2026 AI Giants Ramp Up Defenses: Tool Bans, Talent Poaching, and Opinion Wars Take Center Stage
The first half of 2026 has seen a dramatic shift in the global AI industry—from open collaboration to mutual vigilance. Major players are restricting...
In the first half of 2026, the global AI landscape has undergone a notable transformation. Once characterized by open collaboration, leading firms are now taking defensive measures against each other. As early as March, Google limited Meta's access to its Gemini model due to capacity constraints, declining to fulfill all of Meta's requests for additional computing resources related to Gemini.
Tool restrictions have become a key defensive strategy. On April 22, the Los Angeles Times reported that most Google employees were banned from using competitor tools like Claude Code and Codex for security reasons, with exceptions only granted upon application. However, some DeepMind teams—including those working on Gemini, internal applications, and open-source models—retained access to Claude Code. Microsoft followed suit: The Verge revealed on May 14 that the company canceled most internal licenses for Claude Code, directing developers to switch to its own GitHub Copilot CLI. On June 10, Microsoft further restricted employee use of Claude Fable5, citing Anthropic's data retention requirements.
The talent war has also intensified. Domestic Chinese tech giants like Baidu, Tencent, ByteDance, and Alibaba have rolled out "uncapped salary" policies in recent months to lure top AI talent. According to Red Star Capital Bureau's July 2 report, Sun Tianxiang—core developer of the open-source ChatGPT-like model MOSS—officially joined Baidu on July 1 as head of the Basic Model R&D Department and a member of Baidu's Model Committee. Prior to this, in December 2025, Tencent had appointed Yao Shunyu as its Chief AI Scientist to oversee the development of its Hunyuan large model.
Competition on the application front has turned fierce as well. On June 30, Liu Xing, PR director of Doubao, stated that nearly 100 Xiaohongshu accounts with highly concentrated IP addresses posed as high school students to publish false content smearing Doubao. This incident highlighted the escalating public opinion battles among major players in the AI application space.
Sources
- Los Angeles Times (April 22, 2026)
- The Verge (May 14, 2026; June 10, 2026)
- Red Star Capital Bureau (July 2, 2026)
- JRJ Finance (July 3, 2026)