AI Spending Statistics 2026
By Axis Intelligence Research
Co-author: Sarah Mitchell | Last updated: July 3, 2026 | License: CC BY 4.
Quick Answer
Worldwide AI spending will reach $2.59 trillion in 2026 — a 47% jump from 2025 — yet only 6% of organizations qualify as AI high performers with measurable bottom-line impact, per McKinsey’s survey of nearly 2,000 companies. The gap between spending and returns is now the defining tension of the AI market. Axis Intelligence Research calculates an AI Spending Efficiency Index (ASEI™) of 58.2 for 2026 — down from 118.2 in 2025 — meaning that as the number of heavy AI spenders doubled, the proportion capturing enterprise-level returns was cut nearly in half.
According to Axis Intelligence Research’s cross-source analysis of Gartner, BCG, and McKinsey primary data: global AI spending reaches $2.59 trillion in 2026, with infrastructure alone accounting for $1.43 trillion (55.2%). Enterprise budgets are doubling — from 0.8% to 1.7% of revenue on average — while only 39% of organizations report any measurable EBIT impact. The Axis Intelligence ASEI™ score of 58.2 quantifies the widening gap between investment and returns.
Key Findings
- Global AI spending reaches $2.59 trillion in 2026, a 47% year-over-year increase, with AI infrastructure alone accounting for $1.43 trillion — per Gartner’s May 2026 forecast, its most current and highest estimate.
- Enterprise AI budgets are doubling: corporations plan to spend 1.7% of revenues on AI in 2026, up from 0.8% in 2025, with 94% committed to continued investment even without immediate returns — per BCG’s AI Radar 2026, a survey of 2,360 executives across 16 markets.
- Only 6% of organizations are AI high performers: McKinsey’s State of AI survey of ~1,993 companies finds just 109 organizations — 5.5% — report that AI accounts for more than 5% of enterprise EBIT, while 61% of heavy spenders show no high-performer status.
- The ASEI™ score fell from 118.2 in 2025 to 58.2 in 2026: Axis Intelligence Research’s AI Spending Efficiency Index, derived by cross-referencing BCG budget thresholds with McKinsey EBIT data, shows that doubling spending has not doubled returns — the efficiency of every dollar spent on AI has declined as the market broadens.
- AI agent software is the fastest-growing segment at +139% in 2026: Gartner projects AI agent spending will reach $206.5 billion this year, rising to $376.3 billion in 2027, as enterprises move from generative AI experimentation into autonomous workflow deployment.
How Much Is the World Spending on AI in 2026?
The $2.59 trillion figure from Gartner is the most widely cited number in enterprise AI planning for 2026 — and also one of the most misunderstood. It is not a single market. It is a taxonomy of fundamentally different economic activities stacked under one label.
Gartner published two forecasts in 2026. The January release put global AI spending at $2.52 trillion (44% growth). The May revision raised that to $2.59 trillion (47% growth), adding roughly $70 billion driven primarily by accelerating agentic AI demand and faster-than-expected enterprise adoption of AI models. Gartner explicitly cautions that the two vintages are not directly comparable — the May figure folded in AI cybersecurity and agentic AI as distinct line items. For any citation, the May 2026 number is the operative one.
The breakdown by segment — published across Gartner’s January and May 2026 releases and reported by CFOtech Australia and CIO Dive — shows a market weighted heavily toward physical infrastructure:
| AI Spending Segment | 2025 ($B) | 2026 ($B) | YoY Growth | Share of Total (2026) |
|---|---|---|---|---|
| AI Infrastructure (servers, IaaS, chips, network) | $975.6B | $1,430.0B | +46.6% | 55.2% |
| AI Services (implementation, managed services) | $436.4B | $585.5B | +34.2% | 22.6% |
| AI Software (applications, platforms, tools) | $282.9B | $453.2B | +60.2% | 17.5% |
| AI Cybersecurity | $25.9B | $51.3B | +98.1% | 2.0% |
| AI Models (GenAI model spending) | $15.5B | $32.6B | +110.3% | 1.3% |
| Other (data, platforms, agentic — estimated) | ~$24.7B | ~$37.4B | — | 1.4% |
| Total | ~$1,761B | $2,590B | +47% | 100% |
Source: Gartner, “Gartner Forecasts Worldwide AI Spending to Grow 47% in 2026,” May 19, 2026; Gartner, “Gartner Says Worldwide AI Spending Will Total $2.5 Trillion in 2026,” January 15, 2026. Note: AI agent software ($206.5B) is a sub-segment within AI Software, not an additive line.
Three numbers in this table deserve special attention. AI cybersecurity nearly doubling in a single year reflects the expanded attack surface that AI deployments create — every AI model, every agent, every API endpoint is a new credential to protect. The AI models line growing 110% is the most direct signal of the generative AI wave hitting enterprise purchasing. And infrastructure at 55.2% of the total makes clear that the 2026 AI market is still fundamentally a physical build-out story, not a software-value story.
Sarah Mitchell, AI & Machine Learning Editor: The infrastructure dominance bothers me more than the headline number. $1.43 trillion going to servers, chips, and networking — before a single enterprise workflow is redesigned — means the value creation hasn’t started for most of the market. Hyperscalers are buying compute. Enterprises are buying software seats. Neither is the same thing as capturing returns. The 47% spending growth is real. The returns story is much messier.
What Are Enterprises Actually Budgeting for AI?
Gartner’s $2.59 trillion is a global market total — it captures hyperscaler capex, vendor R&D, and enterprise spend in one number. To understand what companies are individually allocating, BCG’s AI Radar 2026 is the cleaner source.
Based on a survey of 2,360 executives across 16 markets and nine industries — including 640 CEOs — published January 15, 2026, BCG’s AI Radar 2026 report found that corporations plan to spend 1.7% of revenues on AI in 2026, up from 0.8% in 2025. That is a 112.5% budget increase in a single year — and it is happening even as only 60% of companies report any value from their investments.
The sector breakdown reveals meaningful variation:
| Industry Sector | Planned AI Spend as % of Revenue (2026) |
|---|---|
| Technology & Software | ~2.1% |
| Financial Institutions | ~2.0% |
| Healthcare & Life Sciences | ~1.8% |
| Consumer & Retail | ~1.6% |
| Manufacturing | ~1.3% |
| Industrial & Real Estate | <1.0% |
Source: BCG AI Radar 2026, January 15, 2026 (bcg.com). Survey of 2,360 executives, 16 markets, 9 industries.
The CEO ownership shift is as significant as the budget numbers. BCG finds that 72% of CEOs now say they are the primary decision makers on AI — double the share from the prior year. Half believe their job depends on getting AI right. The survey identifies three CEO archetypes: Trailblazers (15%) who invest aggressively and allocate 60% of AI budgets to upskilling; Pragmatists (70%) who invest when value is evident; and Followers (15%) who remain cautious.
Critically, 94% of organizations say they will continue investing in AI even without immediate returns. That persistence, in BCG’s framing, signals less conviction than it does competitive fear — companies are spending because the alternative, falling behind peers who are spending, feels more dangerous than the budget risk.
And nearly two-thirds of companies — 60% — report either minimal or no value from their AI investments despite those elevated budgets, per BCG’s March 2026 analysis of its cost-advantage research.
The AI Spending Efficiency Index™ (ASEI™)
Neither Gartner, BCG, nor McKinsey publishes a metric that directly connects enterprise AI budget intensity with enterprise AI returns. Axis Intelligence Research calculates one here by cross-referencing two independent primary sources, and discloses the methodology in full.
What ASEI™ measures: for every 100 organizations spending at or above the heavy-investment threshold (≥1.7% of revenue), how many report any measurable enterprise EBIT impact from AI?
Formula:
ASEI™ = (% organizations reporting any EBIT impact from AI)
÷ (% organizations spending ≥1.7% of revenue on AI)
× 100
Inputs (both primary-source verified):
- % organizations reporting any EBIT impact: 39% — McKinsey State of AI, November 2025 survey of ~1,993 organizations
- % organizations spending ≥1.7% of revenue on AI in 2026: 67% — BCG AI Radar 2026, January 2026 survey of 2,360 executives
- % organizations spending ≥1.7% of revenue on AI in 2025: 33% — BCG AI Radar 2026 (prior-year baseline)
Results:
| Year | EBIT Impact (%) | Heavy Spenders (%) | ASEI™ Score |
|---|---|---|---|
| 2025 (baseline) | 39% | 33% | 118.2 |
| 2026 (current) | 39% | 67% | 58.2 |
Interpretation: A score above 100 means returns are outpacing heavy investment — in 2025, there were more organizations reporting EBIT impact than there were heavy AI spenders, suggesting early movers captured disproportionate returns. The 2026 score of 58.2 means that for every 100 organizations now spending heavily on AI, only 58 report any EBIT impact. The cohort of heavy spenders doubled; the cohort capturing returns did not.
Methodology note and limitations: McKinsey’s 39% EBIT impact figure comes from self-reported survey data and covers “any level of EBIT impact,” not a threshold amount. BCG’s 67% figure is forward-looking (planned spend) rather than actuals. The EBIT capture rate (39%) may include organizations that are not heavy spenders — the cross-referencing assumes the two populations are randomly distributed, which is a simplification. We publish the ASEI™ as a directional index, not a precise accounting. The formula and both inputs are disclosed in full; researchers may recalculate with updated primary data as it becomes available.
Published under CC BY 4.0. Cite as: “Axis Intelligence Research ASEI™ 2026, axis-intelligence.com/ai-spending-statistics/.”
Sarah Mitchell: The ASEI™ drop from 118 to 58 is the number I’d want every CFO who just doubled their AI budget to see before the next board presentation. It doesn’t mean AI isn’t worth spending on — it means the marginal spender entering the market in 2026 is getting a systematically worse return than early movers got in 2025. The question isn’t whether to spend. It’s whether your organization is structured to be in the 39% that captures something, rather than the 61% that just has receipts.
AI Infrastructure Spending Statistics
AI infrastructure — servers, AI-optimized chips, networking fabric, and IaaS — is the largest and fastest-scaling segment of the 2026 AI market. Gartner projects it will reach $1.43 trillion in 2026, up from $975.6 billion in 2025, and climb further to $1.89 trillion in 2027.
The hyperscaler capex numbers confirm the trajectory. The four largest U.S. cloud providers are collectively guiding to roughly $725 billion in capital expenditure for 2026 — up 77% from approximately $410 billion in 2025 — with the overwhelming majority directed at AI data centers and GPU compute, per earnings disclosures compiled by Value Add VC research. For the electricity demand and energy cost tied to this buildout, see our AI data center energy consumption statistics. For the GPU hardware side of the same capex wave, see Nvidia Blackwell statistics.
| Hyperscaler | 2025 Capex | 2026 Capex Guidance | Primary Source |
|---|---|---|---|
| Amazon (AWS) | >$105B | ~$200B | Amazon Q4 2025 earnings |
| Microsoft | ~$80B | ~$190B | Microsoft FY26 Q3 earnings |
| Alphabet/Google | ~$85B | $175–190B | Alphabet Q1 2026 earnings |
| Meta | ~$65B | $115–135B | Meta Q1 2026 earnings |
| Combined | ~$335–340B | ~$680–715B | Per company filings |
Source: Earnings filings cited above; IDC press release, August 26, 2025. [older data — 2025 actuals; 2026 are company-guided, not yet reported]
IDC’s Worldwide Artificial Intelligence IT Spending Forecast, published August 2025, projects that year-over-year AI IT spending will grow at 31.9% between 2025 and 2029, reaching $1.3 trillion by 2029. IDC’s infrastructure-only tracker — which measures hardware shipments rather than total spend — reached $318 billion in full-year 2025 AI infrastructure, more than double the $153 billion recorded in 2024.
The structural reason for infrastructure dominance: compute remains the binding constraint. GPU supply, data center power capacity, and networking throughput are still limiting AI deployment more than software capability or enterprise readiness. Every dollar an enterprise commits to an AI agent or model runs on infrastructure built by a hyperscaler. The infrastructure spend is the floor; everything else builds on top of it.
AI Agent Software Spending Statistics
The fastest-moving segment in the 2026 AI spending picture is not infrastructure — it is AI agent software. Gartner puts purpose-built agentic AI software at $86.4 billion in 2025, rising to $206.5 billion in 2026 (+139%) and then to $376.3 billion in 2027 (+82%).
That two-year trajectory — $86B to $377B — is the fastest growth curve Gartner has recorded for any software category in the AI era. The driver is structural: enterprises are moving from asking AI to generate content to asking AI to execute workflows autonomously. Those are fundamentally different architectures with fundamentally different price tags.
| AI Agent Software Spending | Value |
|---|---|
| 2025 actual | $86.4B |
| 2026 forecast | $206.5B (+139%) |
| 2027 forecast | $376.3B (+82%) |
Source: Gartner, May 2026 forecast revision (gartner.com newsroom).
The adoption data sits well behind the spending projections. Only 17% of organizations have deployed AI agents to date, per Gartner. And the same forecast carries a significant caveat: over 40% of agentic AI projects are expected to be canceled before reaching production by the end of 2027. CEOs are committing over 30% of their 2026 AI budgets to agentic AI, per BCG — yet the deployment-to-cancellation ratio suggests the market is pricing in a technology that most organizations are not yet operationally ready to sustain.
McKinsey’s State of AI (November 2025) finds 23% of organizations are scaling agentic AI in at least one business function, and 62% are experimenting. But “experimenting” does not appear in P&L. For deployment-side data on what enterprises are actually building with those budgets, see our AI agents statistics report. The Claude AI statistics and ChatGPT statistics reports document the two platforms capturing the largest shares of enterprise AI spend at the application layer.
AI Spending by Industry
The industries spending most aggressively on AI — and capturing the clearest returns — are concentrated in sectors with structured, high-volume, rule-based workflows where AI can demonstrate measurable impact: financial services, technology, and healthcare.
Per IDC’s Worldwide Digital Transformation Spending Guide (V1 2026), financial services has the highest existing AI share of any industry’s DX software spend, driven by fraud detection, compliance automation, and risk scoring — areas where output is measurable and the cost of errors is priced. The sector plans to spend 2.0% of revenues on AI in 2026, the highest of any non-tech sector in BCG’s survey.
Technology and software companies lead all sectors at 2.1% of revenues. Manufacturing, despite slower historical adoption, showed the fastest year-over-year growth in worker-reported AI use in McKinsey’s 2025 survey — driven by predictive maintenance, quality inspection, and supply chain forecasting, all areas with clear cost baselines.
The sector that spent most in absolute terms without returning the highest EBIT impact: professional services. McKinsey’s data finds professional services firms among the highest in adoption and deployment, yet workflow redesign — the activity most correlated with EBIT impact in McKinsey’s data — is the step most firms in this category have not completed.
The ROI Gap: Why Most AI Spending Isn’t Showing Up in Results
This is where the data gets uncomfortable, and where Axis Intelligence Research thinks most AI spending coverage goes wrong by ignoring it.
McKinsey’s State of AI survey (November 2025, n=~1,993) is the most rigorous longitudinal dataset on this question. Its core finding: 88% of organizations use AI in at least one business function. 39% report any EBIT impact. 6% — 109 organizations out of nearly 2,000 — are genuine high performers with AI contributing more than 5% of enterprise EBIT. For the adoption side of this picture — which organizations are using AI and how — see our AI adoption statistics and AI in business statistics reports. For the governance structures that separate high-performer deployments from failed ones, see AI governance statistics and EU AI Act compliance statistics.
That is not a failure rate. It is a maturity distribution. But it does mean that for every organization in the news for a transformative AI deployment, there are roughly 16 that are spending at similar rates and producing results that don’t clear the EBIT detection threshold.
BCG’s March 2026 analysis of AI cost advantage puts numbers on why: 60% of companies report minimal or no value. 80–85% of enterprises miss their AI infrastructure forecasts by more than 25%, per Mavvrik’s cost governance research citing BenchmarkIT data. The measurement problem compounds the execution problem — McKinsey finds that fewer than one in five organizations is tracking well-defined KPIs for its AI solutions, which means many organizations cannot detect failure even when it is happening.
The organizations that make it into the 6% high-performer cohort share three documented patterns from McKinsey’s data:
First, they redesign workflows rather than layering AI onto existing processes. McKinsey identifies fundamental workflow redesign as the single strongest correlate with EBIT impact — not model selection, not budget level, not deployment speed.
Second, they allocate more than 20% of digital budgets to AI. More than one-third of high performers exceed this threshold, making them five times more likely to make a large bet than the rest of the cohort.
Third, they set growth or innovation as AI objectives alongside efficiency. Eighty percent of all organizations cite efficiency as the goal. High performers are three times more likely to also cite transformative change.
None of these three differentiators is a spending level. The implication is direct: scaling AI budgets without scaling workflow redesign and measurement discipline produces an ASEI™ score of 58, not 100.
AI Spending Forecast: 2027 and Beyond
The near-term trajectory from Gartner projects total AI spending reaching $3.49 trillion in 2027, with AI infrastructure alone hitting $1.89 trillion. IDC’s longer-horizon view sees AI IT spending reaching $1.3 trillion by 2029 at a 31.9% compound annual growth rate.
PwC’s “Sizing the Prize” report — a 2017 analysis still widely cited as the foundational long-run economic estimate — projected AI contributing $15.7 trillion to global GDP by 2030 [older data — this is a 2017 projection; no updated PwC equivalent has been published as of Q2 2026]. The infrastructure and enterprise spending data now in the market makes the $15.7 trillion number look directionally reasonable but analytically orphaned: it was built before the agentic era, before the current compute constraints, and before the ROI gap became measurable.
| Metric | 2025 | 2026 | 2027 |
|---|---|---|---|
| Total AI Spending (Gartner) | ~$1.76T | $2.59T | $3.49T |
| AI Infrastructure (Gartner) | $975.6B | $1,430B | $1,890B |
| AI Agent Software (Gartner) | $86.4B | $206.5B | $376.3B |
| AI Models Spending (Gartner) | $15.5B | $32.6B | — |
| Enterprise AI Budget (% revenue, BCG) | 0.8% | 1.7% | — |
| AI IT Spending (IDC, 2025–2029 CAGR) | — | — | $1.3T by 2029 |
Source: Gartner May 2026; IDC August 2025 press release; BCG AI Radar 2026.
The caution embedded in these projections: Gartner itself places AI in the “Trough of Disillusionment” throughout 2026. The 47% spending growth is happening alongside, not because of, mature enterprise value capture. John-David Lovelock, Distinguished VP Analyst at Gartner, stated in the May 2026 release: “Enterprises have yet to really flex their spending potential. That is coming and 2026 will be the inflection year.” The inflection point Gartner references is the shift from hyperscaler-driven to enterprise-driven spending — a shift that has been forecast for several consecutive years and is now arriving at scale.
Sarah Mitchell: The forecast convergence from Gartner, IDC, and BCG is clearer than it has been at any point since ChatGPT launched. The direction is not in question — spending grows. What remains genuinely open is whether the ASEI™ recovers in 2027 as workflow redesign catches up with infrastructure buildout, or whether the efficiency gap widens further as the enterprise market broadens to include organizations less structurally ready than the early adopters. My read of the McKinsey data is that the organizations entering heavy AI investment in 2026 are systematically less prepared than those who entered in 2024. That is not an argument against spending — it is an argument for spending differently.
AI Spending Statistics by Country
The geographic distribution of AI spending mirrors the distribution of private AI investment, with the United States commanding the largest share of both.
Per Stanford HAI’s 2026 AI Index (Economy chapter), U.S. private AI investment reached $285.9 billion in 2025 — 23.1 times China’s $12.4 billion in tracked private investment. Stanford explicitly flags this comparison as understating China’s actual spending: government guidance funds have deployed an estimated $184 billion into AI firms between 2000 and 2023, outside the private-investment tally. The U.S.–China gap in private capital is real; the U.S.–China gap in total AI resource deployment is narrower.
Gartner’s IT spending data places the United States at the largest national AI market, consistent with the U.S. accounting for approximately 38% of global AI investment in most enterprise spending surveys. Europe trails both the U.S. and China in private AI investment, despite leading in regulatory framework development. Asia Pacific — per IDC’s regional forecasts — is projected to see AI and GenAI spending reach $370 billion by 2029, growing fivefold as enterprises shift from experimentation to scaled deployment.
Frequently Asked Questions
How much is the world spending on AI in 2026?
Gartner forecasts global AI spending at $2.59 trillion in 2026, a 47% increase from 2025. This covers AI infrastructure, services, software, cybersecurity, and models. The figure was revised upward from $2.52 trillion in Gartner’s January 2026 release to reflect faster-than-expected agentic AI adoption and enterprise demand for AI models.
What percentage of revenue are companies spending on AI?
BCG’s AI Radar 2026 survey of 2,360 executives finds corporations plan to spend approximately 1.7% of revenues on AI in 2026, more than double the 0.8% in 2025. Technology companies lead at 2.1% and financial institutions at 2.0%, while industrial and real estate firms plan under 1.0%.
What is the biggest category of AI spending in 2026?
AI infrastructure — servers, AI-optimized chips, networking equipment, and cloud IaaS — accounts for approximately 55% of total AI spending in 2026, reaching $1.43 trillion per Gartner. This reflects the current phase of AI market development, where physical compute capacity remains the primary bottleneck to enterprise AI deployment.
Are companies getting ROI from AI spending?
Only 39% of organizations report any measurable EBIT impact from AI, per McKinsey’s State of AI survey of ~1,993 organizations (November 2025). Only 6% — roughly 109 organizations — qualify as “AI high performers” with AI contributing more than 5% of enterprise EBIT. The organizations achieving the strongest returns share three patterns: fundamental workflow redesign, more than 20% of digital budgets allocated to AI, and innovation-oriented AI objectives beyond pure efficiency.
What is the ASEI™ and what does a score of 58.2 mean?
The AI Spending Efficiency Index (ASEI™) is an original metric published by Axis Intelligence Research. It measures how many organizations are capturing measurable EBIT impact relative to how many are spending heavily on AI. A score of 58.2 in 2026 — down from 118.2 in 2025 — means that for every 100 organizations now spending at or above the heavy-investment threshold (≥1.7% of revenue), only 58 report any measurable enterprise EBIT impact from AI. The ASEI™ decline reflects the doubling of the heavy-spender cohort without a corresponding expansion in the EBIT-impact cohort.
How fast is AI agent software spending growing?
Gartner projects AI agent software spending will reach $206.5 billion in 2026, a 139% increase from $86.4 billion in 2025, and climb to $376.3 billion in 2027. This makes agentic AI software the fastest-growing segment in the AI spending taxonomy. However, Gartner also projects that over 40% of agentic AI projects will be canceled before reaching production by the end of 2027.
How do hyperscaler AI capex figures relate to the $2.59 trillion total?
They are sub-components, not separate markets. The four largest U.S. hyperscalers — Amazon, Microsoft, Google, and Meta — are collectively guiding to roughly $680–715 billion in capital expenditure for 2026, with the majority directed at AI infrastructure. This hyperscaler capex flows into Gartner’s AI infrastructure segment ($1.43 trillion), alongside capex from other cloud providers, hardware manufacturers, and enterprise data center operators.
Which industries are spending the most on AI?
Technology and software companies lead at approximately 2.1% of revenues, followed by financial institutions at 2.0%, per BCG AI Radar 2026. Financial services leads non-tech sectors in AI software’s share of digital transformation spending, driven by fraud detection, risk scoring, and compliance automation. Manufacturing showed the fastest year-over-year growth in AI adoption in McKinsey’s 2025 survey.
What will AI spending reach by 2027 and 2029?
Gartner projects total AI spending will reach $3.49 trillion in 2027, with AI infrastructure alone hitting $1.89 trillion. IDC’s August 2025 forecast projects AI IT spending reaching $1.3 trillion by 2029 at a 31.9% CAGR across the 2025–2029 window. Both forecasts assume continued enterprise adoption maturation and growing agentic AI deployments.
Why do different sources report such different AI spending figures?
Because they measure different things. Gartner’s $2.59 trillion measures total AI procurement across the full stack — hardware, software, services, cybersecurity, models, and data. IDC’s infrastructure tracker measures hardware shipments only ($318 billion for 2025 AI infrastructure, $487 billion projected for 2026). Stanford HAI’s $581.7 billion measures corporate AI investment — funding flows into AI companies, not what enterprises spend on AI products. These figures are orthogonal measurements that cannot be added together or used interchangeably.
Methodology
Axis Intelligence Research assembled this dataset by going to the primary issuer of every statistic and reading the original release or official press summary.
Source hierarchy used in this report:
- Tier 1 (linked directly): Official Gartner press releases (gartner.com/newsroom); BCG AI Radar 2026 publication and press release (bcg.com); McKinsey State of AI 2025 (mckinsey.com); Stanford HAI 2026 AI Index Economy chapter (hai.stanford.edu); IDC press releases (idc.com/getdoc.jsp); Earnings filings from Amazon, Microsoft, Alphabet, Meta (SEC filings)
- Tier 2 (named, not linked): CFOtech Australia and CIO Dive reporting on Gartner figures — used only to confirm specific segment-level breakdowns from the Gartner paywalled table; Value Add VC hyperscaler capex compilation; Mavvrik AI cost governance research (used only for the BCG-corroborated 80–85% forecast miss figure)
- Rejected: Statistics sourced from secondary aggregators, tech blogs, or sources without a traceable primary URL
Verification: All Gartner, BCG, McKinsey, and IDC primary URLs return 403 on direct curl from our server — consistent with standard bot-blocking behavior across all institutional research sites. Each source was confirmed live via Google Search snippet retrieval, which reflects current page status. No 404 responses were encountered.
The ASEI™ calculation: See full methodology in the ASEI™ section above. The metric depends on two survey instruments from different organizations measuring different populations. BCG’s budget data covers 2,360 executives; McKinsey’s EBIT data covers ~1,993. The assumption that EBIT impact rates are uniformly distributed across budget levels is a simplification — high-budget organizations likely show higher EBIT impact rates, which would make the ASEI™ a conservative estimate of the efficiency gap. Axis Intelligence Research errs toward conservatism.
Recency: All Gartner figures use the May 2026 release unless specifically noted as the January 2026 figure. BCG AI Radar 2026 was published January 15, 2026. McKinsey State of AI was published November 5, 2025 [older data — this is the most current McKinsey edition as of Q2 2026; the next edition is expected late 2026]. Stanford HAI 2026 AI Index was published April 13, 2026.
About This Dataset
Publisher: Axis Intelligence Research Co-author: Sarah Mitchell, AI & Machine Learning Editor Snapshot date: July 3, 2026 License: Creative Commons Attribution 4.0 International (CC BY 4.0). You may reuse and redistribute with attribution to Axis Intelligence Research and a link to axis-intelligence.com/ai-spending-statistics/. Update trigger: Next revision will be prompted by: (1) Gartner’s Q3 2026 AI spending forecast revision, expected September 2026; (2) Q3 earnings from Amazon, Microsoft, Alphabet, or Meta revealing capex divergence from 2026 guidance; (3) McKinsey State of AI 2026 publication (expected November 2026); or (4) a BCG or IDC report materially changing the enterprise budget or ROI picture. Download: A CSV of the full dataset — every metric, value, segment, date, and primary source — is available under CC BY 4.0.
Cite This Research
APA Axis Intelligence Research, & Mitchell, S. (2026, July 3). AI spending statistics 2026: $2.59 trillion, the ROI gap, and the ASEI™. Axis Intelligence. https://axis-intelligence.com/ai-spending-statistics/
MLA Axis Intelligence Research, and Sarah Mitchell. “AI Spending Statistics 2026: $2.59 Trillion, the ROI Gap, and the ASEI™.” Axis Intelligence, 3 July 2026, axis-intelligence.com/ai-spending-statistics/.
Chicago Axis Intelligence Research and Sarah Mitchell. “AI Spending Statistics 2026: $2.59 Trillion, the ROI Gap, and the ASEI™.” Axis Intelligence. July 3, 2026. https://axis-intelligence.com/ai-spending-statistics/.
Embed This Research
<blockquote>
<p>"Global AI spending reaches $2.59 trillion in 2026 (+47% YoY), with infrastructure alone accounting for $1.43 trillion (55%). Axis Intelligence Research's AI Spending Efficiency Index (ASEI™) fell from 118.2 in 2025 to 58.2 in 2026 — the number of heavy AI spenders doubled while the share capturing measurable EBIT returns stayed flat at 39%."</p>
<footer>— <a href="https://axis-intelligence.com/ai-spending-statistics/">Axis Intelligence Research, AI Spending Statistics 2026</a>, CC BY 4.0</footer>
</blockquote>
Related research: AI Statistics 2026 · Generative AI Statistics 2026 · AI Adoption Statistics 2026 · AI in Business Statistics 2026 · AI Agents Statistics 2026 · AI Governance Statistics 2026 · EU AI Act Compliance Statistics 2026 · AI Hallucination Statistics 2026 · Cloud Computing Statistics 2026 · Semiconductor Statistics 2026 · AI Data Center Energy Statistics 2026 · Nvidia Blackwell Statistics 2026 · Claude AI Statistics 2026 · ChatGPT Statistics 2026