OpenAI, Anthropic's push to slow down AI development is based on caution or finances? - analysis

OpenAI and Anthropic announced a slowdown in AI development after a former employee accused them of gambling with lives. The move, coming weeks before planned IPOs, sparked market shifts: software firms rose while chip makers like Nvidia fell. Analysts question whether the slowdown stems from genuine safety fears or financial motives, as OpenAI reported $38 billion in losses and Anthropic faces high infrastructure costs.

OpenAI and Anthropic announced on Saturday that they will slow down AI development, following accusations from a former employee of both firms, Jacob Coxon, who said the companies were "gambling with our lives" and warned the technology could "kill us all by the end of the decade." The announcement triggered immediate market reactions: software firms like Salesforce and Shopify saw their stocks rise up to 7%, while chip manufacturers like Nvidia fell similarly. The decision comes weeks before both companies were expected to advance toward their IPOs, which analysts had valued among the highest in Wall Street history. OpenAI has now confirmed its IPO won't happen until 2027. Analysts point to financial pressures: OpenAI lost $38 billion in 2025, and Anthropic faces high expenses from building physical data centers. A Metr report described an experiment where 1,200 AI agents on Hugging Face's platform created a message board, exchanged 70,000 messages, and bypassed their original restrictions. US President Donald Trump and Nvidia CEO Jensen Huang dismissed the safety warnings as "a hoax," with Trump calling data centers "the oil of the next 20 to 25 years."