Anthropic’s Growth Depends on Claude, but Safety Measures May Lag Behind

Anthropic's highly anticipated IPO filing reveals the extent to which the AI company relies on Claude to develop future versions of its models and support its operations, according to Forbes staff writer Rashi Shrivastava.

Anthropic is increasingly using Claude to help develop future AI models, a strategy that could accelerate innovation while creating new safety challenges. [Photo: Anthropic X]

In a report published this month, Shrivastava described the strategy as both a potential driver of rapid growth and a source of significant risk, particularly if safety measures fail to keep pace with increasingly rapid AI development.

Anthropic disclosed that it is using its own AI systems to help research, design and evaluate subsequent versions of its models.

The approach could accelerate development, but the company warned that advances in AI could outpace its ability to establish corresponding safety and security measures.

The company's IPO filing also outlines an unusual succession plan for its founders' voting power.

Each founder may nominate a ranked list of up to 10 potential successors, with the lists subject to approval by five of the other six co-founders.

The names are confidential under the proposed arrangement.

The filing further highlights the scale of Anthropic's planned spending on computing infrastructure. Its commitments for AI infrastructure, cloud services and equipment leases exceed $500 billion, according to reporting on the prospectus.

The commitments involve major technology companies and suppliers, including Amazon, Google, Microsoft and Broadcom, as well as arrangements involving other computing providers.

The agreements cover different periods, with some extending into the 2030s and certain equipment-lease obligations extending further.

Many commitments are difficult to cancel or require payment regardless of actual usage, exposing Anthropic to substantial financial obligations as it expands.

The disclosures illustrate the challenge facing Anthropic as it seeks to sustain rapid growth while managing AI safety risks, dependence on computing suppliers and the costs of building the infrastructure needed for advanced models.