OpenAI reported $5.7 billion in revenue for the first quarter of 2026, outpacing rival Anthropic by nearly $1 billion. However, this financial lead may not be as secure as it seems. Beneath the surface, a growing crisis surrounding AI token pricing is straining the budgets of the enterprises that both companies depend on. While OpenAI's revenue looks strong, the underlying economics are becoming increasingly complex and concerning.
Revenue Disparities and Growth Trajectories
Despite OpenAI's success in Q1, the revenue gap between the two firms narrows when viewed over a longer term. Anthropic's annualized revenue is projected to approach $45 billion, significantly exceeding OpenAI's anticipated $25 billion for the same period. Reports suggest that Anthropic is on track to more than double its first-quarter revenue of $4.8 billion to an estimated $10.9 billion in Q2, while OpenAI has yet to reveal its projections for the next quarter. This indicates that while OpenAI may lead in a single quarter, Anthropic’s growth trajectory is more compelling, potentially shifting investor sentiment as both companies prepare for their IPOs.
In terms of valuation, Anthropic is ahead, currently aiming to raise between $30 billion and $50 billion, which would elevate its valuation to as much as $950 billion. In contrast, OpenAI's valuation is around $850 billion. This difference is significant for retail investors, as it may heavily influence the IPO pricing for both firms.
The Budget Strain on Enterprises
Rising token costs are increasingly visible within companies that are major clients of AI services. Uber's CTO, Praveen Neppalli Naga, disclosed that Uber exhausted its entire 2026 AI budget in just four months, largely due to the adoption of Claude Code. This rapid uptake led to monthly API costs per engineer ranging from $500 to $2,000, prompting Uber to rethink its budgeting strategies. Such instances are not unique; Microsoft is also cutting back its use of Claude Code in its Experiences and Devices division, citing financial pressures as a key reason for consolidating toward GitHub Copilot.
Similarly, GitHub announced a change in its billing model for its Copilot AI coding assistant, shifting from flat-rate subscriptions to a usage-based system effective June 1, 2026. Under this new model, developers could see their monthly costs jump from approximately €67 to nearly €966. This unpredictability in pricing comes at a time when enterprise budgets are already under considerable strain.
Understanding the Token Pricing Crisis
The rising costs associated with token usage are tied to the economics of AI infrastructure. Tokens represent the units of computation required for AI models to process inputs and deliver outputs, and every interaction consumes them. According to Anthropic, the average cost per developer using Claude Code is around $6 daily, but this average masks the reality for many enterprises. As of March 2026, 84% of Uber's developers were using agentic workflows, which require significantly more tokens than simpler, single-turn interactions, further compounding cost issues.
Infrastructure costs related to token pricing largely stem from the reliance on high-performance GPUs, with on-demand pricing for the NVIDIA H100 ranging from $1.49 to $6.98 per hour. AI labs must maintain thousands of these GPUs simultaneously to meet enterprise demand, which is directly reflected in API token pricing.
Competitive Strategies in AI
In contrast to the rising costs faced by OpenAI and Anthropic, Google's new Gemini 3.5 Flash model offers faster, cheaper, and more efficient AI capabilities. Google claims that enterprises could save over $1 billion annually by shifting workloads to this model. The cost advantages arise from several factors, including Google's development of its own Tensor Processing Units, which reduces reliance on external GPU pricing. Google’s internal applications are achieving a scale that enhances model efficiency and lowers costs over time.
These developments raise important questions about the sustainability of revenue growth for AI firms like OpenAI and Anthropic. As enterprises deal with escalating AI costs, the likelihood of a shift in spending priorities becomes increasingly apparent. Although OpenAI currently enjoys a revenue lead, the long-term viability of its model may face scrutiny if these financial pressures persist.
The evolving situation highlights the need for AI companies to adjust their pricing strategies and infrastructure approaches to maintain competitive advantages and secure enterprise partnerships. As the market changes, the next moves by these key players in the AI sector will be key in determining their futures in an increasingly challenging environment.
