Dive Brief:
- Roughly one in every four dollars spent on AI goes to waste as enterprises accelerate efforts to bring spending under control, according to Harness data published Wednesday. The company surveyed 700 FinOps and engineering leaders for the report.
- More than half of respondents said their organization lacks a dedicated owner for AI costs, leaving responsibility diffused across engineering, platform and FinOps teams.
- While expenditure is climbing fast, governance and oversight structures needed to manage it haven’t yet caught up, with only one in five organizations able to determine the source of unexpected AI cost spikes within hours.
Dive Insight:
Tech leaders are working hard to rein in runaway AI spending as adoption spreads across the organization. The tricky balance is in enabling innovation without letting unclear guardrails and tool sprawl derail budgets.
Alongside rising spending, Harness’ report found businesses are getting stumped by how AI costs differ from traditional IT spending.
Unlike cloud infrastructure, where costs are tied primarily to servers and storage, AI expenses span multiple categories simultaneously, including infrastructure, foundation models, SaaS subscriptions and managed services.
Most organizations also use three or more major AI providers, each with different pricing structures, making it difficult to have a clear view of spending.
Productivity software — such as AI copilots and coding assistants — has also emerged as one of the largest cost drivers. Since they look like ordinary software licenses rather than infrastructure spend, they are difficult to track.
Vendors are already responding to the rising need for oversight. Major cloud providers, including Oracle and AWS, have rolled out new features and billing structures for AI cost visibility, while the Linux Foundation launched a new group focused specifically on cost management.
Yet lack of visibility remains a barrier for many.
“What surprised me most wasn’t the size of the spend or the speed of the growth, but how consistent the gaps are across every size and geography,” Harish Doddala, VP of cloud and AI cost management at Harness, said in the report. “The visibility problem, the ownership problem, the forecasting problem show up whether you are spending $300K a month or $3M.”
More than half of respondents said they forecast AI spending through guesswork rather than data, while more than 40% continue relying on spreadsheets to manage costs. As AI budgets become material enough for CFO scrutiny, delayed visibility can leave organizations reacting to spending rather than proactively controlling it.
Doddala says time alone will not be enough to close the gaps in visibility. Targeted investments are needed to get a better handle on spending trajectories.
Assigning ownership early, centralizing cost visibility and integrating spending data into engineering workflows are some of the strategies deployed by organizations with stronger cost discipline, according to Harness.
OpenAI offered similar advice earlier this month, urging businesses to increase visibility into AI use and spend, track model outcomes and build governance into day-to-day operations.
