AI budgets surge despite mounting ROI doubts and c-suite tension
The gist
AI budgets are ballooning even as most companies can’t prove they’re getting their money’s worth—and the C-suite is hitting a boiling point over the payback.
What to know
- Surveys show that by late 2026, 67% of supply chain tech budgets and up to $40 billion in enterprise investment went to AI, but 95% of companies saw zero measurable returns.
- Only 14% of organizations can clearly tie all AI initiatives to profit and loss, leaving CEOs and CFOs at odds with tech leaders over accountability and value.
- Despite thin evidence of ROI, 90% of executives expect AI budgets to keep rising in 2027—even as the average reported impact hovers at a lukewarm 5.36 out of 10.
AI ROI: Perception vs. Reality
Despite massive adoption and daily use, most executives admit their companies can’t deliver on AI’s promised returns, exposing a widening credibility gap between investment and impact.
By late 2026, the ROI gap was no longer anecdotal; it was showing up across surveys in different functions and geographies. Gartner found that 55 percent of CSCOs were not sure about the ROI of their companies’ AI investments, even though 67 percent of supply chain digital investments had already been allotted for AI. Writer’s poll of 2,400 employees found that 70 percent of employees and 94 percent of C-suite respondents use AI tools for at least half an hour per day, yet 66 percent of leaders said executives at their company talk a big game about AI but cannot deliver.
Other late-2026 research made the mismatch starker by pairing large spending totals with weak proof of payoff. MIT found that despite up to $40 billion in enterprise investment into Generative AI, 95 percent of companies were getting zero return, while Writer reported that only 29 percent saw significant ROI in GenAI tools and 23 percent in AI agents. IBM found that 79% of executives report seeing productivity gains from AI, while only about 29% can measure its ROI confidently, even as AI Engineer noted a 2x increase in enterprise AI spend year over year. METR’s randomized controlled trial added that experienced developers took 19% longer on real tasks with early-2025 AI tools, though afterward they estimated AI made them about 20% faster.
Missing Links in AI Accountability
AI initiatives stall at the pilot stage because firms lack clear governance and financial ownership, leaving most unable to track value or connect projects to the bottom line.
The bottleneck is not AI access but managerial plumbing: companies are launching initiatives without the governance, ownership and financial logic needed to connect them to profit and loss. BCG found that more than half of CEOs cite linking AI initiatives to the P&L as a key barrier, yet only 14% have clearly defined P&L impact for all AI initiatives, a 42-point gap; it also identified weak value tracking, insufficient funding for people and change management, and a governance imbalance in which only 30% include HR versus 82% including technology.
That missing accountability framework helps explain why deployment so often stalls between experimentation and enterprise change. BCG said nearly two-thirds of CEOs report pursuing AI pilots but only 26% have embedded AI into broader business transformation, while CPA Practice Advisor, citing Protiviti, reported that many finance organizations still have not satisfactorily connected AI investments to measurable business outcomes despite 77% already employing AI; in marketing, Invoca found 74% cannot tell whether agent-handled leads become sales, only 27% measure revenue, bookings or conversions, 56% can’t track once a lead reaches a dealer, franchise or local branch, and tracking often breaks before or after handoff.
CFOs Demand Proof, CIOs Push On
Boardroom rifts deepen as finance leaders question AI’s financial payoff while tech chiefs double down, with most stakeholders now prioritizing hard revenue metrics over hype.
As AI spending rises, the argument inside the C-suite is no longer about whether to experiment but whether the payoff is real. On Business Security Weekly, speakers described CFOs telling CEOs that AI is becoming “one of the most expensive experiments we’ve gone through,” pulling chief executives into ROI scrutiny even as “the CIOs are going to be evangelists”; they also cited a figure that “72% of AI initiatives failed,” reinforcing why CEOs and boards reported less confidence than CIOs about whether current AI programs are actually paying back.
That split is also visible in what finance-oriented stakeholders now want measured. Fidelity found that among 318 respondents, 42% believed the value of some AI-related companies had become too high relative to what those companies were expected to deliver financially, while 27% said AI spending plans were taking longer than initially anticipated; among 449 respondents, 42% said the metric they would watch most closely was AI-driven revenue growth, versus 21% for corporate guidance and 18% for capital expenditures, underscoring finance leaders’ demand for proof over technology leaders’ push to keep deploying.
Spending Surges, Results Lag
Even as most pilots fail to move the needle, companies keep shifting budgets from proven tactics to AI, betting on future gains despite lackluster current results.
Budget momentum has not slowed just because proof is thin; in many cases, AI is gaining share even when overall spending is not. Research & Data cites MIT’s NADA Gen AI Divide Study finding that “95% of enterprise AI pilots delivered no measurable P and L impact,” yet, inside Gartner’s 2025 CMO spend survey where marketing budgets “flatlined at 7.7percent… of company revenue,” organizations were “starting to reallocate to AI tooling and away from media” — a shift framed as “worse than most people realize” because “we’re reallocating away from things that work towards things that mostly don’t.”
That pattern extends beyond marketing: the pressure point is no longer whether companies will keep funding AI, but whether they can justify doing so. FreightWaves said the market is pivoting “from the question of does this work to how can you prove this works and how can you generate more return faster,” even as “another 90% of the survey respondents believe that their investment in technology and AI their budgets are going to increase”; similarly, CLA reported on Nov. 17, 2026 that “more than half of respondents expect to invest in automation and AI over the next year,” although just 21.3% saw highly favorable gains, nearly 28% rated impact 3 or below, and the average score was 5.36.






