AI Absorption Ledger / WMT

Walmart

WMT · Q3 2026 · reported 2026-08-20 · revenue $187.94bn

Assessment

AI disclosure thins. The call's AI passages come down to a closing section of the CEO's remarks and one line from the CFO; the release, the presentation and the 10-Q have none beyond a forward-looking factor. Every spend channel goes unmentioned, as do AI in inventory decisions, AI in advertising and orders through outside AI platforms.

What remains is Sparky and associate work. Customers using Sparky are up on a year earlier and spend more per order than non-users, still a comparison of different customers; the ledger's estimate is . The CEO says AI makes associates' work easier, and the CFO says wages leveraged on associates' use of technology tools and on supply chain automation, without naming AI; labor saved is left unsized this quarter, since nothing ties a cost movement to AI.

Nothing in the quarter's filings attributes a dollar to AI. The 10-Q explains the higher expense rate by liability claims, depreciation and healthcare, and the gross profit rate by tariff refunds and advertising growth; advertising grew with no AI named.

Sized channels against the income statement, Q3 2026

1 of 9 channels sized

Each blue mark is one channel's dollars for the quarter; a bar is the range of an estimate. Grey marks are the company's reported lines. The distance between them is the point: how large the AI channel is next to the line it sits in.

New money and old money, Q3 2026

3 new4 expanded2 relabelled

Each channel is tagged once for whether its money existed before language models, from the company's annual report and call at the start of the period. A bar splits one flow's sized dollars by that tag. The incremental total is the part that would not be there without the models: a new channel counts in full, an expanded one only for what AI added, a relabelled one at zero.

Revenue arriving through AI$29mn to $794mn sized

new not sizedexpanded $29mn to $794mnrelabelled not sized

Incremental total $29mn to $794mnpoint $353mn

The sized total counts every channel the company credits to AI, including relabelled money that existed before language models and the ledger's own estimates for it. The incremental total counts relabelled channels at zero. A flow is split by layer where its dollars sit at more than one: end use, compute sold to builders, and hardware. The same dollar can be a buyer's spend, a cloud's revenue and a chipmaker's revenue, so the layers are never added together.

Paid for AI2 channels · 2 not sized

vendor bill

AI model, partnership and tooling bill

Not sized

No source in the quarter mentions the AI bill.

inscrutabledisclosure: not mentioned· motive: exploratory· before LLMs: new

The quarter names no AI vendor, partner or investment. The channel is carried with no reading of its own. The counterparty is mixed: model partners, vendors of AI tools for associates, and the software and cloud providers whose AI share sits in the bill.

Evidence: 0 quotes, 5 from before coverage

What the company pays outside technology companies for AI: the model partners behind Sparky and agentic commerce, AI tools for associates and the AI share of software and cloud services. Management says it approaches AI development through partnerships. The bill sits in operating, selling, general and administrative expenses and is not split.

Why this motive

Carried from the prior quarter; no AI spend or partnership is mentioned.

Before LLMs: new

At the anchor some technology systems and services were already provided and managed by outside providers, inside operating, selling, general and administrative expenses ( for fiscal 2024), with no AI vendor named; a model and token bill did not exist before LLMs. In the quarter before coverage the CEO named ChatGPT Enterprise licences being rolled out to associates and a partnership with OpenAI.

“Moreover, some of the various technology systems and services on which we rely are provided and managed by third-party service providers.”
Filing, risk factors, 10-K periodic report, 2024-03-15
“Our recent announcement with OpenAI is an example. This new partnership will allow customers and members to purchase items from Walmart and Sam's Club directly through ChatGPT.”
CEO, prepared remarks, earnings call, 2025-11-20
“As we think about new Tech Products and Capabilities, sometimes we build our own tech, and sometimes we partner.”
CEO, prepared remarks, earnings call, 2025-11-20
“We're helping our associates build the skills they'll need to thrive in an AI-Powered workplace through things like embracing OpenAI Certifications and rolling out ChatGPT Enterprise Licenses.”
CEO, prepared remarks, earnings call, 2025-11-20
“We're using AI across the organization to manage costs effectively and to accelerate our growth.”
CFO, prepared remarks, earnings call, 2025-11-20

By quarter

  • Q1 2026described · our inference · exploratory · $11mn to $245mn
  • Q2 2026described · our inference · exploratory · $11mn to $238mn
  • Q3 2026not mentioned · inscrutable · exploratory

engineering

AI inside capital spending on technology

Not sized

No AI statement about capital spending this quarter.

inscrutabledisclosure: not mentioned· motive: exploratory· before LLMs: expanded

The CFO raises expected capital spending for the year and ties depreciation to supply chain automation and speed, without naming AI. The channel is carried with no reading of its own.

Evidence: 0 quotes, 2 from before coverage

The AI part of capital spending on supply chain, customer-facing initiatives and technology: hardware, capitalized software and AI built into the store and supply chain programs. The CFO says investments in AI are incorporated in the capital spending assumptions and does not size them. Capitalized spending is traced and shown and left out of flow totals (methodology, capital spending); it reaches the income statement as depreciation, which the filings name among the causes of higher operating expenses. Supply chain automation is in the same line and is not AI unless management names AI as its cause.

Why this motive

Carried from the prior quarter.

Before LLMs: expanded

At the anchor the fiscal 2024 10-K already named the use of artificial intelligence technology among the company's strategic investments, and capital spending on supply chain, customer-facing initiatives and technology was for the year; no AI share was given then or since, so the level has no traced baseline. From fiscal 2026 the row's label adds 'and other', a widening of the line. The size is the whole of an activity that existed before.

“Our strategy, which includes investments in eCommerce, technology, including the use of artificial intelligence technology, talent, supply chain automation, acquisitions, joint ventures, store remodels and other customer initiatives may not adequately or effectively allow us to continue to grow our eCommerce business”
Filing, risk factors, 10-K periodic report, 2024-03-15
“As we think about new Tech Products and Capabilities, sometimes we build our own tech, and sometimes we partner.”
CEO, prepared remarks, earnings call, 2025-11-20

By quarter

  • Q1 2026described · our inference · exploratory · $52mn to $518mn
  • Q2 2026described · our inference · exploratory · $38mn to $382mn
  • Q3 2026not mentioned · inscrutable · exploratory

Cost displaced by AI3 channels · 3 not sized

operations · cheap to verify

Labor displaced or avoided by AI tools for associates

Not sized

The quarter's only AI words are the CEO's aim of using AI to make associates' work easier, with no tool, deployment or measure (claim c1), and the measured wage leverage is credited to associates' use of technology tools and to supply chain automation without naming AI (claim c4). Nothing ties a Q3 cost movement to AI, so the channel gets no ballpark; the Q1 decomposition is not rolled forward (the methodology rule for AI named beside another cause, read at each quarter's own words).

described, no sizedisclosure: described· motive: exploratory· before LLMs: expanded

The CEO says the company uses AI to make associates' work easier. The CFO says wages leveraged on associates' use of technology tools in stores and on supply chain automation, more than offset by depreciation and self-insurance, and the 10-Q explains the expense rate by claims, depreciation and healthcare; operating expenses rose against revenue growth of . Lower attrition is raising health plan enrollment, a sign the workforce is not shrinking. The channel is unsized this quarter: the step comes from the wording, which names no AI tool or measure, not from a change at the company. The former estimate, , repeated the Q1 decomposition and is no longer the size.

Evidence: 6 quotes, 6 figures, 4 confounds, 6 from before coverage

Payroll that AI tools for store, club, supply chain and office associates may save: the automation of repetitive tasks the 10-K names, AI-powered tools for associate productivity and the associates' agentic tools management describes. Labor is among the company's largest costs and sits in operating, selling, general and administrative expenses. Separate from supply chain automation (robotics and automated distribution centers), which the company credits for labor productivity without naming AI and which is read as a confound here.

Why this motive

AI named with no dollar: the CEO’s AI statement is aspirational and carries no number (claim c1), and no displaced line is shown to shrink because of AI. The CFO credits wage leverage to associates’ use of technology tools and to supply chain automation without naming AI (claim c4), so it is a confound on this channel, not a motive tell. AI named as a cause with no measure and no line moving is the exploratory tell; the quarter quotes no narrative-defensive tell of its own.

Before LLMs: expanded

At the anchor associates were already described as powered by technology, of them, and the company invested in technology and process improvements to raise productivity; labor sat inside operating, selling, general and administrative expenses, for fiscal 2024. In the quarter before coverage the CFO called technology and AI enablers of efficiency gains as operating expenses leveraged, with no amount. AI tools change output per associate, not the existence of the line. The size is the change AI made, not the whole line.

“Our associates – powered by technology – play a critical role in delivering on our purpose to help people save money and live better.”
Filing, business, 10-K periodic report, 2024-03-15
“We invest in technology and process improvements to increase productivity, manage inventory and reduce costs and we operate with discipline by managing expenses and optimizing the efficiency of how we work.”
Filing, mdna, 10-K periodic report, 2024-03-15
“We continue to invest in supply chain automation and utilize a total of 162 distribution facilities which are located strategically throughout the U.S.”
Filing, business, 10-K periodic report, 2024-03-15
“We're helping our associates build the skills they'll need to thrive in an AI-Powered workplace through things like embracing OpenAI Certifications and rolling out ChatGPT Enterprise Licenses.”
CEO, prepared remarks, earnings call, 2025-11-20
“Adjusted SG&A Expenses leveraged slightly in Q3. Expenses are being well managed across the business. Technology and AI have been enablers of efficiency gains.”
CFO, prepared remarks, earnings call, 2025-11-20
“We're using AI across the organization to manage costs effectively and to accelerate our growth.”
CFO, prepared remarks, earnings call, 2025-11-20

Figures

  • Operating, selling, general and administrative expenses as a share of total revenues · 2026-CQ3
  • Operating, selling, general and administrative expenses as a share of total revenues, prior year · 2025-CQ3
  • Operating, selling, general and administrative expenses growth, year over year · 2026-CQ3
  • Total revenues growth, year over year · 2026-CQ3
  • U.S. stores served with some level of automated freight · as-of 2026-07-31
  • Share of eCommerce fulfillment volume processed through automated facilities, floor · 2026-CQ3

What else could explain it

  • transformation program: Supply chain automation, now reaching stores, is credited in the same sentence as the technology tools.
  • other: The wage leverage credited to technology tools names no AI and is not counted as evidence of an AI saving.
  • operating leverage: Wages growing slower than sales is ordinary leverage for a retailer growing transactions.
  • line composition: Depreciation, claims and healthcare move the operating expense line more than any labor saving.

Quotes

“We continue to take a people-led, tech-powered approach. We are using AI to make our work easier and help our associates grow and be at their very best. We believe AI will improve nearly every part of our business by making shopping better and our associates' work easier.”
c1 · CEO, prepared remarks, earnings call, 2026-08-20
“We leveraged wages in Q2 as we continue to improve productivity through increased usage of tech tools by associates in stores and streamlined inventory flow enabled by supply chain automation. More than offsetting these benefits were higher depreciation related to CapEx and increased self-insurance cost.”
c4 · CFO, prepared remarks, earnings call, 2026-08-20
“3,100 of our U.S. stores are now served with some level of automated freight, and we're processing over 50% of our e-commerce fulfillment volume through automated facilities.”
c6 · CFO, prepared remarks, earnings call, 2026-08-20
“If you take the first half of the year, about two-thirds of the increase is from group health. What's happening there is our attrition has gone down quite appreciably in some cases.”
c7 · CFO, qa, earnings call, 2026-08-20
“Partially offset by leverage from increased labor productivity”
c8 · Filing, presentation, 8-K earnings release, 2026-08-20
“The increase for the three months ended July 31, 2026 was primarily driven by higher self-insured general liability claims expense influenced by rising costs to resolve claims across retail and related industries, increased depreciation related to our capital investments and higher associate healthcare benefit costs related to increased enrollment and medical cost inflation.”
c9 · Filing, mdna, 10-Q periodic report, 2026-08-28

By quarter

  • Q1 2026described · our inference · narrative-defensive · $6.9mn to $230mn
  • Q2 2026not mentioned · inscrutable · narrative-defensive
  • Q3 2026described · described, no size · exploratory

operations · cheap to verify

Inventory and fulfillment decisions made with AI

Not sized

No source in the quarter attributes an inventory or fulfillment movement to AI.

inscrutabledisclosure: not mentioned· motive: exploratory· before LLMs: relabelled

Management credits supply chain automation and the optimization of inventory across fulfillment nodes without naming AI. The channel is carried with no reading of its own.

Evidence: 0 quotes, 4 from before coverage

Markdowns, spoilage and fulfillment cost avoided because AI and data help position inventory, choose fulfillment nodes and map store inventory (computer vision on associates' handheld devices). The effect would show in cost of sales and in fulfillment cost; management credits it jointly to technology, AI and supply chain automation and gives no split.

Why this motive

Carried from the prior quarter, read as exploratory; inventory and fulfillment are discussed without AI, and silence is not evidence about motive.

Before LLMs: relabelled

At the anchor the company already invested in technology and process improvements to manage inventory, named artificial intelligence technology among its strategic investments, and was building what the CEO called a more connected, intelligent and automated network. No line, price or volume is shown to move because of AI as distinct from automation, so the activity is read as renamed.

“We invest in technology and process improvements to increase productivity, manage inventory and reduce costs and we operate with discipline by managing expenses and optimizing the efficiency of how we work.”
Filing, mdna, 10-K periodic report, 2024-03-15
“Our strategy, which includes investments in eCommerce, technology, including the use of artificial intelligence technology, talent, supply chain automation, acquisitions, joint ventures, store remodels and other customer initiatives may not adequately or effectively allow us to continue to grow our eCommerce business”
Filing, risk factors, 10-K periodic report, 2024-03-15
“We continue to invest in supply chain automation and utilize a total of 162 distribution facilities which are located strategically throughout the U.S.”
Filing, business, 10-K periodic report, 2024-03-15
“We're building a more connected, intelligent, and automated network.”
CEO, prepared remarks, earnings call, 2023-05-18

By quarter

  • Q1 2026described · described, no size · exploratory
  • Q2 2026described · described, no size · exploratory
  • Q3 2026not mentioned · inscrutable · exploratory

engineering · cheap to verify

Engineering work done with AI coding tools

Not sized

Not mentioned this quarter; no engineering cost line is reported.

inscrutabledisclosure: not mentioned· motive: exploratory· before LLMs: expanded

In the quarter before coverage the CEO said more than of new code was AI-generated or AI-assisted; this quarter's sources do not mention coding or engineering. The channel is carried with no reading of its own.

Evidence: 0 quotes, 2 from before coverage

Software engineering payroll displaced or avoided because a large share of new code is AI-generated or AI-assisted. First evidenced before coverage, in the reference quarter, where the CEO gave the share; no covered quarter mentions it, and firstSeen names the first covered quarter only because the schema has no value for a reference quarter. The tool licences themselves sit in the AI vendor bill.

Why this motive

No motive tell in the covered sources; the reference quarter gives a usage share and licences, not a cost movement. AI named with no measure of cost and no line moving reads exploratory, and nothing in the sources contradicts it; carried.

Before LLMs: expanded

The anchor shows no engineering cost line; technology staff sit inside operating, selling, general and administrative expenses ( for fiscal 2024). In the quarter before coverage the CEO said more than of new code was AI-generated or AI-assisted and that ChatGPT Enterprise licences were being rolled out. AI changes output per engineer, not the existence of the line. The size is the change AI made, not the whole line.

“When AI is used for Software Development, more than 40% of the new code is either AI-Generated or AI-Assisted.”
CEO, prepared remarks, earnings call, 2025-11-20
“We're helping our associates build the skills they'll need to thrive in an AI-Powered workplace through things like embracing OpenAI Certifications and rolling out ChatGPT Enterprise Licenses.”
CEO, prepared remarks, earnings call, 2025-11-20

By quarter

  • Q1 2026not mentioned · inscrutable · exploratory
  • Q2 2026not mentioned · inscrutable · exploratory
  • Q3 2026not mentioned · inscrutable · exploratory

Revenue arriving through AI3 channels · $29mn to $794mn sized · $29mn to $794mn incremental · 2 not sized

search discovery · cheap to verify

Sales gained through Sparky, the company's AI shopping agent

0.02% to 0.42% of the quarter’s revenue

Incremental total: counts in full.

our inferencedisclosure: direction only· motive: exploratory· before LLMs: expanded

Customers using Sparky are up on a year earlier and spend more per order than non-users; the gap is wider than the given in the prior two quarters and is now stated per order rather than as average order value. Walmart U.S. eCommerce net sales were , up . The size shown is the ledger's own estimate, , on the prior quarter's assumptions.

Evidence: 4 quotes, 4 figures, 3 confounds, 3 from before coverage

Sales gained because customers shop with Sparky, the company's own AI shopping agent in its app, on its website and in stores, which builds baskets, plans meals and reorders items. The agent carries no price; the money is the increment in eCommerce and store sales it causes, not the sales of the customers who use it.

Why this motive

Carried: the CEO gives user growth and the spend-per-order gap (claim c2) and the CFO speaks of leading in agentic experiences (claim c5), with no lift measured and no sales attributed.

Before LLMs: expanded

At the anchor customers found products through search on the company's own site and app, which the head of Walmart U.S. credited with helping advertising grow; Walmart U.S. eCommerce net sales were for fiscal 2024, with no shopping assistant; by the quarter before coverage Sparky was live in the app. Sparky changes how existing digital orders are built, so the AI part is the increment it causes, not the line. The size is the change AI made, not the whole line.

“Increasingly, customers are using computers, tablets and smart phones to shop with us and with our competitors and to do comparison shopping.”
Filing, risk factors, 10-K periodic report, 2024-03-15
“What's driving it, of course, over time will be better, a stronger, bigger marketplace. More marketplace and helping them connect to customers and then more assortment that's easier to find with search also helps the advertising business grow.”
Executive, qa, earnings call, 2023-05-18
“We are really excited about some of the capabilities that our customers are engaging in with what we call Sparky today, our digital agent that is live in the app.”
Executive, qa, earnings call, 2025-11-20

Figures

  • Growth in the number of customers using Sparky, year over year · 2026-CQ3
  • Spend per order of Sparky users above non-users · 2026-CQ3
  • Walmart U.S. eCommerce net sales · 2026-CQ3
  • Walmart U.S. eCommerce net sales growth, year over year · 2026-CQ3

What else could explain it

  • mix shift: The spend gap compares adopters with non-adopters.
  • bundling: Sparky is part of the app at no price.
  • other: The comparison is restated per order this quarter; the ledger does not treat the wider gap as a change in effect.

Quotes

“We continue to take a people-led, tech-powered approach. We are using AI to make our work easier and help our associates grow and be at their very best. We believe AI will improve nearly every part of our business by making shopping better and our associates' work easier.”
c1 · CEO, prepared remarks, earnings call, 2026-08-20
“Sparky is a great example. The number of customers using Sparky is up 70% from last year, and the customers and members who use Sparky for shopping spend 40% more per order than others who do not.”
c2 · CEO, prepared remarks, earnings call, 2026-08-20
“Sparky even recognized the ingredients they had recently purchased, both online and in-store, so they did not buy something they already had. It is building trust.”
c3 · CEO, prepared remarks, earnings call, 2026-08-20
“In closing, our teams continue to focus on what we do best, serving customers and members with everyday great value, exceptional convenience, and speed, all while pushing our business model forward, diversifying our profit mix, and leading in agentic experiences.”
c5 · CFO, prepared remarks, earnings call, 2026-08-20

By quarter

  • Q1 2026direction only · our inference · exploratory · $30mn to $710mn
  • Q2 2026direction only · our inference · exploratory · $27mn to $732mn
  • Q3 2026direction only · our inference · exploratory · $29mn to $794mn

distribution

Orders arriving through third-party AI platforms

Not sized

Not mentioned this quarter, and never given a volume, share or revenue.

inscrutabledisclosure: not mentioned· motive: channel-defensive· before LLMs: new

The CFO's closing reference to leading in agentic experiences names no partner or outside platform and is read under Sparky. The channel is carried with no reading of its own.

Evidence: 0 quotes, 2 from before coverage

Orders that customers place through outside AI assistants, built with the partners management names (OpenAI and Alphabet), rather than on the company's own site or app. Management describes the partnerships and gives no volume, share or revenue.

Why this motive

Carried from the prior quarter; outside AI platforms and the partnerships are not mentioned.

Before LLMs: new

AI platforms as a place to shop are not in the anchor: the fiscal 2024 10-K describes customers shopping and comparing prices on computers, tablets and smart phones. Such orders would arrive inside eCommerce net sales (Walmart U.S. for fiscal 2024), with no part from AI platforms then. In the quarter before coverage the CEO announced a partnership letting customers purchase from Walmart and Sam's Club directly through ChatGPT, starting with checkout, with no volume given.

“Increasingly, customers are using computers, tablets and smart phones to shop with us and with our competitors and to do comparison shopping.”
Filing, risk factors, 10-K periodic report, 2024-03-15
“Our recent announcement with OpenAI is an example. This new partnership will allow customers and members to purchase items from Walmart and Sam's Club directly through ChatGPT.”
CEO, prepared remarks, earnings call, 2025-11-20

By quarter

  • Q1 2026described · our inference · channel-defensive · $30mn to $148mn
  • Q2 2026not mentioned · inscrutable · channel-defensive
  • Q3 2026not mentioned · inscrutable · channel-defensive

marketing · cheap to verify

AI features in advertising tools

Not sized

No AI feature in advertising is mentioned this quarter.

inscrutabledisclosure: not mentioned· motive: product-defensive· before LLMs: relabelled

Global advertising grew , credited to Walmart Connect and Flipkart Ads, and the company bought Vibe for self-service advertising tools; neither is tied to AI. The channel is carried with no reading of its own.

Evidence: 0 quotes, 1 confound, 5 from before coverage

Advertising revenue gained because AI features in the tools sold to advertisers, such as dynamic adjustment of content mix to optimize campaigns, lift what suppliers and marketplace sellers spend. Advertising is recorded in net sales or as a reduction of cost of sales; the AI part is not split. The acquisition of Vibe, a self-service advertising business, is an acquisition confound and not part of this channel: management does not describe it as an AI company.

Why this motive

Carried from the prior quarter.

Before LLMs: relabelled

At the anchor the advertising tools already targeted groups of customers, ran an auction and attributed in-club sales to digital ad spend, and the anchor call answered a question about AI in advertising with that targeting; global advertising grew in that quarter. The covered quarters name AI features without measuring a lift from them.

“What's driving it, of course, over time will be better, a stronger, bigger marketplace. More marketplace and helping them connect to customers and then more assortment that's easier to find with search also helps the advertising business grow.”
Executive, qa, earnings call, 2023-05-18
“There's been a considerable momentum really that started last year when we launched our second place auction capability.”
Executive, qa, earnings call, 2023-05-18
“Now, this ability that the team has developed for sellers and suppliers to reach groups of customers that are targeted, it's really improving, and I think that's definitely driving the results there.”
Executive, qa, earnings call, 2023-05-18
“Advertisers are responding to our recently launched in-club sales attribution feature, which provides advertisers with clear insights on the returns of digital ad spend both online and in clubs while enhancing member experience.”
CFO, prepared remarks, earnings call, 2023-05-18
“Our global advertising business delivered strong growth of over 30% in Q1.”
CFO, prepared remarks, earnings call, 2023-05-18

What else could explain it

  • acquisition: Vibe enters the advertising base; management does not describe it as an AI business.

By quarter

  • Q2 2026described · our inference · product-defensive · $6.4mn to $96mn
  • Q3 2026not mentioned · inscrutable · product-defensive

Cost imposed, or revenue lost, by others’ AI1 channel · 1 not sized

distribution

Store traffic and sales lost to AI shopping platforms

Not sized

Not mentioned this quarter; lost demand has no line.

inscrutabledisclosure: not mentioned· motive: imposed· before LLMs: new

The 10-Q's only AI passage is the forward-looking factor on disruptions from the deployment of AI; the call does not raise the risk.

Evidence: 0 quotes, 1 from before coverage

Store and club traffic, and the cross-selling that comes with it, that may be lost as AI-enabled platforms take over product search, discovery, advertising and purchasing, together with competition from rivals' agentic shopping tools. The 10-K names the risk; there is no line and no figure.

Why this motive

Carried: toll channel, the company did not choose the risk.

Before LLMs: new

The fiscal 2024 10-K listed competitors in eCommerce, digital advertising, fulfillment and other services and named no AI platform among them; demand lost to AI shopping tools had no line or figure then or since.

“Increasingly, customers are using computers, tablets and smart phones to shop with us and with our competitors and to do comparison shopping.”
Filing, risk factors, 10-K periodic report, 2024-03-15

By quarter

  • Q1 2026described · inscrutable · imposed
  • Q2 2026not mentioned · inscrutable · imposed
  • Q3 2026not mentioned · inscrutable · imposed

Reported lines, year-over-year growth

Revenue +5.9%

Q3 2026. Growing slower than revenue: cost of sales (+4.5%). A displaced cost shows up as a line that stays under the dashed revenue line. These are the audited lines, as first reported; nothing here is attributed to AI by the filing.

Cost of salesOperating, selling, general and administrative expensesRevenue
0%2%4%6%8%10%Q1 2025Q2 2025Q3 2025Q4 2025Q1 2026Q2 2026Q3 2026Operating, selling, general and administrative expensesRevenueCost of sales
Reported values and filings
LineQ2 2025Q3 2025Q4 2025Q1 2026Q2 2026Q3 2026
Total revenues$165.61bn$177.40bn$179.50bn$190.66bn$177.75bn$187.94bn
Cost of sales$124.30bn$132.77bn$134.71bn$143.62bn$133.06bn$138.80bn
Operating, selling, general and administrative expenses$34.17bn$37.34bn$38.09bn$38.33bn$37.20bn$39.75bn