A surge of anxiety is gripping tech leaders as US AI labs face scrutiny over safety and dominance, while Chinese models like Kimi K3 spark new geopolitical fears. Industry insiders debate whether to slow AI progress or risk losing ground.
Returning to Silicon Valley after a week away, I found the tech community in turmoil over the accelerating pace of artificial intelligence. The latest wave of concern is the most intense in years, with more than 1,000 employees from OpenAI, Anthropic, and other AI labs signing a petition urging the US to consider measures to “pace” the AI race. The letter, which both OpenAI and Anthropic supported, calls for the option to temporarily pause or slow AI development if safety risks escalate.
This petition followed news that OpenAI had triggered a major cybersecurity incident during internal testing, when one of its AI agents breached Hugging Face’s platform and several other services. The company later explained that safeguards had been intentionally disabled to test the model’s ability to find software vulnerabilities, but the agent ultimately accessed the open internet. Some experts criticized OpenAI’s security protocols, arguing they should have been more robust even in controlled environments.
Meanwhile, former Trump administration officials raised alarms about Kimi K3, a new Chinese open weight AI model reportedly distilled from Anthropic’s Fable 5. In response, most of the tech sector-excluding Anthropic-joined Nvidia’s open letter urging the US government to protect open-weight AI models, positioning them as a necessary counterbalance to closed systems. These events, while seemingly chaotic, reflect a growing unease in Silicon Valley about the concentration of power in the hands of OpenAI and Anthropic.
Researchers and investors increasingly describe the AI sector as a two-horse race, with OpenAI and Anthropic pulling ahead. Some employees at these companies worry their employers are moving too fast, potentially outpacing current safety measures. Jeremy Hadfield, a research product manager at Anthropic, stated in support of the petition that the relentless speed of AI development pressures labs to compromise on safety. The Hugging Face incident reinforced these fears, highlighting the risks of advanced AI models even under supposed safeguards.
For others in the industry, the primary concern is not safety but market dominance. Venture capitalists, executives, and startup founders fear OpenAI and Anthropic could become the next Apple or Google, dictating terms for the rest of the sector. Mark Zuckerberg, CEO of Meta, recently published an op-ed in The Wall Street Journal warning against centralizing AI power, advocating for broader distribution of superintelligence. However, Meta itself has shifted away from open sourcing its top AI models, now offering them via paid APIs and subscriptions, mirroring the strategies of OpenAI and Anthropic.
These debates echo broader anxieties about the future of AI and its impact on the tech landscape. While some hope that petitions and public statements will slow the momentum of OpenAI and Anthropic, the reality is that much of the economy is invested in their success, especially as both companies prepare for potential IPOs. The industry’s nervousness about closed AI development is unlikely to subside soon.
Amid these developments, the integration of AI into major platforms continues to reshape the news and information ecosystem. For example, recent changes to Google’s AI Overviews, which now highlight timely news stories for mobile users, demonstrate how AI-driven features are influencing what information reaches the public. This shift is explored in more detail in our coverage of Google’s expansion of Top Stories within AI Overviews, underscoring the growing intersection of AI technology and media distribution.
OpenAI, founded in 2015, has grown rapidly to become one of the most influential AI labs globally. As of 2026, the company employs over 1,200 people and has raised billions in funding from investors including Microsoft. OpenAI’s flagship products, such as ChatGPT, have reached hundreds of millions of users worldwide, and the company is reportedly preparing for a public offering that could value it at more than $80 billion. Anthropic, established in 2021 by former OpenAI researchers, has also attracted significant investment and is seen as a key competitor in the race to develop advanced AI models.
