OpenAI is weighing significant cuts to what it charges developers and businesses for access to its AI models, according to a report from the Wall Street Journal. The deliberations center on token pricing — the per-unit cost companies pay every time they send text to or receive text from an AI system — and the potential reductions being considered are described as drastic.
According to Reuters, the move is being driven by an anticipated battle for users with Anthropic, the AI startup behind the Claude family of models. The two companies are increasingly direct rivals for the enterprise and developer market, and pricing has emerged as a key front in that competition.
Benzinga notes that both OpenAI and Anthropic are eyeing what it calls a "trillion-dollar IPO dream," adding a financial dimension to what might otherwise look like a straightforward price war. Capturing and locking in a large base of paying users before a potential public offering would strengthen each company's growth story for investors.
Token prices matter because they function like a metered utility bill for AI. The cheaper tokens get, the easier it becomes for startups and enterprises to build products on top of these models at scale — which in turn drives adoption and, eventually, revenue for the underlying platform.
If OpenAI follows through, it could pressure Anthropic to respond in kind, setting off a broader repricing across the industry. That's good news for developers and businesses building with AI, but it raises questions about when — or whether — either company achieves sustainable profitability. The stakes extend well beyond the two firms: whoever wins this user base could shape which AI platform becomes the default infrastructure of the next decade.