TL;DR
AI providers like OpenAI, Anthropic, and Google are currently losing money on enterprise subscriptions by heavily subsidizing costs. As usage shifts toward agentic AI, costs are expected to rise sharply, posing significant financial risks for companies relying on these services without adequate planning.
Major AI providers are currently offering enterprise subscriptions at a loss, subsidizing costs to foster widespread adoption and integration into business workflows. This strategy, while beneficial for market penetration now, is unsustainable long-term and could lead to significant cost increases for companies that rely heavily on these services.
Leading AI companies such as OpenAI, Anthropic, and Google are operating at a financial loss on enterprise subscriptions, with estimates suggesting they are subsidizing thousands of dollars per user per month. For example, OpenAI’s ChatGPT Plus has maintained a $20 monthly fee for three years, despite models becoming more capable and feature-rich, which substantially increases the cost of service provision.
The economics are particularly strained with the rise of agentic AI, where extended autonomous sessions burn through tokens at rates far exceeding simple chat interactions. GitHub Copilot, for instance, is shifting to usage-based billing due to the unsustainable costs associated with agentic workloads, which can exhaust rate limits within hours.
Why It Matters
This situation poses a serious financial risk for enterprises that have integrated AI deeply into their workflows, often without accounting for the rising costs. As providers move away from subsidized models, companies may face bill shocks that could dwarf current SaaS expenses, potentially threatening profitability and operational stability.
Furthermore, the shift indicates a fundamental change in AI economics, from a subsidized growth strategy to a cost-recovery or profit-driven model, which could alter the availability and pricing of AI tools in the future.
Built Not Bought: How AI Could Break the Subscription Empire and Change the Future of Business
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Background
Over the past two years, AI providers have heavily subsidized enterprise and consumer subscriptions to accelerate adoption. OpenAI, for example, has not increased ChatGPT Plus pricing despite significant improvements in model capabilities. Meanwhile, companies like Google and Meta have offered AI services at subsidized rates or free, funded through other revenue streams like advertising.
Recent developments show a pivot, with companies like GitHub announcing plans to implement usage-based billing for AI tools, acknowledging that the previous flat-rate models are no longer economically viable given the rise of agentic AI workloads. This shift is driven by the exponential increase in token consumption when AI operates autonomously over extended periods.
“Our subscription pricing was something we stumbled into. We’re considering phasing out unlimited plans, similar to unlimited electricity.”
— OpenAI VP of Product Nick Turley
“Starting June 1, 2026, Copilot will transition to usage-based billing because agentic workloads are causing costs to explode.”
— GitHub CEO
Cannabis AI Revolution Playbook: A Toolkit for Smarter Cultivation, Sustainable Resource Management and Consistent Product Quality (Cannabis Business Blueprint)
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
What Remains Unclear
It remains unclear how quickly providers will pass on these costs to enterprises and whether new pricing models will be adopted universally. The full financial impact on companies heavily reliant on AI workflows is still unfolding, and some organizations may be underestimating future expenses.
usage-based AI billing solutions
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
What’s Next
Next steps include providers formalizing new billing structures, likely moving toward usage-based models, and enterprises reassessing their AI strategies and budgets. Monitoring how these changes impact enterprise operations and costs over the coming months is critical.
Sustainable Cloud Development: Optimize cloud workloads for environmental impact in the GenAI era
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Key Questions
How much are AI providers currently losing on enterprise subscriptions?
Estimates suggest that providers like OpenAI and Anthropic are losing tens of dollars per user per month, with some reports indicating losses exceeding $20 or more per user, especially with agentic AI workloads.
Will enterprise AI costs increase significantly?
Yes, as providers shift to usage-based billing for agentic AI, costs are expected to rise sharply, potentially making current subsidized models unsustainable for many companies.
What should companies do to prepare for rising AI costs?
Organizations should audit their AI usage, model future expenses under new billing schemes, and consider optimizing workflows or negotiating new contracts to mitigate potential cost shocks.
Could providers revert to subsidized pricing?
It is unlikely in the long term, as the current subsidies are unsustainable. Providers may attempt to balance profitability with competitive pricing, but the trend appears toward cost recovery through usage-based models.
