Enterprise AI Statistics 2026: ROI, AI Agents & Spending

Last updated: June 2026. Reviewed by Report AI editorial. Every figure on this page is linked to its primary source.

ENTERPRISE AI 2026 — KEY DATA POINTS

$37B

Enterprise GenAI spend
2025 (3.2× YoY)
Menlo Ventures

23%

Scaling agentic AI
+39% experimenting
McKinsey

40%+

Agentic projects canceled
projected by 2027
Gartner

Enterprise generative-AI spend, 2024–2025 (Menlo Ventures) · See all comparison charts →

Enterprise generative-AI spending reached roughly $37 billion in 2025 — more than triple the prior year (Menlo Ventures). At the same time, 23% of organizations are now scaling agentic AI with another 39% experimenting (McKinsey), yet Gartner warns that more than 40% of agentic AI projects will be canceled by 2027 (Gartner). Below are the key enterprise AI statistics for 2026, each dated and sourced.

Executive summary

  • Fast Fact: Enterprises spent $37B on generative AI in 2025 — split roughly $19B applications / $18B infrastructure — with coding the single largest departmental category (~55% of departmental application-layer spend).
  • Primary Driver: Agentic AI moved from concept to production in 2025–26. 23% of organizations are scaling agentic systems and Gartner expects 33% of enterprise applications to embed agentic AI by 2028 (up from <1% in 2024).
  • The Bottleneck: Only 39% of organizations attribute any EBIT impact to AI, and Gartner projects 40%+ of agentic projects will be canceled by 2027 due to unclear ROI and weak governance.
  • ROI Gap: Expectations are high — PagerDuty’s 2025 survey finds companies anticipate 171% average ROI (192% in the U.S.) on agentic AI — but those are projections, not realized returns.

Enterprise AI spending & deployment

MetricFigureSource
Enterprise GenAI spending (2025)$37B (3.2× 2024)Menlo Ventures
Application vs. infrastructure split$19B / $18BMenlo Ventures
Organizations scaling agentic AI23% (+39% piloting)McKinsey
Expected agentic AI ROI171% avg (192% U.S.)PagerDuty
Enterprise apps with agentic AI by 202833% (from <1% in 2024)Gartner
Agentic AI projects canceled by 202740%+Gartner

Where enterprise GenAI budgets go

Within the $19 billion 2025 application-layer spend, the departmental breakdown is concentrated in engineering — coding tools alone account for roughly $4 billion (~55% of departmental spend), per Menlo Ventures:

DepartmentShare of departmental AI spend
Coding / engineering~55% ($4.0B)
IT~10%
Marketing~9%
Customer success~9%
Design~7%
HR~5%

Deep dive: why coding is the biggest enterprise use case

Coding has clean, measurable productivity outcomes and shows the fastest time-to-value among enterprise GenAI workloads. The category includes AI pair-programmers, agentic IDE assistants, code review, and automated migration tooling — all of which compound on a developer base that organizations are already paying for, making per-seat ROI relatively easy to justify.

The rise of AI agents

Agentic AI is the defining enterprise trend of 2026. McKinsey finds 23% of organizations already scaling an agentic system, with another 39% experimenting. Gartner expects 33% of enterprise software applications to embed agentic AI by 2028, up from less than 1% in 2024 — while cautioning that much of today’s activity is early-stage and that “agent washing” (rebranding non-agentic products as agents) is widespread.

ROI: high expectations, uneven reality

Expectations are sky-high: in PagerDuty’s 2025 survey of 1,000 executives, companies anticipate an average 171% ROI on agentic AI (192% in the U.S.), and more than half say they’ve already deployed agents. But these are projected returns. The realized picture is more sober — McKinsey finds only 39% of organizations attribute any EBIT impact to AI so far, and a small group of roughly 6% “AI high performers” captures a disproportionate share of the value. The gap between expected and realized ROI is the defining enterprise-AI story of 2026.

Top barriers: governance & cost

  • Governance & risk controls. Gartner projects 40%+ of agentic AI projects will be canceled by 2027 due to unclear ROI, escalating costs, and inadequate controls.
  • Agent washing. Gartner estimates only a small fraction of vendors claiming “agentic” capabilities actually deliver them — buyer due diligence is now critical.
  • Workflow redesign. Only the small group of high performers reports meaningful EBIT impact — they redesign workflows around AI rather than bolt it on.
  • Runaway inference cost. Although unit inference is cheap, total cost can scale rapidly with agentic loops and multi-step reasoning.

Frequently asked questions

How much do enterprises spend on AI?

Enterprises spent roughly $37 billion on generative AI in 2025, up 3.2× from $11.5 billion in 2024 — split approximately $19B on applications and $18B on infrastructure (Menlo Ventures).

What is the ROI of enterprise AI in 2026?

Expectations are high — PagerDuty’s 2025 survey finds companies anticipate an average 171% ROI on agentic AI (192% for U.S. enterprises) — but these are projections. In practice, McKinsey reports only 39% of organizations attribute any EBIT impact to AI, with returns concentrated among a small group of high performers.

How many enterprises use AI agents?

McKinsey finds 23% of organizations are scaling an agentic AI system and 39% are experimenting. Gartner forecasts 33% of enterprise software applications will embed agentic AI by 2028, up from under 1% in 2024.

Why do enterprise AI projects fail?

The main causes are weak governance, unclear ROI, and runaway costs. Gartner projects over 40% of agentic AI projects will be canceled by the end of 2027.

Data sources & methodology

This page indexes primary research from leading consultancy reports, venture-capital industry surveys, and analyst forecasts.

  1. Menlo Ventures2025: The State of Generative AI in the Enterprise, December 2025.
    Verified data points: $37B enterprise GenAI spend; 3.2× YoY growth; $19B/$18B application/infrastructure split; departmental spend breakdown.
    Source: menlovc.com/perspective/2025-the-state-of-generative-ai-in-the-enterprise
  2. McKinsey QuantumBlackThe State of AI, November 2025.
    Verified data points: 23% scaling agentic AI (+39% experimenting); 39% EBIT impact attribution; ~6% AI high performers.
    Source: mckinsey.com/quantumblack/state-of-ai
  3. GartnerPredicts Over 40% of Agentic AI Projects Will Be Canceled by End of 2027, June 2025.
    Verified data points: 40%+ agentic AI cancellation forecast; 33% enterprise apps with agentic AI by 2028.
    Source: gartner.com
  4. PagerDutyAgentic AI Survey 2025.
    Verified data points: 171% avg expected ROI (192% U.S.); 1,000 executives surveyed; majority report at least one agent deployed.
    Source: pagerduty.com/newsroom/agentic-ai-survey-2025

Related pages: AI Adoption Statistics 2026 · Generative AI Statistics 2026 · Methodology

Related Reports & Resources

Other reports in this cluster (Business & Enterprise): LLM Market Report 2026 — provider share, enterprise LLM spend, and the Anthropic-vs-OpenAI shift driving enterprise AI budgets.

Compare year-over-year: Enterprise GenAI spend chart · LLM market share chart · State of AI by year.

Background reading: AI Adoption Statistics 2026 · Generative AI Statistics 2026 · AI Investment & Funding 2026.

Key concepts: AI Agent · Generative AI · Large Language Model.

Browse the category: All Statistics & Reports → Business & Enterprise cluster.