Realtime AI News
Dong-A Ilbo: Chinese value models outgrow three US firms combined in usage share
Chinese AI models now hold a combined usage share larger than three US companies put together, as low prices drive adoption, Dong-A Ilbo reported on October 18, calling it a second "DeepSeek shock". The report frames the shift as a pricing story but does not disclose its methodology, so the share figure is best treated as directional.
Chinese AI models are winning share on price, with their combined usage share now exceeding that of three US companies put together, according to a Dong-A Ilbo report published on October 18. The Korean outlet described the shift as a second "DeepSeek shock".
The story is about pricing. Chinese vendors have pushed inference prices very low, trading peak capability for volume, while top US labs keep spending toward the capability frontier. Where the two strategies meet in the same market, the share curve starts to move.
The "DeepSeek shock" label refers to the moment a low-cost model first forced the market to reprice the trade-off between training spend and inference output. Dong-A's use of "again" signals that it sees the current situation not as a single release but as a structural drift built up by sustained price competition.
It is worth being precise about what the report measures. Market share in usage is not revenue or profit. Whether cheap tokens convert into a business depends on retention, enterprise contracts and paid services, none of which appear in a call-volume comparison.
For US vendors the pressure is not only on the price tag. Frontier model pricing bundles reliability, compliance tooling and enterprise support, and many customers pay a premium for those. But when a good-enough model performs comparably on a large share of everyday tasks, each line item of that premium gets re-examined.
The report does not disclose its methodology, data source or time window, so the share figure is better read as a directional signal than a measurement. Confirming the trend would require third-party usage data, enterprise procurement cycles and the compliance progress of Chinese vendors in overseas markets.
The thing to watch is the slope of two curves: usage share for low-cost models, and revenue share for frontier ones. If the first keeps climbing while the second holds, the market is segmenting. If both bend toward Chinese vendors, that is a change in the competitive order rather than a price skirmish.
Why it matters
Price advantage is shifting the contest from capability demos to unit economics, forcing US vendors to justify their premium item by item while Chinese vendors must show that cheap usage converts into durable revenue.
Nearby Updates
All10/19, 08:00
TorchNeuron turns AWS Trainium into a native PyTorch device for Hugging Face models
Hugging Face detailed TorchNeuron, a PyTorch plugin that exposes AWS Trainium as a native device with eager execution, torch.compile and native distributed APIs. Transformers, Diffusers, Accelerate, PEFT and TRL can now run on Trainium directly, with out-of-the-box support for 89% of the top 100 text-generation architectures, up from 46%.
10/19, 08:49
OpenWorld to acquire AI agent startup Blocksight to power $1B tokenization push
OpenWorld has announced plans to acquire Blocksight, an AI agent startup, in a deal it frames as powering a $1 billion tokenization push. Dealroom reported the move, which names both companies and the strategic intent but leaves the financial terms undisclosed.
10/19, 10:55
当Coding Agent开始动态并行，软件开发变得更好了吗？ yeeyi yeeyi
当Coding Agent开始动态并行，软件开发变得更好了吗？ yeeyi yeeyi. 当Coding Agent开始动态并行，软件开发变得更好了吗？ yeeyi yeeyi
10/19, 04:37
Anthropic commits 5% of its shares to a new public benefit fund
Anthropic said on October 18 that it will place 5% of its shares in a public benefit fund structured as a wholly owned entity inside the company. The commitment includes an option to add 2% more for every 50% rise in Anthropic's value, while the fund's governance is still to be detailed in the coming months.