Enterprise budgets for artificial intelligence have exploded. Global spending is on track to hit $2.59 trillion in 2026, a 47 percent jump from the prior year, according to forecasts cited across multiple reports. Yet surveys show most organizations still struggle to prove the money pays off.
Only about 25 percent of AI initiatives deliver the expected returns, per an IBM CEO study referenced in recent analyses. Just 16 percent reach enterprise-wide scale. PwC’s 2026 CEO Survey found 56 percent of leaders saw neither revenue gains nor cost cuts from AI in the past year. Only 12 percent achieved both.
Boards and investors are pressing harder. Kyndryl’s 2025 Readiness Report showed 61 percent of executives feel more pressure to demonstrate ROI than a year earlier. Teneo’s Vision 2026 survey noted 53 percent of investors want positive returns within six months. Many firms lack the benchmarks to judge success at all, Wedbush analysts observed at their recent Disruptive Technology Conference.
MIT Sloan Management Review laid out three practical paths companies are taking. The function-focused approach starts with one high-value, repeatable use case inside a single department. The coordinated approach brings consistency across units once pilots multiply. The enterprise portfolio approach treats AI spending like any other strategic investment, with governance at scale. Most organizations move through these stages over time, the researchers found after interviewing more than 30 senior leaders.
Measurement remains inconsistent. Some track hours saved or tickets deflected. Others chase revenue attribution or risk reduction. The Business Manual highlighted metrics such as agent-to-human ratio optimization and the share of tasks that auto-resolve without escalation. These translate speed and volume into cost-per-outcome figures.
Recent RBC Capital Markets CIO survey data, reported by Business Insider on June 26, 2026, paints a picture of accelerating adoption. All respondents allocate budget to AI and large language model projects. Ninety-one percent created entirely new AI budgets rather than reallocating existing funds. More than half said AI is already in production; another 35 percent expect production within six months. Token costs have not triggered the feared panic. Nearly nine in ten called budgets manageable, and most plan to spend more, not less.
OpenAI dominates enterprise usage in that survey. Fifty-seven percent named ChatGPT their primary model-based service. Only 12 percent pointed to Anthropic’s Claude. Performance perceptions follow the same split: 44 percent rated OpenAI highest versus 24 percent for Anthropic. Enterprise revenue now accounts for more than 40 percent of OpenAI’s total and is on pace to reach parity with consumer by year-end.
Broader data shows mixed results. Forbes reported in January 2026 that fewer than 1 percent of executives see significant ROI of 20 percent or greater. Fifty-three percent report only 1-5 percent returns. IBM data echoed low confidence, with just 29 percent of executives saying they can measure ROI reliably. Deloitte’s State of AI in the Enterprise found two-thirds of organizations report productivity gains, but revenue growth remains largely aspirational—only 20 percent have achieved it while 74 percent hope to.
Internal tools often deliver faster payback than customer-facing features. One 2026 product-leader survey found coding assistants and internal AI tooling cited far more often for measurable impact than shipped products. Larger companies with over 1,000 employees show broader adoption and stronger results, NVIDIA’s State of AI report noted. Financial services, retail, and healthcare lead in both deployment and returns.
Companies that succeed tend to set baselines before deployment, then track changes across cost removal, revenue influence, and risk reduction. They tie AI directly to P&L outcomes rather than activity metrics. Hybrid pricing models that blend seat licenses with usage fees have gained rapid traction as the preferred buying structure.
Pressure is mounting as spending scales. Gartner projects AI application software spending alone will exceed $270 billion in 2026. The shift from pilots to production is real for many, yet the gap between investment and verifiable financial impact persists for the majority. Firms without explicit measurement frameworks continue to lag those that treat AI ROI with the same discipline applied to any other capital project.
WebProNews is an iEntry Publication