In a nutshell: Rising inference costs threaten the profitability of AI business models because unit economics turn negative at broad scale.
The massive investments in AI development and data centers are not amortizing economically at the expected pace. Companies operating AI services are facing financial strain due to rising token costs.
The cost structure of large AI models is becoming the central challenge for business models that rely on a broad user base. While the development of generative models already costs billions, the economically critical costs emerge only in the productive phase: every query to an LLM incurs costs for computing power and token processing that grow exponentially with massive usage.
For CEOs, the central question of unit economics arises: can revenue per user sustainably exceed inference costs per query? Many funding rounds calculated on assumptions about user conversion rates and average queries that do not hold up under real operational pressure. The equation “many users × small margin = big business” does not work if the margin threshold is undercut.
This leaves companies banking on scaling in a dilemma: either they raise prices for end users or AI APIs, but risk competitiveness and customer churn, or they accept losses and hope for technological breakthroughs that lower costs. The optimistic phase of AI financing, in which growth trumped all else, is showing first cracks.
Source: borncity.com · Published June 6, 2026
Lumi AI News — AI-assisted curation in accordance with Article 50 EU AI Act. Paraphrase and classification by Lumi News Pipeline v1.6.5.