Anthropic weighs new model release ahead of November listing

Anthropic is reportedly evaluating the launch of a new artificial intelligence model prior to its initial public offering. According to a report published on 18 September by Reuters, which cited three sources familiar with the matter, the company is considering this move to counter the momentum gained by OpenAI. The rival firm recently debuted its GPT-6 Astra model, which has attracted significant interest from business customers. This development has led investors to reassess Anthropic’s standing as a leader in enterprise AI tools. One source indicated that Anthropic is currently assessing the safety of the prospective model as it deliberates on a possible release.

The decision-making process involves balancing the investment required for deploying new models against the need to bolster profitability. As interest rates remain elevated, backers are increasingly focused on the timing of anticipated profits. A separate report by the Financial Times noted that investors are questioning whether Anthropic can sustain its growth trajectory amidst intensifying competition, customer price sensitivity, and ongoing concerns regarding AI safety. Despite these worries, the startup’s annualized revenue reached 65 billion dollars in July. Backers predict this figure will exceed 120 billion dollars by the end of the year, a pace that would make Anthropic the fastest-growing business in history. However, considerable uncertainty remains over the company’s ability to maintain this growth, reflected in the wide range of estimated IPO valuations, which span from 1.5 trillion to 4 trillion dollars.

The startup had previously targeted late October for its listing but is now aiming for November, according to a Wall Street Journal report. This strategic timing follows recent statements by Anthropic CEO Dario Amodei, who has called for the AI industry to scale back the launch of new capabilities due to safety concerns. In a 3,800-word essay published on 12 September, Amodei described a dire scenario involving AI-related attacks. He expressed concern that within six to 12 months, a swarm of bots could potentially take over the entire internet, causing hundreds of billions of dollars in damage if AI becomes more powerful without necessary guardrails.

Experts have offered varying perspectives on these safety concerns. Maya Mikhailov, CEO and co-founder of Savvi AI, noted a gap between the worries of AI labs and the actual usage patterns of enterprise customers. She observed that most business clients are not using state-of-the-art models. This disparity is also evident in revenue data, with Ramp’s AI Index indicating that the top 1 per cent of businesses account for the bulk of model companies’ enterprise revenue. Mikhailov suggested that training these advanced models costs billions of dollars, yet most enterprise clients do not utilise the full capacity of these systems, providing labs with a financial incentive to moderate their development pace. PYMNTS has contacted Anthropic for comment but has not yet received a reply.