A Korean Claude user reports receiving two erroneous, massive invoices from Anthropic totaling $16,627,739.70, despite having zero API usage and no payment method on file. The initial invoice for $1,669,875.90 was followed 24 hours later by a second demand for $16,627,739.70. While the user initially suspected a sophisticated scam, the notifications originated from Anthropic’s official domain and utilized the company’s legitimate Stripe invoicing system.

Although the user had no active credit card registered with the platform, the victim’s bank recorded and declined two attempted overseas transactions originating from the US-based AI company on July 8. No actual funds were successfully transferred as a result of these billing attempts.

As of July 11, 2026, Anthropic has not provided an official explanation or clarification regarding whether the incident resulted from a security breach, a technical billing error, or an AI hallucination (phenomenon where an AI produces factually incorrect or nonsensical information). The incident has sparked widespread debate in the tech community regarding the reliability of automated billing systems in AI services.

A Korean Claude user reports receiving two erroneous, massive invoices from Anthropic totaling $16,627,739.70, despite having zero API usage and no payment method on file. The initial invoice for $1,669,875.90 was followed 24 hours later by a second demand for $16,627,739.70. While the user initially suspected a sophisticated scam, the notifications originated from Anthropic’s official domain and utilized the company’s legitimate Stripe invoicing system.

Although the user had no active credit card registered with the platform, the victim’s bank recorded and declined two attempted overseas transactions originating from the US-based AI company on July 8. No actual funds were successfully transferred as a result of these billing attempts.

As of July 11, 2026, Anthropic has not provided an official explanation or clarification regarding whether the incident resulted from a security breach, a technical billing error, or an AI hallucination (phenomenon where an AI produces factually incorrect or nonsensical information). The incident has sparked widespread debate in the tech community regarding the reliability of automated billing systems in AI services.