and what gets shaken by this collapse is the easy financing that fueled the rise of the hardware sector. This alters who can compete, at what speed, and at whose expense. Some things moved in the US-China AI race, and some stayed put.
SurveyAlso read: Smartwatch vs screenless fitness band: How do they track heart rate, sleep and recovery?
The models did not blink
Start with what the crash cannot touch. Moonshot AI released Kimi K3 in July – an open-weight, 2.8 trillion-parameter model – that Axios claims outperformed Anthropic’s Fable 5 and OpenAI’s GPT-5.6 Sol in front-end coding tests held by Arena. Just how far behind is China? Depending on whom you ask, the CAISI from the US government placed DeepSeek V4 Pro at eight months, Chris McGuire from CFR places it at seven months, and Stanford’s 2026 AI Index estimates the performance difference at 2.7 percent. Take your pick; none of them was affected by the STAR50 crash.
What changes is the money
It is those named who get hit the hardest by the crash. Zhipu and MiniMax have lost a total of $150 billion from the highs. As reported by Reuters, Hong Hao of Lotus Asset Management told Reuters that the selling wave is going to prevent many firms from listing. Equity funded by retail investors was supposed to be the way for financing the boom. It’s going to bring frustration, and according to Reuters, it is expected to hamper funding from technology companies.
However, private capital is something else. As mentioned in the 36Kr analysis of the Bloomberg report, DeepSeek has already secured the amount of 800 billion yuan from such firms as CATL and Tencent, while Moonshot secured pre-IPO at $500 billion valuation. Industrial and state capital is not linked to the share price. Think about it. This crash might just strengthen the top few firms.
Compute is still the choke point
However, money was never the only constraint. According to Nvidia, its market share in AI data center chips in China has dwindled to nearly zero, while according to TipRanks, citing IDC, China shipments have declined by 44 percent during H1. There are ways around it. As per CFR analysts, DeepSeek trained its V4 with black-market Blackwell chips, while sources told 36Kr that DeepSeek is building a data center in Inner Mongolia that will house at least 160,000 Huawei Ascend chips.
Also read: China AI Bubble Burst explained: What caused the $480 billion stock crash
It is all about scale. China is estimated to account for around $80 billion out of global compute spending this year, whereas Alphabet, Microsoft, Meta, and Amazon collectively account for more than $700 billion.
The US has a bill coming too
However, do not construe this as a complete victory for the United States government. In mid-October, the Nasdaq had its worst week since June on concerns about the financing of the AI rollout. According to FourWeekMBA, Meta has about $454.5 billion worth of non-cancelable commitments in its latest filing. It is difficult to justify those commitments given free or cheap AI models from China, and Axios even states that Kimi K3 jeopardizes the very foundation of the American AI revolution.
What to watch next
Three things to watch. Shares from both Zhipu and MiniMax will be unlocked in January. Both DeepSeek and Moonshot are readying themselves for listing. The U.S. government is considering imposing restrictions on tech equipment which has already hit optical firms Innolight and Eoptolink. The success of these listings will tell how bad the crash was.
Also read: OLED vs IPS gaming monitors: Which panel do you need for your games
