Contacts
1207 Delaware Avenue, Suite 1228 Wilmington, DE 19806
Contact
Business Address: 1207 Delaware Avenue, Suite 1228 Wilmington, DE 19806

AI Spending Statistics 2026–2027: Who Is Paying for the $2.67 Trillion AI Build-Out

AI spending statistics 2026 chart showing $2.67 trillion worldwide AI spending by segment, Global AI spending forecast 2027 of $3.64 trillion and Big Tech AI capex in Q3 2026

AI Spending Statistics 2026

By Axis Intelligence Research

Co-author: Sarah Mitchell | Last updated: November 7, 2026 | License: CC BY 4.0

Worldwide spending on artificial intelligence will reach $2.67 trillion in 2026, up 49.5%, and $3.64 trillion in 2027, according to Gartner’s September 2026 forecast. Axis Intelligence Research’s AI Spend Revision Index (ASRI) reads +7.9: six major 2026 AI spending estimates and plans were raised during the year, and none was cut.


Quick Answer

  • $2.67 trillion: worldwide AI spending in 2026 (Gartner, September 2026), with $1.48 trillion (55.6%) going to AI infrastructure.
  • $3.64 trillion: Gartner’s 2027 forecast, up 36.2%. In January 2026, Gartner put 2027 at $3.34 trillion: the forecast has gained $300.6 billion in eight months.
  • +7.9: the AI Spend Revision Index (ASRI), an Axis Intelligence Research measure of how far the year’s AI spending estimates and capex plans have been raised since they were first published.
  • $202.0 billion: what Amazon, Alphabet, Microsoft and Meta spent on property and equipment in a single quarter (July–September 2026), 115.1% of the cash their operations generated.

Key Findings

  1. Gartner forecasts worldwide AI spending of $2.67 trillion in 2026, up 49.5% from 2025, and $3.64 trillion in 2027 (September 16, 2026 forecast).
  2. Axis Intelligence Research’s AI Spend Revision Index (ASRI) stands at +7.9 as of October 29, 2026: all six tracked 2026 AI spending estimates and capex plans were revised up, none down.
  3. Axis Intelligence Research finds that Amazon, Alphabet, Microsoft and Meta spent $202.0 billion on property and equipment in Q3 2026, equal to 115.1% of their combined operating cash flow.
  4. Axis Intelligence Research finds that 11.2% of monthly AI users worldwide pay for an AI subscription, based on GWI’s Q2 2026 survey (9.5% of internet users pay; 85.2% use AI tools).
  5. 55.9% of U.S. businesses on Ramp’s platform paid for an AI product in September 2026, up 12.1 percentage points in a year, per the Ramp AI Index.

Reading guide. “AI spending” in this report means money paid for AI products and capacity: hardware, cloud, software, models and services. It is not the same as “AI market size” (vendor revenue) or “AI investment” (venture, private equity and M&A). Each table names the measure it uses.

How Much Is the World Spending on AI in 2026?

Gartner forecasts $2.67 trillion of worldwide AI spending in 2026, a 49.5% increase on 2025’s $1.79 trillion. That figure comes from Gartner’s September 16, 2026 forecast, its third 2026 reading after $2.53 trillion in January and $2.60 trillion in May.

Worldwide AI spending by segment, 2025–2027

Segment2025202620272026 growth2026 shareSource
AI infrastructure$981.9B$1,484.4B$1,977.7B51.2%55.6%Gartner, Sep 2026
AI services$434.0B$576.5B$745.7B32.8%21.6%Gartner, Sep 2026
AI software$288.2B$461.6B$656.4B60.2%17.3%Gartner, Sep 2026
AI cybersecurity$25.9B$51.3B$86.0B98.1%1.9%Gartner, Sep 2026
AI agents and assistants$16.5B$29.2B$65.5B77.3%1.1%Gartner, Sep 2026
Generative AI models$13.0B$28.3B$51.6B117.1%1.1%Gartner, Sep 2026
AI platforms for data science and machine learning$19.4B$26.4B$35.6B36.3%1.0%Gartner, Sep 2026
AI application development platforms$6.9B$9.5B$12.5B38.6%0.4%Gartner, Sep 2026
AI data$0.8B$3.1B$6.5B278.5%0.1%Gartner, Sep 2026
Total AI spending$1,786.7B$2,670.5B$3,637.3B49.5%100%Gartner, Sep 2026

Growth rates and shares are Axis Intelligence Research calculations from Gartner’s Table 1; Gartner itself states the 49.5% total. In September, Gartner separated cross-functional agents and assistants from AI software and added consumer agents and assistants, so the software line is not directly comparable with Gartner’s earlier 2026 forecasts.

Sarah Mitchell’s read: The headline number is real, but it is a forecast built mostly on hardware orders. Infrastructure alone is 55.6% of the 2026 total. The part that tells you whether AI is working inside companies, software, is growing faster (60.2%), but it is still less than a third the size of the hardware bill.

How IDC’s numbers compare

IDC measures a narrower slice: servers and storage bought for AI. IDC raised its 2026 AI infrastructure forecast to $497 billion in July 2026, after $318 billion in 2025 and $153 billion in 2024 (IDC, July 21, 2026). IDC expects that market to pass $1.08 trillion in 2029 and $1.21 trillion in 2030.

The gap between Gartner’s $1.48 trillion “AI infrastructure” and IDC’s $497 billion is definitional, not a disagreement. IDC counts AI servers and storage. Gartner’s infrastructure segment covers AI-optimized IaaS, AI-optimized servers, AI network fabric, AI processing semiconductors and devices. In Gartner’s January 2026 forecast, AI-optimized servers alone made up 17% of total AI spending. Neither number should be added to the other.

How Much Will AI Spending Reach in 2027?

Gartner forecasts $3.64 trillion of worldwide AI spending in 2027, up 36.2% on 2026. AI infrastructure is projected at $1.98 trillion, AI services at $745.7 billion and AI software at $656.4 billion (Gartner, September 2026).

Two shifts inside the 2027 forecast matter more than the total:

  • Agents grow faster than any other large segment. Gartner’s AI agents and assistants segment more than doubles, from $29.2 billion in 2026 to $65.5 billion in 2027 (+124.1%).
  • Inference overtakes training in cloud spending. Gartner’s August 2026 forecast puts AI-optimized infrastructure-as-a-service (IaaS, cloud computing capacity rented on demand) at $42.3 billion in 2026 and $66.1 billion in 2027. Inference, the cost of running models once they are trained, is $23.3 billion of the 2026 total against $19.0 billion for training, and rises to 59% of the segment in 2027.

Sarah Mitchell’s read: Training spend is a bet on the next model. Inference spend is a usage bill paid every time someone queries a model. When inference passes training, AI spending stops being a bet on the next model and starts tracking how much people actually use the current ones. That is the most useful single signal in the 2027 forecast.

Are AI Spending Forecasts Keeping Up? The AI Spend Revision Index (ASRI)

The AI Spend Revision Index (ASRI) measures how far the main 2026 AI spending estimates and corporate capex plans have moved since they were first published. Its baseline reading, as of October 29, 2026, is +7.9. A positive reading means estimates were raised; zero means they held; a negative reading means they were cut.

ASRI is an Axis Intelligence Research construction. It takes six published figures, compares the first 2026 value of each with its latest value, and averages the percentage changes with equal weights.

ASRI componentFirst estimate (date)Latest estimate (date)RevisionSource
Gartner, worldwide AI spending 2026$2,527.8B (Jan 15)$2,670.5B (Sep 16)+5.6%Gartner
Gartner, worldwide AI spending 2027$3,336.7B (Jan 15)$3,637.3B (Sep 16)+9.0%Gartner
Gartner, AI infrastructure 2026$1,366.4B (Jan 15)$1,484.4B (Sep 16)+8.6%Gartner
IDC, AI infrastructure 2026$487B (Apr 16)$497B (Jul 21)+2.1%IDC
Meta, 2026 capex outlook (midpoint)$125B (Jan 28)$140B (Oct 21)+12.0%Meta
Amazon, 2026 capex plan~$200B (Feb 5)~$220B (Oct 29)+10.0%Amazon
ASRI (equal-weighted mean)+7.9Axis Intelligence Research

Formula: ASRI = (1/6) × Σ [(latest value ÷ first value) − 1] × 100. Inputs: Meta Q4 2025 results ($115–135 billion outlook) and Meta Q3 2026 results ($135–145 billion); Amazon Q4 2025 release (“about $200 billion”) and Amazon Q3 2026 release (“approximately $220 billion”); IDC’s April and July 2026 infrastructure forecasts; Gartner’s January and September 2026 forecasts.

What +7.9 means. Across forecasters and spenders alike, the 2026 numbers published in the first weeks of the year were too low by roughly 8%, and not one of the six was revised down. The spenders moved further than the forecasters: Meta and Amazon raised their own plans by 10–12%, while the analyst forecasts moved 2–9%. When the companies writing the checks revise faster than the firms modeling them, the models are trailing the purchase orders.

Where forecasters were most wrong

Comparing Gartner’s three 2026 tables segment by segment shows what the revisions were made of (Axis Intelligence Research calculation; all tables from Gartner):

Gartner segment2026 revision, Jan → May2026 revision, Jan → Sep2027 revision, Jan → May
AI infrastructure+4.8% (+$65.1B)+8.6% (+$118.0B)+8.1% (+$142.1B)
AI application development platforms0.0%+13.4%0.0%
AI services−0.5%−2.1% (−$12.2B)−0.2%
AI platforms for data science and machine learning−3.8%−15.0%−4.1%
AI models (January and May definition)+23.6%not comparable+36.2%
Total+2.7%+5.6% (+$142.6B)+4.7%

Gartner redefined its software, agent and model segments in September; those lines are compared only where the definition held.

Sarah Mitchell’s read: The pattern is consistent. The upgrades went to chips, data centers, model APIs and tools for building custom AI applications. The downgrades hit the layer delivered by people and specialist platforms: consulting, integration and data-science tooling, which lost 15% of its 2026 forecast in eight months. Gartner’s own explanation in September is that enterprises are turning to service providers less often and leaning on the AI features of software they already own. If you sell AI services, the forecast is telling you that buyers are choosing capacity over advice.

What Is AI Money Spent On? Infrastructure vs. Software vs. Services

For every $1 spent on AI services in 2026, $2.57 goes to AI infrastructure, up from $2.26 in 2025 (Axis Intelligence Research calculation from Gartner’s September 2026 forecast: $1,484.4 billion ÷ $576.5 billion; $981.9 billion ÷ $434.0 billion).

The hardware tilt is also visible in which chips are bought. IDC counted $89.7 billion of AI infrastructure spending in Q1 2026, up 33.1% year over year, and 97.6% of it was servers. Within accelerated servers, systems built on Arm processors ($53.0 billion) outsold x86 systems ($34.6 billion) for the quarter (IDC, July 2026).

For the chip supplier view of the same money, see our Nvidia statistics and semiconductor statistics. For the physical build, see the AI data center tracker and data center construction spending.

How Much Are Microsoft, Alphabet, Amazon and Meta Spending on AI?

Amazon, Alphabet, Microsoft and Meta spent a combined $202.0 billion on property and equipment in the July–September 2026 quarter. The companies do not split out the AI share of that spending.

Big Tech capex vs. operating cash flow, Q3 2026

CompanyQ3 2026 capexQ3 2026 operating cash flowCapex as % of operating cash flowSource
Amazon$77.3B$42.4B182.1%Amazon Q3 2026 release
Alphabet$51.4B$42.3B121.4%Alphabet Q3 2026 results (SEC)
Microsoft$39.4B$58.0B67.9%Microsoft FY27 Q1
Meta$33.9B$32.8B103.5%Meta Q3 2026 results
Combined$202.0B$175.5B115.1%Axis Intelligence Research

Capex = cash purchases of property and equipment (Microsoft: “additions to property and equipment”; Microsoft’s figure including finance leases is $45.2 billion). Microsoft’s fiscal Q1 2027 covers July–September 2026.

Axis Intelligence Research finds the four companies spent $1.15 on capex for every $1.00 of operating cash flow in Q3 2026. A year earlier, Alphabet’s ratio was 49.5% and Amazon’s 98.8%; in Q3 2026 they were 121.4% and 182.1%. Alphabet reported negative free cash flow of $9.1 billion for the quarter, and Amazon’s trailing twelve-month free cash flow was an outflow of $42.4 billion.

At the Q3 pace, the four companies’ capex runs at about $808 billion a year (Axis Intelligence Research estimate: Q3 combined × 4; not company guidance). Company plans for calendar 2026: Amazon about $220 billion, Microsoft about $170 billion, and Meta $135–145 billion.

What the spending is buying, in revenue terms: Google Cloud added $15.6 billion of quarterly revenue year over year (to $30.7 billion, up 103%), AWS added $13.0 billion (to $46.0 billion) and Azure added $10.8 billion (to $33.2 billion, up 48%). Microsoft reported $677 billion of commercial remaining performance obligation (contracted revenue not yet recognized), up 73%, and Alphabet’s CEO said the Gemini app passed 1 billion monthly active users.

Sarah Mitchell’s read: A year earlier, Amazon and Alphabet still funded capex out of their own operating cash flow. In Q3 2026 three of the four crossed 100%, which means the difference has to come from cash reserves, debt or leases. That is not proof of a bubble. Contracted backlogs like Microsoft’s $677 billion are the counterweight. It does mean 2027 spending depends on credit markets in a way 2025 spending did not.

Company-by-company commitments, project by project, live in the AI capex tracker. Valuation, credit and return-on-capital data are covered in AI bubble statistics.

How Much Are Businesses Spending on AI?

Companies planned to spend about 1.7% of revenue on AI in 2026, more than double the 0.8% of 2025, according to BCG’s AI Radar 2026, a survey of 2,360 executives, including 640 CEOs, in 16 markets. Technology companies and financial institutions planned about 2%; industrial and real estate companies, under 1%. The same survey found 94% of organizations plan to keep investing in AI in 2026 even without returns.

McKinsey’s State of AI 2026 survey (1,719 participants, May 4–June 8, 2026) shows where the budget pressure is going:

  • 28% of respondents say their organizations spend more than 10% of their total information and communication technology budget on AI.
  • 60% expect their organization’s AI investment to increase over the next year; 8% expect an increase of 30% or more.
  • 20% say AI-related operating costs, including token costs, have constrained their AI use.
  • 32% say their organization decided against buying at least one software product because it could build it in-house with agentic coding tools.
  • 37% attribute at least some EBIT (earnings before interest and taxes) impact to AI, about the same share as a year earlier.

Which industries will raise AI budgets fastest in 2027?

The table pairs two McKinsey measures by industry: the share of respondents expecting AI investment to rise by 10% or more next year, and the share whose AI use is already constrained by operating costs. The budget momentum ratio (an Axis Intelligence Research calculation) divides the first by the second: above 2, budget growth is outrunning cost pressure; near 1, the two are converging.

IndustryExpect AI investment +10% or moreAI use constrained by costsBudget momentum ratioSource
Insurance55%13%4.2McKinsey; Axis calc.
Consumer goods and retail43%12%3.6McKinsey; Axis calc.
Pharma and medical products48%18%2.7McKinsey; Axis calc.
Media and telecom43%20%2.1McKinsey; Axis calc.
Healthcare payers and providers40%20%2.0McKinsey; Axis calc.
Professional services42%21%2.0McKinsey; Axis calc.
Advanced manufacturing36%18%2.0McKinsey; Axis calc.
Financial institutions45%24%1.9McKinsey; Axis calc.
Travel and logistics40%22%1.8McKinsey; Axis calc.
Technology41%25%1.6McKinsey; Axis calc.
Energy and materials27%18%1.5McKinsey; Axis calc.
Public and social sector18%12%1.5McKinsey; Axis calc.
Engineering and construction30%22%1.4McKinsey; Axis calc.
All respondents38%20%1.9McKinsey; Axis calc.

Sarah Mitchell’s read: Technology companies have the highest cost-constraint rate (25%) of any industry. They were early, they run the most tokens, and they are the first to feel the inference bill. Insurers and retailers sit at the other end: big planned increases, few cost complaints. That is where 2027 enterprise AI budgets have the most room to grow before finance teams push back.

Share of U.S. businesses paying for AI, September 2026

55.9% of U.S. businesses on Ramp paid for at least one AI product or service in September 2026, up 12.1 percentage points from a year earlier (Ramp AI Index, based on card and bill-pay transactions from more than 85,000 U.S. businesses).

SegmentShare paying for AISource
All businesses55.9%Ramp AI Index
Large businesses65.7%Ramp AI Index
Medium-sized businesses62.5%Ramp AI Index
Small businesses50.0%Ramp AI Index
Information79.9%Ramp AI Index
Finance & insurance73.7%Ramp AI Index
Manufacturing60.9%Ramp AI Index
Retail trade49.9%Ramp AI Index
Construction43.2%Ramp AI Index
Healthcare42.5%Ramp AI Index
Accommodation & food33.0%Ramp AI Index

By provider, 43.8% of Ramp businesses paid Anthropic and 40.6% paid OpenAI in September 2026; xAI stood at 4.1%. Because 43.8% + 40.6% exceeds the 55.9% paying for any AI product, Axis Intelligence Research calculates that at least 28.5% of U.S. businesses on Ramp pay both Anthropic and OpenAI.

Sarah Mitchell’s read: Budgets as a share of revenue doubled, but adoption on Ramp is still close to a coin flip, and it barely moved between August (55.2%) and September (55.9%). So the budget growth has to come mostly from companies that already pay for AI paying for more of it, which fits the inference story above. Multi-vendor buying, with at least 28.5% paying two model providers, points the same way. For one of the fastest-growing line items, see AI coding assistant statistics.

How Many People Pay for AI Tools?

9.5% of internet users aged 16 and over worldwide pay for an AI tool subscription each month, while 85.2% use AI tools, according to GWI’s Q2 2026 survey. Axis Intelligence Research calculates a worldwide AI paid conversion rate of 11.2%: roughly one paying subscriber for every nine monthly users.

The paid conversion rate divides the share of internet users who pay for an AI subscription by the share who use AI tools, both from the same GWI survey wave, so the two inputs share one method and one sample.

AI paid conversion rate by country, Q2 2026

CountryUse AI tools monthlyPay for an AI subscriptionPaid conversion rate (Axis)Source
Kenya97.3%18.5%19.0%GWI
Israel94.0%14.1%15.0%GWI
China89.4%12.8%14.3%GWI
United Arab Emirates95.7%13.3%13.9%GWI
India84.8%11.4%13.4%GWI
Saudi Arabia93.3%11.8%12.6%GWI
Worldwide85.2%9.5%11.2%GWI
Brazil93.4%9.9%10.6%GWI
United States74.6%6.8%9.1%GWI
Australia72.0%6.5%9.0%GWI
South Korea82.7%6.4%7.7%GWI
Canada72.1%5.3%7.4%GWI
United Kingdom68.0%4.9%7.2%GWI
France67.4%4.4%6.5%GWI
Germany75.4%4.8%6.4%GWI
Italy76.1%4.0%5.3%GWI
Japan58.6%2.0%3.4%GWI

All 53 markets in GWI’s survey are in the CSV. Ghana shows the highest ratio (24.3%) because only 35.8% of its internet users report using AI tools.

Paying skews young and male. Among internet users aged 25–34, 12.8% of men and 10.7% of women pay for an AI subscription; among those aged 55–64, 7.2% and 6.5%.

Sarah Mitchell’s read: The rich-country assumption fails here. A Kenyan AI user is more than twice as likely to pay as a German one, and India converts better than the United States. One reading that fits the table: where AI tools replace help that is expensive or hard to find, people pay; where AI arrives bundled into phones and office software they already own, it stays free. Japan, at 3.4%, is the market where that free tier wins most clearly. For usage beyond payment, see our generative AI statistics.

Which Countries Spend the Most on AI?

The United States accounted for 75.7% of global AI infrastructure spending in Q1 2026 ($67.9 billion of $89.7 billion), against 8.7% for China ($7.8 billion), according to IDC. Western Europe spent $5.1 billion; Middle East and Africa spending rose 233% year over year to $1.1 billion.

Investment tells the same story. U.S. AI investment reached $285.88 billion in 2025, against $20.92 billion for Europe and $12.41 billion for China, according to Stanford HAI’s 2026 AI Index. Axis Intelligence Research calculates that the United States invested 23.0 times as much as China and 13.7 times as much as Europe. Total global corporate AI investment reached $581.69 billion, of which $344.66 billion was private investment and $214.44 billion mergers and acquisitions.

RegionAI infrastructure spend, Q1 2026 (IDC)AI investment, 2025 (Stanford HAI)
United States$67.9B (75.7%)$285.88B
China$7.8B (8.7%)$12.41B
Western Europe / Europe$5.1B$20.92B
Asia/Pacific excl. Japan and China$5.8B—
Middle East and Africa$1.1B—

Sarah Mitchell’s read: Read these as measures of commercial hardware purchases and private capital, which is what they are built to track; state-directed spending is outside their scope. Even so, the consumer table above shows Chinese AI users paying at a higher rate (14.3% conversion) than Americans (9.1%). The country gap is about infrastructure capital, not about demand.

Frequently Asked Questions About AI Spending

Is $2.67 trillion the size of the AI market?

No. Gartner’s $2.67 trillion measures spending on AI-related products and capacity in 2026, including AI-optimized servers, chips, devices and services. “AI market size” figures usually count vendor revenue from AI software alone and come out far smaller. Quote the measure with the number.

Why are Gartner’s and IDC’s AI spending numbers so different?

They measure different things. Gartner’s 2026 figure ($2.67 trillion) covers infrastructure, software, services, models and security. IDC’s $497 billion covers AI servers and storage only. Both were raised during 2026.

How much of AI spending goes to data centers and hardware?

About 56% in 2026. Gartner puts AI infrastructure at $1.48 trillion of $2.67 trillion. Axis Intelligence Research calculates that infrastructure spending is 2.57 times AI services spending, up from 2.26 times in 2025.

Is more money now spent running AI models than training them?

In rented cloud capacity, yes. Gartner forecasts $23.3 billion of inference spending on AI-optimized IaaS in 2026 against $19.0 billion for training, with inference reaching 59% of the segment in 2027 (Gartner, August 2026).

How much should a company budget for AI as a share of revenue?

BCG’s AI Radar 2026 found companies planned about 1.7% of revenue for AI in 2026, up from 0.8% in 2025. Technology and financial companies planned about 2%, industrials and real estate under 1%. In McKinsey’s 2026 survey, 28% of respondents spend more than 10% of their technology budget on AI.

Are AI costs starting to limit how much companies use AI?

For one company in five. In McKinsey’s State of AI 2026 survey, 20% of respondents said AI operating costs, including token costs, constrained their AI use, rising to 25% in technology companies. Even so, 60% expect their AI investment to increase next year.

Does all Big Tech capex go to AI?

None of the four companies reports an AI-only capex figure. The $202.0 billion spent by Amazon, Alphabet, Microsoft and Meta in Q3 2026 also pays for warehouses, offices and general-purpose servers, so treat it as an upper bound on Big Tech’s AI capex, not a measure of it.

Which AI provider do U.S. businesses pay most often?

Anthropic. In September 2026, 43.8% of U.S. businesses on Ramp paid Anthropic and 40.6% paid OpenAI, per the Ramp AI Index. By Axis Intelligence Research’s calculation, at least 28.5% pay both.

What would show AI spending slowing in 2027?

Three data points would come first: a negative ASRI reading (forecasts or capex plans cut), Big Tech 2027 capex guidance below 2026 levels in the Q4 2026 results, and inference spending growth falling below infrastructure growth. As of November 7, 2026, none of the three has happened.

Methodology

Scope. This report covers AI spending (money paid for AI hardware, cloud capacity, software, models and services), corporate capex by the four largest U.S. cloud and platform companies, enterprise AI budgets, U.S. business AI purchasing, consumer AI subscriptions and AI investment by country. Figures from different producers are shown side by side and never summed.

Sources. Every figure was taken from the organization that produced it and checked on November 7, 2026: Gartner (forecasts of January, May, August and September 2026), IDC (AI infrastructure tracker and forecasts, April and July 2026), company earnings releases and filings (Amazon, Alphabet, Meta, Microsoft), BCG (AI Radar 2026), McKinsey (State of AI 2026), Ramp (AI Index, September 2026), Stanford HAI (2026 AI Index) and GWI (Q2 2026 survey wave, published in the Digital 2026 October Global Statshot). Vendor research is used with its incentive in view: Gartner and IDC sell advisory services to the companies whose spending they forecast; Ramp sells spend-management software.

AI Spend Revision Index (ASRI). ASRI = (1/6) × Σ [(latest published value ÷ first published value for 2026 or 2027) − 1] × 100, across six components: Gartner total AI spending 2026, Gartner total AI spending 2027, Gartner AI infrastructure 2026, IDC AI infrastructure 2026, Meta 2026 capex outlook midpoint and Amazon 2026 capex plan. Weights are equal. The baseline reading is +7.9 as of October 29, 2026. Future readings use the same components and weights; the 2027 cycle starts with the first 2027 estimates published in early 2027.

Capex-to-operating-cash-flow ratio. Cash purchases of property and equipment ÷ net cash from operating activities, calendar Q3 2026 (July–September), × 100. Microsoft’s fiscal Q1 2027 covers the same months.

Budget momentum ratio. Share of McKinsey respondents in an industry expecting AI investment to rise 10% or more next year (Exhibit 9: “≥30%” plus “10–29%”) ÷ share reporting AI use constrained by operating costs (Exhibit 8). Both measures come from the same McKinsey survey.

AI paid conversion rate. Share of internet users aged 16+ paying for an AI tool subscription each month ÷ share using AI tools each month × 100, both from GWI’s Q2 2026 wave. GWI removed Russia from its survey from Q1 2026, so worldwide values are not comparable with GWI figures published before 2026.

About This Dataset

The CSV ships 538 rows in long format: one row per observation, with metric, value, unit, as-of date, geography, segment, entity, source organization, source document, source URL, retrieval date, a primary-source flag, an Axis-calculation flag and the formula for every calculated row. Every number on this page is a row in the file. The dataset is licensed under CC BY 4.0: reuse it with credit to “Axis Intelligence Research, AI Spending Statistics 2026–2027.” It is updated when a source publishes new data, not on a fixed calendar.

Cite This Research

APA: Axis Intelligence Research. (2026, November 7). AI spending statistics 2026–2027: Who is paying for the $2.67 trillion AI build-out. Axis Intelligence. https://axis-intelligence.com/ai-spending-statistics/

MLA: Axis Intelligence Research. “AI Spending Statistics 2026–2027: Who Is Paying for the $2.67 Trillion AI Build-Out.” Axis Intelligence, 7 Nov. 2026, axis-intelligence.com/ai-spending-statistics/.

Chicago: Axis Intelligence Research. “AI Spending Statistics 2026–2027: Who Is Paying for the $2.67 Trillion AI Build-Out.” Axis Intelligence, November 7, 2026. https://axis-intelligence.com/ai-spending-statistics/.

[Copy-to-clipboard buttons for each format]

Related research: AI statistics · AI capex tracker · AI bubble statistics · Cloud market share statistics · AI IPO tracker

Recent Posts

Stablecoin Statistics 2026-2027: Market Cap, Supply by Year and the 8.8% Drawdown From the May Peak

Stablecoin Statistics 2026-2027 By Axis Intelligence Research Co-author: Sarah Davis | Last updated: November 20, 2026 |

Stablecoin Payment Statistics 2026-2027: $482 Billion Paid in Nine Months on a Float That Barely Grew

Stablecoin Payment Statistics 2026-2027 By Axis Intelligence Research Co-author: Sarah Davis | Last updated: November 20

PC Gaming Statistics 2026–2027: Players, Prices & Hardware

PC Gaming Statistics 2026-2027 By Axis Intelligence Research Co-author: Noah Mc (gaming) | Last updated: November 19, 20

Axis Intelligence Research

Stay ahead on tech & data

Get notified when we publish or update datasets, trackers, research, and reports across technology, business, AI, cybersecurity, finance, infrastructure, energy, and more.

Research updates only. No spam. Unsubscribe anytime.