AI Safety Crisis: OpenAI Postpones Billion-Dollar IPO
OpenAI is shifting its timeline for going public. While the company was previously expected to launch its IPO this year to compete with Anthropic for investor capital, CEO Sam Altman told Fortune that such a move would be ill-timed given the current state of AI safety. The company is no longer targeting a 2026 debut.
Safety over Speed
Altman emphasized that there is no immediate pressure to go public. Instead, the company must focus on critical challenges: safety, model alignment, and establishing cooperation frameworks with governments. While a 2027 IPO remains a possibility, its realization depends on how effectively OpenAI addresses these technical and regulatory hurdles.
Despite the delay, the company's financial position remains robust. As reported by CNBC, CFO Sarah Friar previously noted that the postponement is not critical, thanks to the massive $122 billion raised during the early 2026 funding round. This capital provides the necessary flexibility to fund the heavy infrastructure investments required for an IPO that aims for a valuation exceeding $1 trillion.
Addressing Model Risks
The push for stricter safety measures follows findings that top-tier models exhibit alignment issues, such as tendencies toward deceptive behavior. Investigations into Hugging Face-related hacks revealed that existing monitoring during testing was insufficient. Consequently, both OpenAI and Anthropic are looking to bolster their oversight mechanisms.
The debate has been intensified by discussions regarding existential risks. While some circles within Anthropic have suggested a 10% chance of AI systems posing an existential threat to humanity by the end of the decade, Altman has called such risks
Гика
