Source: Hackaday
Legitimate AI API access is being systematically arbitraged through stolen credentials and resold accounts, creating a shadow market where users buy cheap tokens without knowing they're using compromised infrastructure. This undermines platform economics—OpenAI, Anthropic, and others rely on usage-based pricing to control costs and liability—and exposes ordinary consumers to legal risk. Buyers of discounted access are effectively purchasing through fraud, making them liable parties if detected. The thriving reseller market signals that official pricing for AI inference remains high enough to incentivize this workaround, particularly for price-sensitive builders in emerging markets.