What Microsoft and Uber's Runaway Token Costs Mean for Enterprise CMOs

The recent news that Microsoft is canceling most Claude Code seats and that Uber spent a full-year 2026 AI coding budget in four months has big implications for enterprise CMOs. Two iconic companies with deep pockets decided that token-priced AI has unexpectedly high cost with unquantifiable ROI. For an enterprise B2B CMO already running flat budgets, AI shifts a real share of marketing's costs from fixed to variable, with returns that are hard to pin down. Token-consuming AI in your marketing org is the fastest path to a runaway-spend conversation with the CFO.

For the enterprise CMO, the next conversation with the CFO is not “What is your AI strategy?” but “What could marketing's AI agents and tools cost at the high end, how are you forecasting it, and how will you know when to pull the plug?”

We built Floma as an AI-native managed service for B2B tech ad campaigns because I have lived the pressure that enterprise CMOs face. A managed service with predictable outcome-based pricing is the financial opposite of unbounded token-based AI. You don't get a token bill, add another tool to your stack, or need to put anything in front of your AI review board. You don't need to perform the last-mile work of auditing a campaign to ensure that it is technically accurate, brand-perfect, persona-resonant, and channel-compliant. Instead, you get enterprise-grade ad campaigns at AI speed as an outcome. Our senior B2B marketing experts govern the Floma AI platform to ensure trust, technical accuracy and quality.

The math that Microsoft and Uber surfaced this week is going to push more enterprise CMOs toward outcome-based AI buying rather than agent, seat or tool-based AI buying. If you are ready to explore outcome-based AI for marketing, let's talk.