Chinese AI models gain ground as US firms chase cheaper alternatives

US companies are increasingly turning to Chinese AI models — notably Z.ai's GLM-5.2 — as a cheaper alternative to pricier American systems, Futurism and Business Insider reported. Silicon Valley figures have called GLM-5.2 nearly as capable as leading US models at a fraction of the cost, and the trend is accelerating as the domestic price war makes buyers cost-conscious.

Mechanically, the appeal is straightforward economics: Chinese open and hosted models undercut Western frontier pricing by an order of magnitude, and for many commodity tasks the capability gap is now small enough to be irrelevant. A Business Insider hands-on test of the free GLM-5.2 found impressive answer quality despite painful wait times and some unrealistic outputs (e.g., hotel-pricing gaps) — a mixed but improving picture.

The financial world is taking notice: Goldman Sachs initiated coverage on three preferred Chinese AI models, including Hong Kong-listed Zhipu (Knowledge Atlas) with a HK$1,880 price target implying roughly 15% upside — a signal that institutional investors see durable value in the Chinese AI stack.

Competitively, this is one of the week's defining themes: cheap Chinese models (GLM-5.2, DeepSeek V4 Flash, Qwen) are structurally pressuring OpenAI, Anthropic and Google on price and forcing the US price war. Skeptics note latency, reliability, data-residency and geopolitical/compliance concerns that limit enterprise adoption of Chinese models for sensitive workloads. Watch whether US enterprises move beyond experimentation to production deployment, and how export/regulatory dynamics shape access.