China’s AI race accelerates, putting pressure on US companies
A new report points to 16 Chinese AI models released in one month, the growth of cheaper open-weight systems, and possible consequences for US pricing power and lofty valuations.
China’s AI competition is entering a more intense phase as developers release new models at a rapid pace. According to the report, Chinese developers introduced 16 AI models in a single month, adding pressure to competition with OpenAI, Anthropic and Google.
Cheaper, open-weight alternatives
Companies including DeepSeek and Xiaomi are developing systems described as increasingly capable, less expensive and, in some cases, open-weight. Wider availability of such models could improve access for developers while weakening the pricing power of major US providers.
The contest is not purely technical. Anthropic CEO Dario Amodei has warned about the risks of rushing the development of advanced AI. Some Chinese developers reject calls to slow down and question whether powerful models would remain accessible and affordable if their availability were controlled primarily by US companies.
Large valuations bring large risks
On the financial side, the report says Moonshot AI has reached a reported 0 billion valuation and links DeepSeek to a possible 2 billion funding round. These figures are reported claims and should not be treated as confirmation of completed transactions.
Lower-cost Chinese models could compress the revenue and margins of US rivals at a time when billions are being invested in infrastructure and model development. If revenue and profits do not meet investor expectations, today’s high valuations could face a correction.
The competitive landscape therefore extends beyond model quality to cost, access, release strategy and economic sustainability. Model counts alone do not establish technical leadership, however; independent evidence is still needed to compare capability, safety and real-world adoption.
