Shopify reported Q2 2026 earnings on August 5th, and the financials were strong: GMV of $115.6 billion, revenue of $3.58 billion, free cash flow margin at 18%. But the real story for merchants is buried in the AI numbers. AI-referred traffic and orders each tripled year over year, and the pattern of what is selling through AI channels should change how you think about your catalog before Q4.
The headline numbers
Three figures set the tone for the quarter, and all three beat expectations.
- GMV: $115.6 billion, up 32% year over year
- Revenue: $3.58 billion, up 34%
- Free cash flow margin: 18% of revenue, still climbing
Merchant Solutions revenue grew 37% and subscriptions grew 22%, so the growth is coming from merchants actually selling more, not just from plan pricing. Solid quarter. Now the part that matters for how you run your store.
AI-referred traffic and orders tripled
AI-referred traffic and AI-referred orders each grew 3x year over year in Q2. That is not a rounding error on a small base becoming a slightly bigger small base. Shopify also shared that roughly half of AI-referred sessions land directly on product pages, versus about 20% for traditional search. AI agents skip your homepage and your collection pages. They deliver a shopper straight to a product, which means the product record itself is doing all the selling.
75% of AI orders came from the long tail
Here is the number most merchants will skim past: 75% of AI-attributed orders came from outside Shopify's top 100 product categories. Broad, popular categories are where agent recommendations get used the least. The queries showing up in Shopify's AI purchase data look like this: brake rotor for a 2019 F-150, front, drilled. Specific, constrained, high-intent. That is the long tail, and it is exactly where AI agents are converting.
If you sell in a spec-heavy category like automotive aftermarket parts, this is the strongest signal yet that AI channels favor your kind of catalog. An agent answering a fitment question does not want a bestseller list. It wants the one part that fits.
AI channels bring first-time buyers
New-buyer orders, meaning first purchases from customers who had never bought from a store before, came in at nearly twice the rate through AI channels compared to other channels. Read that as an acquisition story: AI agents are functioning as a discovery engine, sending shoppers to merchants they would never have found through search or paid social. For merchants with strong catalog data, this is net-new demand.
The catch: your metafields do not reach agents
Now the uncomfortable part. All of that long-tail, spec-driven demand runs on structured product data, and your fitment metafields do not reach AI agents by default. Shopify Catalog Mapping routes your custom data into exactly three fields: title, description, and category. The agent-facing Catalog API exposes no merchant metafields at all. If the attribute an agent needs to answer a query is not in one of those three fields, the agent cannot see it, and your product silently drops out of the answer.
That means years of carefully built fitment data, year-make-model tables, spec metafields, and compatibility apps are invisible to the fastest-growing referral channel in commerce unless that data also lives in your titles, descriptions, and category assignments. We covered the tactical side of this in our guide to optimizing Shopify products for AI search.
What to do before Q4
The playbook follows directly from the data. First, audit which attributes actually drive purchase decisions in your category and confirm they appear in title, description, or category, not only in metafields. Second, fix attribute completeness across the catalog, because agents cannot recommend what they cannot parse. Third, verify your AI channel configuration so your catalog is flowing to agent surfaces at all.
At Ambaum we work with aftermarket parts merchants on exactly this: catalog structure, attribute completeness, and AI channel configuration. If you want to know where your data stands before Q4 traffic arrives, reach out for a fitment and AI readiness audit.
Read the full transcript
Shopify reported second quarter earnings on August 5th and the AI numbers were the story. Here is what actually matters for your store.
Start with the headline. GMV hit $115.6 billion, up 32% year over year. Revenue came in at $3.58 billion, up 34%, and free cash flow margin climbed to 18%.
Now, the AI takeaways. AI-referred traffic and AI-referred orders each grew three times year over year in the quarter. Not a rounding error, tripled.
Here is the part most merchants miss. 75% of AI-attributed orders came from outside Shopify's top 100 product categories. Broad categories are where agent recommendations get used the least. The queries showing up in Shopify's AI purchase data look like this: brake rotor for a 2019 F-150, front, drilled. That is the long tail and it is exactly where agents are converting.
And these are new customers. First purchases through AI channels came in at nearly twice the rate of other channels. Agents are bringing you buyers who have never bought from you before.
So here is the catch. Your fitment metafields do not reach agents by default. Shopify Catalog Mapping routes your custom data into exactly three fields: title, description, and category. And the agent-facing Catalog API exposes no merchant metafields at all. If the attribute is not in those three fields, the agent cannot see it.
At Ambaum, we work with aftermarket parts merchants on exactly this: catalog structure, attribute completeness, and AI channel configuration. Find out where your data stands before Q4. Reach out for a fitment and AI readiness audit.
