OpenAI, Anthropic court Wall Street while managing AI safety expectations ahead of IPOs

OpenAI and Anthropic are preparing for potentially record-breaking IPOs, balancing investor demands for growth with government and public concerns over AI safety. OpenAI announced its Astra model will be released broadly but with a 'critical' cybersecurity threshold, limiting its most powerful capabilities to trusted testers. The move signals a strategic effort to preempt regulatory backlash while maintaining commercial momentum.