Chinese companies and investors are intensifying efforts to catch up with—and in some cases surpass—leading U.S. artificial intelligence models. The activity spans in-house development, regulatory-driven changes in ownership, and intensifying domestic competition.
Key corporate developments
-
ByteDance is reportedly developing an AI model intended to rival Anthropic’s most advanced Mythos system, aiming to create powerful language and multimodal capabilities in-house.
-
DeepSeek is promoting its work on a challenger to Claude Code, focusing on models optimized for code generation and developer-oriented tasks.
Transactions and regulatory intervention
Responding to regulatory pressure aimed at tightening oversight of the domestic tech sector, the ownership of Manus AI shifted this week: some of Manus AI’s former investors in China repurchased the startup from Meta after Chinese regulators ordered the original acquisition to be unwound.
This move is part of a broader pattern in which Beijing attempts to erect a firewall around domestically developed technologies and impose stricter scrutiny on foreign acquisitions and investments.
Market risks and policy challenges
Chinese policymakers are also confronting persistent volatility in high-value tech stocks, a dynamic that complicates efforts to maintain a steady, upward trend in the markets. At the same time, an intensifying price war among domestic AI firms risks eroding corporate profits—beneficial to consumers in the short term, but potentially damaging to long-term R&D investment and industry sustainability.
Why it matters
A stronger Chinese AI ecosystem could reshape global competition if local players successfully match or exceed the capabilities of leading U.S. models. Yet regulatory interventions, market instability, and aggressive price competition are key constraints that will influence whether that advancement is sustainable and profitable.
(Reporting based on Brendan Ruberry)
