FTSE 100 AI spend stays hidden as agent costs begin to spiral

New Flexera analysis finds only 2% of FTSE 100 firms disclosed an AI spending figure, even as global AI debt tops $450bn.

By Disrupts Editorial Team Tue, 10 November 2026 AI level 2 of 5

AI level 2 of 5: drafted by our AI editorial assistant from the company's announcement; fact-checked, edited and signed off by Mark Walker, Editorial Director. What the levels mean

A control room with a curved wall of twelve large monitors displaying various data visualizations, rows of empty desks and chairs in the foreground, and bright ambient light from large windows and overhead panels.

Flexera, a technology spend and risk intelligence firm, has published analysis showing that just 2% of FTSE 100 companies disclosed a specific AI expenditure figure in their 2025 annual reports. Only 3% reported a measurable financial return from AI. The finding creates a sharp accountability gap: 36% of the FTSE 100 reported substantive AI deployment, and 32% linked the technology to a named business outcome, yet almost none put a verifiable number on what it cost or returned.

The timing matters. The Bank of England's Financial Policy Committee recently cited Morgan Stanley estimates placing global AI-related debt issuance at approximately $450bn by early September 2026, more than double the total for 2025. That capital is flowing into corporate AI programmes that, according to this analysis, are largely opaque to shareholders.

The agent problem raises the stakes

The disclosure gap is about to become more difficult to close. OpenAI's always-on "Dots" agents, which autonomously carry out background tasks, introduce a consumption model that is structurally harder to budget for than traditional software licences. Each agent instruction can spawn repeated compute cycles without direct human intervention, compounding costs in ways that quarterly finance reviews are poorly designed to catch.

Marlon Oliver, Senior Vice President for EMEA at Flexera, put the problem plainly: "Companies are burning through annual AI budgets in months, and always-on agents will put those budgets under even more pressure. Token consumption, compute and licence numbers tell you what AI costs and how much it is being used. They do not tell you whether it is creating value."

One anecdote from Flexera's research illustrates the scale of the hidden-cost problem: a single enterprise discovered after the fact that just four users were responsible for 35% of its AI spend. At FTSE 100 scale, that kind of concentration could represent tens of millions in unattributed cost.

Cross-sector read-across: from IT budgets to capital markets

For cross-sector investors, the disclosure void is a due-diligence problem that cuts across every sector in which FTSE 100 companies operate, from financial services and consumer goods through to energy and defence procurement. When 54% of the FTSE 100 simultaneously identify AI as a material or company-specific risk, yet almost none quantify the associated expenditure, the information asymmetry between boards and shareholders widens precisely as AI run-rates accelerate.

The Bank of England's October 2026 agents' summary adds a further dimension: firms reporting AI-driven labour savings are increasing spending on software and cloud services in parallel. That dynamic suggests AI productivity gains are, at least in part, being recycled into the infrastructure that sustains them, a reinvestment loop that does not appear in most AI reporting frameworks and that standard accounting categories struggle to capture.

The practical pressure on audit committees and CFOs is growing. As agentic AI systems shift enterprise technology from fixed-licence to consumption-based pricing, the gap between internal cost visibility and external disclosure will attract increasing scrutiny from institutional investors, credit analysts, and, eventually, regulators. The Financial Reporting Council has not yet issued specific guidance on AI cost disclosure; how long that position holds will be a key signal for anyone allocating capital into UK-listed equities with significant AI exposure.