AI Absorption Ledger / SHOP

Shopify

SHOP · Q2 2026 · reported 2026-08-05 · revenue $3.58bn

Assessment

The spend side repeats Q1 with more words and still no number. The 10-Q names AI-related usage inside a increase in cloud and infrastructure costs and a increase in research and development computer hardware and software, and adds that cloud and infrastructure is the significant majority of cost of subscription solutions. The CFO places most of Sidekick's AI cost in subscription gross profit, held level while usage scaled, and the majority of internal AI spend in research and development. Cost of subscription solutions came in above its prior-year ratio to subscription revenue, the visible part of the AI cost. The ledger puts the merchant inference bill in that line at and internal usage at .

The narrative-defensive reading holds and its efficiency counterweight weakened. The tie between flat headcount and AI is not repeated: the CFO credits leverage to headcount discipline and AI to better output, and the President speaks of an AI reflexive company rebuilding teams around AI. The 10-Q shows sales and marketing employee-related costs down with no AI named. The workforce channel stays expanded, but with nothing this quarter tying the avoided payroll to AI it is left unsized; the step comes from the wording, not from a change at the company.

The revenue channels stepped. AI-platform commerce moves from directional to bounded: management calls agentic volume small relative to of GMV while orders rose times, and the ledger keeps its ballpark share, . Merchant AI tools move from described to directional on the first outcome attributed to Sidekick, a increase in new merchants reaching early orders, with pricing kept indirect. The agentic plan for brands without a Shopify store, described in Q1, is not mentioned. No toll appears: management reads AI as a complement to search, which it says is still growing.

Sized channels against the income statement, Q2 2026

5 of 8 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, Q2 2026

4 new4 expanded

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.

Paid for AI$55mn to $254mn sized

new $24mn to $129mnexpanded $31mn to $125mn

Incremental total $24mn to $254mnpoint $60mn$67mn in 1 channel has no traced baseline
Revenue arriving through AI$3.5mn to $57mn sized

new $2.8mn to $14mnexpanded $717k to $43mn

Incremental total $3.5mn to $57mnpoint $16mn

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 AI3 channels · $55mn to $254mn sized · $24mn to $254mn incremental

cost of-revenue

Model and inference bill for merchant-facing AI, the part in cost of subscription solutions

0.31% to 1.3% of the quarter’s revenue

Incremental total: counts in full.

Matched line moved : the whole line, not this channel.

our inferencedisclosure: direction only· motive: product-defensive· before LLMs: new

The 10-Q names AI-related usage inside a increase in cloud and infrastructure costs, which it calls the significant majority of cost of subscription solutions; the ledger reads that level at on the plain majority range. The CFO says most AI cost for merchant use of Sidekick sits here, while Sidekick handled about conversations. At the prior-year ratio to subscription revenue the line would have been , a visible excess of . The size shown is the ledger's own estimate, , for the part in cost of subscription solutions; the remainder of the Sidekick AI cost outside that line is not sized. The counterparty is mixed: model providers for model calls, and the cloud providers whose AI-related usage sits in cloud and infrastructure costs.

Evidence: 4 quotes, 7 figures, 2 confounds, 5 from before coverage

What the company pays to run the AI products merchants use, most of all Sidekick: model calls and the AI-related usage inside cloud and infrastructure costs, which the CFO says sit in subscription solutions gross profit. The channel is the part booked in cost of subscription solutions: the CFO places the vast majority of the Sidekick AI cost there, and the remainder, wherever it sits, is not sized here. The 10-Q names AI-related usage inside the year-over-year increase in cloud and infrastructure costs and never splits it. Merchants pay nothing extra for these products.

Why this motive

Inference cost in cost of revenue with no price change: the CFO says most of the AI cost of Sidekick sits in subscription gross profit and was absorbed with margins held level (claim c4), and agentic and AI features carry no separate pricing (claim c15). Carried from the prior quarter, with the same tell repeated.

Before LLMs: new

At the anchor Sidekick and the Shopify Magic features already existed, but the annual report described cost of subscription solutions ( for FY2024, against subscription revenue of ) as third-party infrastructure, hosting and support and explained its growth by cloud and infrastructure costs without naming AI. No model or token bill was given then.

“For example, Shopify Magic provides a suite of AI-enabled features integrated across the Shopify platform and Sidekick offers merchants an AI-enabled commerce assistant.”
Filing, business, 10-K periodic report, 2025-02-11
“Shopify does not currently develop its own foundational AI models, rather, we develop ways to incorporate AI-powered tools into select products we offer to merchants, in order to support elements of their business operations as well as into certain of our internal business operations.”
Filing, risk factors, 10-K periodic report, 2025-02-11
“Cost of subscription solutions consists primarily of third-party infrastructure, hosting costs and other direct costs, an allocation of costs incurred by both the operations and support functions, including personnel-related costs directly associated with subscription solutions such as salaries, benefits and stock-based compensation, processing fees related to billing our merchants, payments for domain registration and amortization of acquired intangible assets.”
Filing, mdna, 10-K periodic report, 2025-02-11
“Cost of subscription solutions increased for the year ended December 31, 2024 compared to the same period in 2023. The increase was due mainly to an increase in cloud and infrastructure costs and increase in payment processing fees on merchant billings.”
Filing, mdna, 10-K periodic report, 2025-02-11
“Gross margin for subscription solutions was 81.4% compared to 78.0% in Q1 of 2023. The increase stems from pricing changes on standard plans and, to a lesser extent, continued support and hosting efficiencies.”
CFO, prepared remarks, earnings call, 2024-05-08

Figures

  • Increase in cloud and infrastructure costs, which includes AI-related usage, year over year · 2026-CQ2
  • Cloud and infrastructure costs within cost of subscription solutions, from the 10-Q's "significant majority" · 2026-CQ2
  • Cost of subscription solutions growth, year over year · 2026-CQ2
  • Subscription solutions revenue growth, year over year · 2026-CQ2
  • Sidekick conversations in the quarter, approximate · 2026-CQ2
  • Cost of subscription solutions at the prior-year ratio to subscription revenue · 2026-CQ2
  • Cost of subscription solutions above the prior-year ratio to subscription revenue · 2026-CQ2

Reported line it is matched to

Cost of subscription solutions rose , , against subscription revenue growth of ; the 10-Q puts of the increase in cloud and infrastructure costs, which include AI-related usage and are the significant majority of the line.

2026-CQ2: 2025-CQ2:

What else could explain it

  • line composition: Cloud and infrastructure costs hold ordinary hosting for the whole platform; the AI part is not split.
  • operating leverage: The CFO credits cost efficiencies and support with holding margin level, so the line understates the AI bill by whatever those efficiencies saved.

Quotes

“Cost of subscription solutions increased for the three months ended June 30, 2026 compared to the same period in 2025. The increase was driven by a $37 million increase in cloud and infrastructure costs which includes AI-related usage. Cloud and infrastructure costs remains the significant majority of our cost of subscription solutions.”
c1 · Filing, mdna, 10-Q periodic report, 2026-08-05
“As a reminder, the vast majority of AI costs related to merchant use of Sidekick appear in subscription solutions gross profit. We were able to hold gross margins at a relatively consistent level quarter-over-quarter while Sidekick usage scaled, which reflects some cost efficiencies and support, as well as our ability to continue providing merchants unique AI solutions like Sidekick while diligently managing cost.”
c4 · CFO, prepared remarks, earnings call, 2026-08-05
“In the second quarter, daily active merchants using Sidekick were up 3.6x year-on-year, and daily sessions were up 4.8x. It handled nearly 34 million conversations, and it was used to create more than 36,000 custom apps, up from 12,000 in Q1.”
c13 · Executive, prepared remarks, earnings call, 2026-08-05
“Actually, what's really cool is we now have a number of distilled models where we take a teacher model, usually a big frontier model, then teach it a specific use case to a smaller model, which results in much faster, less costly, and actually sometimes even better at the narrow task.”
c17 · Executive, qa, earnings call, 2026-08-05

By quarter

  • Q1 2026direction only · our inference · product-defensive · $4.2mn to $30mn
  • Q2 2026direction only · our inference · product-defensive · $11mn to $48mn

vendor bill

AI tools and model usage inside the company

0.35% to 2.3% of the quarter’s revenue

Incremental total: counts in full.

our inferencedisclosure: direction only· motive: narrative-defensive· before LLMs: new

The CFO says the majority of internal AI spend is in research and development and explains an uptick in its growth; the 10-Q names AI-related usage inside a increase in computer hardware and software costs there, against more in employee-related costs. Usage runs through an internal proxy with distilled models for repeated tasks. The size shown is the ledger's own estimate, . The counterparty is mixed: frontier model providers, vendors of coding agents, and the hardware and software providers behind the internal proxy and distilled models.

Evidence: 7 quotes, 2 figures, 2 confounds, 3 from before coverage

What the company pays for AI used by its own staff: coding agents, the internal coding partner in Slack, the internal model proxy, frontier and smaller distilled models. The CFO says the majority of internal AI spend sits in research and development, where the 10-Q names AI-related usage inside the increase in computer hardware and software costs.

Why this motive

The tells still conflict. The President credits the CEO with making the company AI reflexive and says work is being rebuilt around AI (claims c18 and c16), the mandate tell; the CFO now speaks of cost-managed use and model tiers (claim c5) and frames the benefit as quality of output (claim c8), not as a displaced line. Carried as narrative-defensive, the less durable reading.

Before LLMs: new

The anchor says AI-powered tools were built into certain internal business operations and that research and development ( for FY2024) carried internal hosting; no internal AI bill or usage cost was named.

“Shopify does not currently develop its own foundational AI models, rather, we develop ways to incorporate AI-powered tools into select products we offer to merchants, in order to support elements of their business operations as well as into certain of our internal business operations.”
Filing, risk factors, 10-K periodic report, 2025-02-11
“Research and development expenses consist primarily of employee-related expenses for product management, product development, product design, data analytics, contractor and consultant fees as well as internal use hosting costs and corporate overhead allocations.”
Filing, mdna, 10-K periodic report, 2025-02-11
“We continue to remain disciplined on headcount, with total headcount remaining essentially flat for the past three quarters, all while maintaining and, in fact, accelerating our product innovation capabilities and continuing the top-line momentum of our business. How we leverage AI internally is an important element of how we are able to do that.”
CFO, prepared remarks, earnings call, 2024-05-08

Figures

  • Increase in computer hardware and software costs within research and development, which includes AI-related usage, year over year · 2026-CQ2
  • Research and development growth, year over year · 2026-CQ2

What else could explain it

  • line composition: Computer hardware and software costs within research and development carry non-AI software and equipment.
  • bundling: Staff use and merchant products may share one proxy and one set of vendor agreements; the split between this channel and the merchant inference bill is the ledger's.

Quotes

“Research and development expenses increased for the three months ended June 30, 2026 compared to the same period in 2025, due to increases of $34 million in computer hardware and software costs, which includes AI-related usage, and $11 million in employee-related costs.”
c2 · Filing, mdna, 10-Q periodic report, 2026-08-05
“This operating leverage provides the financial flexibility to continue to invest in our platform, including in our internal AI capabilities. We've moved from a place of just reflexive use of AI to a place of AI leverage. Our AI philosophy is straightforward: maximum leverage paired with thoughtful cost management. We use the best model for the job, frontier intelligence where it matters, less expensive models where it doesn't.”
c5 · CFO, prepared remarks, earnings call, 2026-08-05
“R&D, the majority of our internal AI spend is allocated here, so you've seen a modest uptick in year-over-year growth.”
c7 · CFO, prepared remarks, earnings call, 2026-08-05
“We believe widespread adoption of AI tooling already is and will continue to yield benefits in the quality of our output.”
c8 · CFO, prepared remarks, earnings call, 2026-08-05
“Over the past year or so, we've really moved in from experimenting with AI to just rebuilding teams and our work around what it can do. The way we think about these models is we use the best model for the job. We use frontier intelligence where it matters. We use less expensive models where it's simply not needed. Everything that we've done runs through an internal proxy, which gives us incredible visibility, gives us control, gives us security.”
c16 · Executive, qa, earnings call, 2026-08-05
“Actually, what's really cool is we now have a number of distilled models where we take a teacher model, usually a big frontier model, then teach it a specific use case to a smaller model, which results in much faster, less costly, and actually sometimes even better at the narrow task.”
c17 · Executive, qa, earnings call, 2026-08-05
“Shopify is probably the most AI-pilled company in the world. Certainly, credit to that goes to Tobi, who I think has been thinking about this longer than most and certainly put Shopify on the path of being AI reflexive before almost anyone else.”
c18 · Executive, qa, earnings call, 2026-08-05

By quarter

  • Q1 2026direction only · our inference · narrative-defensive · $9.7mn to $65mn
  • Q2 2026direction only · our inference · narrative-defensive · $13mn to $81mn

engineering

Internal investment in AI capability (Catalog, protocol, AI infrastructure)

0.87% to 3.5% of the quarter’s revenue

Incremental total: counts at zero at the point and in full at the high end, with no traced baseline.

our inferencedisclosure: described· motive: exploratory· before LLMs: expanded

The build named this quarter is the Catalog search index for agents, the commerce protocol and Catalog API now open to every builder, and distilled models. Research and development grew to . The size shown is the ledger's own estimate, , an assumed share of the line's payroll.

Evidence: 5 quotes, 3 figures, 1 confound, 4 from before coverage

Engineering and product payroll committed to building AI into the platform: the Catalog search index for agents, the Universal Commerce Protocol, the agentic admin and attribution, distilled models, and what the CFO calls investment in AI infrastructure and internal AI capabilities. It sits in research and development and is not split.

Why this motive

Investing for the long term, carried: the President describes years of investment in a search index for agents and opening the Catalog API to every builder (claims c19 and c27), and the CFO lists internal AI capabilities among what leverage funds (claim c5). No return on the build is measured.

Before LLMs: expanded

At the anchor the company was already investing in expanding AI capabilities in its products and in its own machine learning models for marketing, inside research and development ( for FY2024), with no AI share given then or since. The size is the whole of an activity that existed before.

“Shopify does not currently develop its own foundational AI models, rather, we develop ways to incorporate AI-powered tools into select products we offer to merchants, in order to support elements of their business operations as well as into certain of our internal business operations.”
Filing, risk factors, 10-K periodic report, 2025-02-11
“We are making investments in expanding the AI capabilities available in our products, including the ongoing deployment and improvement of existing machine learning and AI technologies.”
Filing, risk factors, 10-K periodic report, 2025-02-11
“Research and development expenses consist primarily of employee-related expenses for product management, product development, product design, data analytics, contractor and consultant fees as well as internal use hosting costs and corporate overhead allocations.”
Filing, mdna, 10-K periodic report, 2025-02-11
“Our teams have created and leveraged advanced models using AI and machine learning, which now allows us to target our Audiences with unprecedented precision.”
Executive, prepared remarks, earnings call, 2024-05-08

Figures

  • Research and development · 2026-CQ2
  • Research and development, prior year · 2025-CQ2
  • Research and development growth, year over year · 2026-CQ2

What else could explain it

  • line composition: Research and development holds all product and engineering payroll and the AI usage bill; the AI build is not split.

Quotes

“This operating leverage provides the financial flexibility to continue to invest in our platform, including in our internal AI capabilities. We've moved from a place of just reflexive use of AI to a place of AI leverage. Our AI philosophy is straightforward: maximum leverage paired with thoughtful cost management. We use the best model for the job, frontier intelligence where it matters, less expensive models where it doesn't.”
c5 · CFO, prepared remarks, earnings call, 2026-08-05
“Actually, what's really cool is we now have a number of distilled models where we take a teacher model, usually a big frontier model, then teach it a specific use case to a smaller model, which results in much faster, less costly, and actually sometimes even better at the narrow task.”
c17 · Executive, qa, earnings call, 2026-08-05
“For nearly two years, we've been investing in the search index, ensuring over a billion products and 20 years of commerce experience is distilled for agents.”
c19 · Executive, prepared remarks, earnings call, 2026-08-05
“Every builder can now access UCP and the Catalog API across millions of merchants, so they can build commerce experiences with the same infrastructure as our major AI partners.”
c27 · Executive, prepared remarks, earnings call, 2026-08-05
“On top of this, thousands of developers all over the world are embedding commerce into their own applications using the same primitives we built.”
c28 · Executive, prepared remarks, earnings call, 2026-08-05

By quarter

  • Q1 2026described · our inference · exploratory · $31mn to $122mn
  • Q2 2026described · our inference · exploratory · $31mn to $125mn

Cost displaced by AI2 channels · 2 not sized

engineering · cheap to verify

Headcount held flat with AI across functions

Not sized

No Q2 source ties the avoided payroll to AI: the CFO credits operating leverage to headcount discipline without naming AI (claim c6) and credits AI only with the quality of output (claim c8), and the 10-Q names no AI cause for the employee-related cost changes (claims c2 and c3). The test reads each quarter’s own sources, so the AI share held from Q1 is not carried; the step comes from the wording, not from a change at the company. Payroll in the three operating lines, inside , is the ceiling.

described, no sizedisclosure: described· motive: narrative-defensive· before LLMs: expanded

The quarter does not repeat the tie between flat headcount and AI: the CFO credits leverage to headcount discipline and AI to better output, and the President says teams are being rebuilt around AI. The 10-Q shows sales and marketing employee-related costs down and research and development employee-related costs up , neither attributed to AI. With nothing tying the avoided payroll to AI this quarter, the channel is left unsized; the former estimate, which held the prior quarter’s AI share, is no longer used. The channel is expanded and sized as an increment when it is sized.

Evidence: 6 quotes, 3 figures, 2 confounds, 4 from before coverage

Payroll the company says it avoids because staff work with AI: flat headcount while revenue grows, a large share of code written by AI, app review on AI testing. Management describes it for every function; engineering carries the measured claims. The lines it would show in are research and development, sales and marketing and general and administrative.

Why this motive

Carried, and the efficiency tell weakened: this quarter the CFO credits operating leverage to headcount discipline without naming AI (claim c6) and frames AI as improving the quality of output (claim c8), while the President speaks of the company as AI reflexive (claim c18), the mandate tell. The 10-Q shows sales and marketing employee-related costs falling and names no AI cause (claim c3).

Before LLMs: expanded

Flat headcount predates coverage: at the anchor call headcount had been flat for quarters, which the President credited to structure and automation and the CFO in part to internal AI, and the annual report planned growth without significant additional hiring, with employees at the end of 2024. In coverage LLM tools are deployed in the work and management credits flat headcount to AI; the 10-Qs attribute no saving to AI, though sales and marketing employee-related costs fell in Q2 2026. The size is the change AI made, not the whole line.

“Research and development expenses consist primarily of employee-related expenses for product management, product development, product design, data analytics, contractor and consultant fees as well as internal use hosting costs and corporate overhead allocations.”
Filing, mdna, 10-K periodic report, 2025-02-11
“While we intend to expand our business without significant additional hiring in the near term, we may be unable to achieve that expansion without increasing our headcount.”
Filing, risk factors, 10-K periodic report, 2025-02-11
“which has enabled us to consistently deliver 25% revenue growth excluding logistics, all while keeping our headcount flat for three straight quarters. More importantly, because of the structure and the automation we have worked to put in place, we think we can continue to operate against very limited headcount growth while achieving a continued combination of consistent top-line growth and profitability.”
Executive, prepared remarks, earnings call, 2024-05-08
“We continue to remain disciplined on headcount, with total headcount remaining essentially flat for the past three quarters, all while maintaining and, in fact, accelerating our product innovation capabilities and continuing the top-line momentum of our business. How we leverage AI internally is an important element of how we are able to do that.”
CFO, prepared remarks, earnings call, 2024-05-08

Figures

  • Decrease in employee-related costs within sales and marketing, year over year · 2026-CQ2
  • Increase in employee-related costs within research and development, year over year · 2026-CQ2
  • Research and development, sales and marketing, and general and administrative combined · 2026-CQ2

What else could explain it

  • transformation program: Headcount discipline predates coverage: headcount was flat for quarters at the anchor call, credited to structure and automation, with internal AI named as one element; the AI share in the estimate discounts it.
  • operating leverage: Operating lines growing slower than revenue of is ordinary leverage for a platform whose revenue follows GMV.

Quotes

“Research and development expenses increased for the three months ended June 30, 2026 compared to the same period in 2025, due to increases of $34 million in computer hardware and software costs, which includes AI-related usage, and $11 million in employee-related costs.”
c2 · Filing, mdna, 10-Q periodic report, 2026-08-05
“Sales and marketing expenses increased for the three months ended June 30, 2026 compared to the same period in 2025, due to increases of $85 million in overall marketing program spend and $10 million in payouts related to our affiliate partner programs, offset by a $16 million decrease in employee-related costs.”
c3 · Filing, mdna, 10-Q periodic report, 2026-08-05
“This reflects continued leverage as we grew gross profit dollars faster than our expenses, primarily by maintaining our discipline in headcount.”
c6 · CFO, prepared remarks, earnings call, 2026-08-05
“We believe widespread adoption of AI tooling already is and will continue to yield benefits in the quality of our output.”
c8 · CFO, prepared remarks, earnings call, 2026-08-05
“Over the past year or so, we've really moved in from experimenting with AI to just rebuilding teams and our work around what it can do. The way we think about these models is we use the best model for the job. We use frontier intelligence where it matters. We use less expensive models where it's simply not needed. Everything that we've done runs through an internal proxy, which gives us incredible visibility, gives us control, gives us security.”
c16 · Executive, qa, earnings call, 2026-08-05
“Shopify is probably the most AI-pilled company in the world. Certainly, credit to that goes to Tobi, who I think has been thinking about this longer than most and certainly put Shopify on the path of being AI reflexive before almost anyone else.”
c18 · Executive, qa, earnings call, 2026-08-05

By quarter

  • Q1 2026bounded · our inference · narrative-defensive · $1.5mn to $46mn
  • Q2 2026described · described, no size · narrative-defensive

customer support · cheap to verify

Merchant support handled with AI

Not sized

No source in the quarter attributes a support cost movement to AI or gives a support cost line; the AI share of support interactions was last given at the anchor call.

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

Support appears only as one of the efficiencies that held subscription margin level while Sidekick usage scaled. Neither the call nor the 10-Q ties it to AI, as the anchor call did. The channel is carried with no reading of its own.

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

Support staff cost avoided because AI assists or resolves merchant support interactions. The cost sits in cost of subscription solutions. The covered quarters cite efficiencies in support next to the LLM bill without attributing them to AI.

Why this motive

Carried from the prior quarter, read as exploratory: the CFO again cites efficiencies in support without saying what produced them (claim c4, cited under the inference bill).

Before LLMs: expanded

At the anchor call over of merchant support interactions were assisted with AI, AI had added round-the-clock support in languages, and support efficiencies were already a cause of subscription margin; support cost sat in cost of subscription solutions ( for FY2024). The size is the change AI made, not the whole line.

“Cost of subscription solutions consists primarily of third-party infrastructure, hosting costs and other direct costs, an allocation of costs incurred by both the operations and support functions, including personnel-related costs directly associated with subscription solutions such as salaries, benefits and stock-based compensation, processing fees related to billing our merchants, payments for domain registration and amortization of acquired intangible assets.”
Filing, mdna, 10-K periodic report, 2025-02-11
“During Q1, over half of our merchant support interactions were assisted with AI and often fully resolved with the help of AI.”
CFO, prepared remarks, earnings call, 2024-05-08
“Additionally, our support staff has experienced a significant reduction in the amount of toil that is part of their jobs. We are improving the merchant support process and achieving much greater efficiency than ever before.”
CFO, prepared remarks, earnings call, 2024-05-08
“Gross margin for subscription solutions was 81.4% compared to 78.0% in Q1 of 2023. The increase stems from pricing changes on standard plans and, to a lesser extent, continued support and hosting efficiencies.”
CFO, prepared remarks, earnings call, 2024-05-08
“The increase was primarily due to an increase in payment processing fees on merchant billings and cloud and infrastructure costs, partially offset by a decrease in support costs.”
Filing, mdna, 10-K periodic report, 2025-02-11

What else could explain it

  • operating leverage: Cost efficiencies and support are cited together as what held subscription margin; neither is attributed to AI.

By quarter

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

Revenue arriving through AI3 channels · $3.5mn to $57mn sized · $3.5mn to $57mn incremental · 1 not sized

product revenue · cheap to verify

AI tools for merchants (Sidekick, Pulse, Shopify Magic, AI Toolkit)

0.02% to 1.2% of the quarter’s revenue

Incremental total: counts in full.

our inferencedisclosure: direction only· motive: product-defensive· before LLMs: expanded

Management gave the first outcome attributed to Sidekick: onboarding guidance led to a increase in new merchants reaching early order milestones. Usage kept growing, daily active merchants times a year earlier and over custom apps, and the President said pricing stays indirect. The size shown is the ledger's own estimate, .

Evidence: 7 quotes, 5 figures, 3 confounds, 3 from before coverage

Revenue the company may gain because merchants use its AI assistant and tools to set up stores, build custom apps and automations, and run analytics: more merchants reaching early sales, more retention, more GMV. The tools are included at no extra price; management says monetization is indirect, through merchant success.

Why this motive

Included at no additional cost, carried: monetization stays indirect, through merchant success (claim c15), and the cost is absorbed in subscription margin (claim c4). The first measured outcome attributed to Sidekick, more new merchants reaching early orders (claim c14), is a merchant milestone, not a movement in the company's price or revenue per merchant, so the offensive tell is not met.

Before LLMs: expanded

Sidekick and the Shopify Magic suite were already offered to merchants at the anchor, inside subscription solutions revenue ( for FY2024), with no price of their own and no measured effect on revenue. The size is the change AI made, not the whole line.

“For example, Shopify Magic provides a suite of AI-enabled features integrated across the Shopify platform and Sidekick offers merchants an AI-enabled commerce assistant.”
Filing, business, 10-K periodic report, 2025-02-11
“Shopify does not currently develop its own foundational AI models, rather, we develop ways to incorporate AI-powered tools into select products we offer to merchants, in order to support elements of their business operations as well as into certain of our internal business operations.”
Filing, risk factors, 10-K periodic report, 2025-02-11
“Currently, the most practical applications of AI are found in tools that simplify business operations and enhance productivity, all of which we have been developing deeper capabilities with our AI product suite Shopify Magic.”
Executive, prepared remarks, earnings call, 2024-05-08

Figures

  • Daily active merchants using Sidekick, multiple of the prior year · 2026-CQ2
  • Daily active merchants using Sidekick, multiple of the prior year (Q&A figure) · 2026-CQ2
  • Sidekick conversations in the quarter, approximate · 2026-CQ2
  • Custom apps created with Sidekick in the quarter, floor · 2026-CQ2
  • Increase in new merchants reaching an early order milestone, attributed to Sidekick onboarding guidance · 2026-CQ2

What else could explain it

  • bundling: The tools are part of the subscription at no added price; any effect is inside retention and GMV.
  • other: The prepared remarks and the Q&A give different multiples for daily active merchants ( times and times); the ledger records both and uses neither in the size.
  • mix shift: Merchants who adopt an assistant early may be the ones who would have reached early orders anyway; the onboarding figure does not say how the comparison was drawn.

Quotes

“As a reminder, the vast majority of AI costs related to merchant use of Sidekick appear in subscription solutions gross profit. We were able to hold gross margins at a relatively consistent level quarter-over-quarter while Sidekick usage scaled, which reflects some cost efficiencies and support, as well as our ability to continue providing merchants unique AI solutions like Sidekick while diligently managing cost.”
c4 · CFO, prepared remarks, earnings call, 2026-08-05
“In the second quarter, daily active merchants using Sidekick were up 3.6x year-on-year, and daily sessions were up 4.8x. It handled nearly 34 million conversations, and it was used to create more than 36,000 custom apps, up from 12,000 in Q1.”
c13 · Executive, prepared remarks, earnings call, 2026-08-05
“Sidekick's personalized guidance for new merchants during onboarding led to an 8% increase in merchants reaching five orders within 15 days.”
c14 · Executive, prepared remarks, earnings call, 2026-08-05
“Look, we reserve the right to wake up smarter every single day. In terms of monetization now, the business model is predicated on merchants doing well. The better merchants do, the better Shopify does, and Sidekick helps with that.”
c15 · Executive, qa, earnings call, 2026-08-05
“At the same time, we've built connectors to agents including Claude, ChatGPT, Perplexity, Manus, Replit, and Vercel with our AI Toolkit.”
c25 · Executive, prepared remarks, earnings call, 2026-08-05
“If you look at five years in, setup accounts for 8% because they're already at scale, and that's where analytics and reporting make up more than 40%.”
c30 · Executive, qa, earnings call, 2026-08-05
“What's really interesting is if you look at daily active merchants using Sidekick, it grew 3x year-over-year.”
c31 · Executive, qa, earnings call, 2026-08-05

By quarter

  • Q1 2026direction only · described, no size · product-defensive
  • Q2 2026direction only · our inference · product-defensive · $717k to $43mn

distribution · expensive to verify

Orders arriving through AI platforms (ChatGPT, Copilot, Google AI, Perplexity)

0.08% to 0.39% of the quarter’s revenue

Incremental total: counts in full.

our inferencedisclosure: bounded· motive: channel-defensive· before LLMs: new

Management now limits the level: agentic volume is small relative to quarterly GMV of , while traffic and orders each rose times in a year, of AI-referred sessions land directly on a product page, conversion from AI search runs above organic search, and of AI-attributed orders come from outside the top categories. The words table does not convert small, so the size shown is the ledger's own estimate, , at the prior quarter's share read through the phrase table from the travel peer's bound.

Evidence: 14 quotes, 6 figures, 1 confound, 1 from before coverage

Merchant orders that start in an outside AI assistant, through checkout inside the assistant or referral to the storefront, earning the same payments and subscription economics as any online store order. Includes Shop Campaigns placements on ChatGPT. Management gives growth multiples and conversion ratios, not a level.

Why this motive

Integrations to be present where buyers start, carried: the company opens Catalog and the commerce protocol to AI partners and builders (claim c27) and earns the same economics on each order (claim c11). Conversion ratios are given (claims c20 and c21) but no revenue movement is measured.

Before LLMs: new

At the anchor the annual report listed storefronts, physical retail and social media as sales channels and named no AI platform; no traffic or orders from AI assistants were reported. Total revenue was in FY2024.

“This platform allows merchants to seamlessly manage, market and sell their products across various sales channels, including online storefronts, physical retail spaces, social media and more.”
Filing, business, 10-K periodic report, 2025-02-11

Figures

  • Orders arriving through AI channels, multiple of the prior year · 2026-CQ2
  • Conversion from AI search above traditional organic search · 2026-CQ2
  • Share of AI-attributed orders from outside the top categories · 2026-CQ2
  • GMV in the quarter, against which agentic volume is called small · 2026-CQ2
  • Share of AI-referred sessions landing directly on a product page · 2026-CQ2
  • AI-referred product-page landing rate relative to traditional search · 2026-CQ2

What else could explain it

  • mix shift: Some orders through AI assistants would have come through search; the President says search sessions are still growing and calls AI a complement, which the ledger cannot test.

Quotes

“While the volume from agentic commerce is still small relative to our massive GMV, the growth trends are impressive. Both AI-driven traffic and also orders to Shopify stores tripled year-over-year in the second quarter. New buyer orders are coming in at nearly twice the rate of other channels.”
c9 · Executive, prepared remarks, earnings call, 2026-08-05
“Search remains one of our largest sources of buyer traffic to our merchants, and it's still growing. Traditional search sessions are up 1.3x over the past two years, holding roughly a third of all storefront sessions. That is AI as a complement to search rather than a substitute for it.”
c10 · Executive, prepared remarks, earnings call, 2026-08-05
“You've seen this, of course, but agentic transactions carry the exact same economics as an online store transaction. There's no new fees. There's no separate pricing. More agentic GMV, it means more Shopify revenue, and that's the model, and it's been working really well for almost two decades.”
c11 · Executive, qa, earnings call, 2026-08-05
“I will say, I mentioned in my earlier comments that we are seeing incremental dollars flow through agentic.”
c12 · Executive, qa, earnings call, 2026-08-05
“We're seeing that AI searches powered by Catalog converted twice the rate of those using scraped data.”
c20 · Executive, prepared remarks, earnings call, 2026-08-05
“When you zoom it even further, conversion from AI search runs nearly 80% higher than traditional organic search as well.”
c21 · Executive, qa, earnings call, 2026-08-05
“We saw that AI search was starting to disproportionately benefit the long tail in 2025, and that trend has continued, with 75% of AI-attributed orders in the second quarter coming from outside our top 100 categories in Q2.”
c22 · Executive, prepared remarks, earnings call, 2026-08-05
“It is early, and the volume relative to $116 billion GMV is small.”
c23 · Executive, qa, earnings call, 2026-08-05
“Number one, I think consumers are finding, we have the data to back it up. Traffic on agentic is 3x year-on-year. Orders are up 3x as well.”
c24 · Executive, qa, earnings call, 2026-08-05
“I think we talk about Catalog from the consumer perspective quite a bit, but it also is a driver to come to Shopify.”
c26 · Executive, qa, earnings call, 2026-08-05
“Every builder can now access UCP and the Catalog API across millions of merchants, so they can build commerce experiences with the same infrastructure as our major AI partners.”
c27 · Executive, prepared remarks, earnings call, 2026-08-05
“On top of this, thousands of developers all over the world are embedding commerce into their own applications using the same primitives we built.”
c28 · Executive, prepared remarks, earnings call, 2026-08-05
“Early indications show that AI search has been particularly helpful to some of the smaller brands that form the long tail of commerce.”
c29 · Executive, prepared remarks, earnings call, 2026-08-05
“Buyer shopping journeys are being compressed as half of all AI-referred sessions are landing directly on a product description page. That is 2.5x more than what we see with traditional search.”
c32 · Executive, prepared remarks, earnings call, 2026-08-05

By quarter

  • Q1 2026direction only · our inference · channel-defensive · $2.4mn to $12mn
  • Q2 2026bounded · our inference · channel-defensive · $2.8mn to $14mn

product revenue · cheap to verify

Agentic plan: Catalog listing for brands not on Shopify

Not sized

The plan is not mentioned this quarter, and it never had a price, a count of brands or revenue.

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

The call speaks of Catalog as a reason retailers come to Shopify and of the Catalog API opened to builders, but not of the plan that lists brands without a Shopify store. The channel is carried with no reading of its own.

Evidence: 0 quotes, 1 from before coverage

A plan launched in March 2026 that lets brands on any platform make their products available to AI agents through Shopify Catalog without a Shopify store. The sources give no price, no count of brands and no revenue; management presents it also as a path to full migrations.

Why this motive

Carried from the prior quarter; the quarter's sources do not mention the agentic plan for brands without a Shopify store.

Before LLMs: new

Nothing like it exists at the anchor: the annual report names no product that lists brands without a Shopify store on AI platforms, and selling to brands off the platform was then a matter of checkout and commerce components.

“This platform allows merchants to seamlessly manage, market and sell their products across various sales channels, including online storefronts, physical retail spaces, social media and more.”
Filing, business, 10-K periodic report, 2025-02-11

By quarter

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

Reported lines, year-over-year growth

Revenue +33.7%

Q2 2026. Growing slower than revenue: research and development (+12.9%), sales and marketing (+20.0%), general and administrative (+11.5%). 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. Not drawn: general and administrative, which one-time items move by more than 60% in a quarter; the values are in the table below.

Cost of revenuesResearch and developmentSales and marketingRevenue
0%10%20%30%40%Q1 2025Q2 2025Q3 2025Q4 2025Q1 2026Q2 2026Cost of revenuesRevenueSales and marketingResearch and development
Reported values and filings
LineQ1 2025Q2 2025Q3 2025Q4 2025Q1 2026Q2 2026
Revenue$2.36bn$2.68bn$2.84bn$3.67bn$3.17bn$3.58bn
Cost of revenues$1.19bn$1.38bn$1.45bn$1.98bn$1.62bn$1.88bn
Research and development$377mn$394mn$375mn$390mn$437mn$445mn
Sales and marketing$405mn$415mn$410mn$433mn$496mn$498mn
General and administrative$109mn$122mn$115mn$125mn$115mn$136mn