Interview

Harley Finkelstein on Shopify's Q2 blowout: $87B GMV, 12% US e-commerce share, and agentic commerce infrastructure

Aug 6, 2025 with Harley Finkelstein

Key Points

  • Shopify controls 12% of US e-commerce at checkout, second only to Amazon, giving it transaction-data advantages no individual retailer can match.
  • Enterprise brands including Michael Kors, Canada Goose, and Starbucks are migrating to Shopify as macro uncertainty forces stress-testing of legacy tech stacks for AI readiness.
  • Shopify is building agentic commerce infrastructure through Catalog, Universal Cart, and Checkout Kit to become the fulfillment layer as AI assistants emerge as a commerce surface.
Harley Finkelstein on Shopify's Q2 blowout: $87B GMV, 12% US e-commerce share, and agentic commerce infrastructure

Summary

Shopify delivered a strong Q2, with GMV reaching $87 billion, up 31% year-over-year. Revenue came in at $2.7 billion, and free cash flow hit $422 million, representing 16% of revenue. The company has now posted 11 consecutive quarters of positive free cash flow and 8 consecutive quarters of double-digit free cash flow growth. President Harley Finkelstein described the quarter as Shopify 'operating on all cylinders,' with merchants outperforming the broader e-commerce market despite macro headwinds around tariffs, de minimis policy changes, and consumer confidence concerns.

Shopify now controls 12% of US e-commerce, measured at checkout, making it the second-largest online checkout platform in the country behind Amazon. That number carries strategic weight beyond bragging rights: the scale of transaction data gives Shopify a checkout performance advantage that no individual retailer building their own stack can replicate. Finkelstein argued that a unified view of millions of merchants effectively makes Shopify the second-largest online retailer in America if treated as a single entity.

Enterprise Brand Momentum

New enterprise additions announced this quarter include Michael Kors, Canada Goose, Starbucks, Burton, and Camel. The Burton win carries symbolic significance: Shopify's origin story traces directly to founder Tobi Lütke building e-commerce software in 2004 to sell snowboards through a store called Snow Devil, software that eventually became Shopify itself. Birkenstock, Hunter Douglas, and Mattel were cited as examples of long-established brands — some founded in the 1700s and 1900s — now running on the platform.

Finkelstein noted that macro uncertainty is paradoxically accelerating enterprise migration. Larger brands that previously tolerated legacy or homegrown stacks are now stress-testing whether their technology is AI-ready and future-proofed, and many are concluding it is not. L Catterton, the PE firm backed in part by Bernard Arnault's family and run by Michael Chu, now routinely asks newly acquired portfolio brands when they are migrating to Shopify, a telling signal of how the enterprise perception has shifted.

Agentic Commerce Infrastructure

Shopify is positioning itself as the backend infrastructure layer for AI-driven commerce through three products. The first is Catalog, a search API launched this past quarter that uses large language models to index and categorize every SKU across Shopify's merchant base, making the full catalog accessible to any agent via MCP server or direct API. The second is Universal Cart, a new product that holds items from multiple Shopify stores in a single cart during an agent conversation, solving the multi-merchant problem in agentic shopping flows. The third is Checkout Kit, launched last year and updated this week, which embeds Shopify's checkout — including Shop Pay — directly into third-party agents, with theming now supported so the checkout matches the host application's look and feel. Microsoft Copilot is already integrated via Checkout Kit.

The strategic logic is straightforward: as AI assistants become a commerce surface, Shopify wants to be the fulfillment and transaction layer, the same role it plays in social commerce integrations across Instagram, YouTube, Roblox, and Spotify. Fenty Beauty was cited as an example of a brand already generating meaningful revenue through a virtual Shopify storefront inside Roblox, which draws over 20 million daily active users in the US, predominantly under 18.

Stablecoins and Cross-Border Payments

Shopify has supported cryptocurrency payments since 2012 via a Coinbase integration, but Finkelstein drew a clear line between that era's speculative use case and the current stablecoin opportunity. The current focus is on USDC through Coinbase, framed as a utility play rather than a crypto bet. The primary use case is cross-border settlement: merchants receive dollars with no new wallet required, no added consumer friction, faster settlement, and blockchain-based transparency. Finkelstein sees cross-border as a default operating mode for the best modern merchants rather than a feature, and stablecoins as a natural fit for that reality.

Monetization Questions in Agentic Commerce

Finkelstein acknowledged that the monetization model for AI-driven commerce remains unresolved. He is skeptical that paid placement will dominate agentic shopping the way it does in search, arguing that the value proposition of an AI shopping agent depends entirely on it acting in the consumer's interest rather than the highest bidder's. The analogy he used: a personal shopper who consistently recommends the highest-commission item loses the client. Shopify's positioning is to be present across every channel where commerce can happen rather than to bet on any single monetization model winning.

IPO Market Commentary

Ahead of the Shopify discussion, Finkelstein offered a direct endorsement of public markets, now 42 quarters post-IPO. He pushed back on the prevailing view among private companies that an IPO is a last resort, calling it one of the best decisions Shopify made for operational discipline and transparency. His advice to Figma CEO Dylan Field: maintain the same product culture and communication consistency post-IPO rather than treating the listing as a transformation event.