OpenAI and Anthropic Feel Pressure as AI Spending Shifts to ROI
Corporate AI budgets are tightening as businesses demand returns, posing a growth challenge for leading AI labs.
A quiet but consequential shift is underway in how companies approach artificial intelligence spending. After an era characterized by expansive, exploratory AI usage — sometimes called "tokenmaxxing," in which organizations threw compute at every conceivable problem — enterprises are now demanding demonstrable returns on their investments. That pivot toward efficiency over experimentation carries real consequences for the revenue trajectories of OpenAI and Anthropic, the two dominant players in large-language-model services.
The dynamic reflects a broader maturation in enterprise technology adoption cycles. Early enthusiasm for generative AI prompted procurement teams to approve sweeping platform contracts, often with little accountability for outcomes. Now, as those contracts come up for renewal, finance departments are scrutinizing utilization rates and business impact. The result is a more disciplined buyer, one who is optimizing token consumption rather than maximizing it — a behavioral change that could compress per-customer revenue even as the total number of enterprise adopters continues to grow.
Read more China's Zhipu AI Narrows Gap With US Leaders on Cost Efficiency →
For OpenAI and Anthropic, the timing is significant. Both companies have been scaling infrastructure and hiring aggressively in anticipation of sustained revenue growth. A deceleration in consumption growth, driven not by disillusionment with AI but by a rational push for efficiency, introduces a new variable into financial projections that lean heavily on expanding usage. The distinction matters: customers are not abandoning these platforms, but they are using them more deliberately, which is a fundamentally different growth environment than the one both labs may have modeled.
The longer-term question is whether the shift toward efficiency ultimately strengthens or weakens the AI industry's commercial foundations. A case can be made that disciplined adoption produces stickier, more defensible enterprise relationships — clients who have integrated AI into measurable workflows are harder to displace than those who experimented loosely. But in the near term, the transition from volume-driven to value-driven consumption is likely to create a growth headwind that neither company can simply engineer its way out of.
Continue reading at US Top News and Analysis.
