AI Researchers Warn Companies Are Racing Ahead of Safety Testing

AI researchers said companies developing self-improving systems are moving faster than their safety testing can keep pace with, according to reporting cited this week. The warning comes as the AI investment cycle keeps accelerating: chip stocks have risen on Anthropic's spending outlook, and AMD just acquired World Labs to expand its AI capability. The tension is showing up in market commentary too. One analysis flagged that Anthropic's spending plans carry what was described as '$2 trillion IPO warning' language, while a separate piece on Microsoft argued the company remains a strong business but that 'AI raises serious questions' about how durable current spending levels are. Cerebras, meanwhile, was highlighted for what one analyst called explosive growth, illustrating how divided views on AI's near-term payoff have become even among bullish observers. What is unresolved is whether the safety concerns researchers are raising will translate into slower deployment or regulatory scrutiny that could crimp the spending growth investors are currently rewarding. No specific policy action or company response has been announced yet.