Token Billing Exposes AI's Missing ROI And Puts Billion-Dollar Bets At Risk

Anthropic just closed a $65 billion Series H at a $965 billion valuation. The same week, one of its enterprise customers accidentally spent $500 million in a single month on its models after failing to set spend limits. The gap between those two numbers is the story. For most of the generative AI era, enterprise pricing was subsidized and opaque. Flat-fee subscriptions absorbed unlimited token burn, and the actual cost of any given task remained invisible to finance teams. That changed in Q1 2026, when Anthropic and OpenAI quietly moved enterprise customers to token-based billing. The shift turned a diffuse budget line into a measurable, per-task cost. What it revealed is making investors uncomfortable. Uber is the most public case. The company burned through its entire 2026 AI coding tools budget by April after rolling out AI tools at near-total scale across its engineering organization. COO Andrew Macdonald then acknowledged at a May 25 conference that despite 95% of engineers using AI monthly, he could not draw a line between that token spend and meaningful consumer-facing product improvements. "That link is not there yet," Macdonald said. Microsoft, facing Claude Code bills running $500 to $2,000 per engineer monthly, began canceling direct Claude Code licenses and routing engineers back to GitHub Copilot. The ROI problem has two layers. The first is output quality: LLMs hallucinate, loop, and fail in ways that are difficult to predict, and every failed run costs tokens regardless of outcome. The second is structural: there is no standard unit for measuring the cost of an AI task because the same task can consume wildly different token counts depending on the prompt, the model version, the context window, and whether the agent makes wrong turns. Token-based billing made the spend visible without making it legible. This matters for venture capital because the current wave of AI infrastructure investment is predicated on enterprise AI becoming a durable, recurring revenue line. Gartner projects AI agent software spending will hit $207 billion in 2026, up 139% from 2025. That trajectory assumes enterprises continue to expand AI spend. The Uber signal, and the pattern of companies quietly pulling back token consumption, suggest the trajectory is under pressure at the margin. As The Street noted, the companies selling tokens benefit from current adoption regardless of whether buyers can show ROI. The question is how long that asymmetry holds once CFOs can see the line item. Anthropic CEO Dario Amodei has acknowledged the timing risk explicitly, in a different context. In a February interview, he warned that if AI revenue growth forecasts are off by even a year, "then you go bankrupt" — which is why he has kept capital expenditure more conservative than the hyperscalers. He was referring to Anthropic's own infrastructure bets, but the logic applies to his enterprise customers too. If token-based billing reveals that the productivity gains do not justify the cost, enterprises do not go bankrupt; they just stop renewing. GitHub Copilot's June 2026 move to token-based billing provided the clearest retail-level evidence yet. Users on the promotional tier reported burning 30 to 60 percent of monthly credits in a handful of prompts. One user said Copilot went from their favorite subscription to their most stressful overnight. These are developers, the cohort with the highest AI literacy and the strongest motivation to make AI tools work. If the cost-value calculation is breaking down for them, the enterprise rollout projections are built on shakier ground than the valuation multiples suggest. For investors, the token billing transition is the first real price discovery mechanism the AI industry has produced. Flat-fee subscriptions created a comfortable fiction: costs were low, adoption was high, and ROI was a question for later. Usage-based billing asks the question now. Anthropic's path to justifying a near-trillion-dollar valuation runs directly through enterprises proving, to their own finance teams, that tokens are worth buying.