A few weeks ago, I traveled to Philadelphia for a conversation with a prospective CloudZero customer. We’d been working with the prospect’s engineering team for some weeks, demoing our platform in view of the RFP they’d drawn up. This stage had gone well, and so the next step was talking it over with the prospect’s CFO.
We expected a conversation centered around the key criteria in the RFP. Instead, the CFO told us, in no uncertain terms: “The most important thing, the only thing that matters, is AI cost allocation.”
The most important thing, the only thing that mattered, had somehow appeared nowhere on the RFP.
The Two Opposing AI Pressures
We’re seeing an unprecedented push from executives to update the central engine of business operations. By this, I of course mean: Adopt AI, whether you’re an experienced engineering leader, a sales development representative (SDR) fresh out of college, or anything in between. AI-nativeness promises profound new levels of output and efficiency and no executive wants to be caught holding the AI-skeptic bag.
On the other hand, under these directions, people have no reason to moderate their AI usage, and so they blow through AI budget estimates. Finance leaders suspected their estimates might have been wrong, but they’re winding up wrong by greater orders of magnitude than they thought possible. As in the early days of the cloud, these teams have no method for tying AI costs to revenue-producing activities, so it becomes near-impossible to assess their ROI, much less decide how to modulate them.
Executives exert downward pressure on mid-level managers to get their teams to ramp up their AI usage, fast. In turn, this AI ramp-up creates an upward pressure on executives, who see rapidly rising AI costs coming from every department, and need to make intelligent decisions about where to increase or decrease capital expenditures.
Executives, like this prospect’s CEO, wind up between a rock and a hard place. They don’t want to discourage AI usage, but they also don’t want to encourage reckless spending, or spending that can’t be tied back to revenue-producing business initiatives.
This may sound like an unresolvable catch-22. But it isn’t — and it’s those same mid-level leaders who hold the key to its resolution.
Free Guide
2026 Buyer’s Guide To FinOps Platforms
Find the right FinOps platform to grow your cloud maturity.
The Limits Of ‘Strategic Unprofitability’
AI costs have a lot in common with their spiritual predecessor (and fellow budget-breaker), cloud costs. The cloud made it possible for engineers to bypass the hardware-provisioning process altogether and innovate on much tighter timelines. It was therefore easy for leadership to ramp cloud investments, and given the magnitude of ROI those investments promised, hard to modulate them. It was an era of “strategic unprofitability”: accepting some suboptimal balance sheets to capitalize on a changing status quo.
When the status quo had changed, modulation became necessary. Cloud financial operations (FinOps) became a staple of engineering organizations.
Like the cloud, AI has made it possible for people to build without the restrictions they faced in the past. But unlike the cloud, these newly unrestricted people are not just engineers. They’re marketers, product managers, sellers, security pros; they’re anyone and, per executive mandate, everyone.
Between the executives and the people incurring these costs sit finance leaders like me. We’re responsible for P&Ls, for budgeting and forecasting, for being the single source of truth of how expenses in any area correlate with returns. We know that managing cloud costs is an urgent matter, and we should know that managing AI costs has become a C-level concern. It may not have reverberated throughout the organization yet, but it will, and it will come from executives desperate for dependable answers.
The directive you will get, if you haven’t already: Tie AI costs to the capital-eligible projects they support, and give me a sense of where to increase or decrease my investments.
How can you do that?
An All-Seeing Eye For AI Costs
CloudZero is the market leader in AI cost visibility. For years, we’ve helped companies like Nubank, Coinbase, DoorDash, and more manage their cloud costs through a data model that offers unparalleled allocation and unit economics capabilities. We help the world’s most sophisticated cloud spenders organize their cloud costs, eradicate inefficiencies, manage their margins, and identify the most promising areas of investment.
CloudZero does the same for AI. We collect usage data that allows us to allocate every penny of your AI costs to the teams, products, and/or features driving them, allowing you to associate dollar investments with value-based returns. How that new outbound email automator correlated with pipeline growth, for example, or how new AI-powered features impact with the costs to support your biggest customers — and how to gauge the contract accordingly.
AI costs are an executive-level concern that arise from natural market pressures. FinOps leaders sit at the center of these swirling pressures, and can play an integral role in their resolution.
FinOps In The AI Era: A Critical Recalibration
What 475 executives told us about AI and cloud efficiency.
