The origin story is told so often it has worn smooth. In 2004 a German programmer who had moved to Canada the year before started an online snowboard shop called Snowdevil with Daniel Weinand and Scott Lake, found the available e-commerce software intolerable, and built his own on Ruby on Rails. By June 2006 the snowboards were gone and the software had become Shopify.
Everyone files this under pivot: a lucky accident, a founder who stumbled into a better business than the one he set out to run. That reading misses what actually happened, and it misses the thing Tobias Lütke has done four or five more times since.
He did not stumble. He gave something up. He has kept giving things up ever since — and each time, the thing he abandoned was the part that looked like the business, while the thing he kept was the layer underneath it.
The snowboards were the easy part
Walking away from Snowdevil is not, on its own, impressive. Selling snowboards online in 2004 was a small business with a small ceiling, and swapping it for infrastructure was the obvious trade once you had built the infrastructure.
What is worth noticing is that Lütke was already the kind of engineer who builds the layer below. He sat on the Ruby on Rails core team and wrote Active Merchant, the payments library, according to public records of his open-source work. A founder who left school at 16 for a programming apprenticeship, as Forbes tells it, tends to see a storefront as an interface problem sitting on top of a much more valuable plumbing problem.
The tell is what he did next, once the plumbing was worth real money.
The empire he declined to build
Shopify went public in May 2015, raising more than $131 million. At that point the platform had every excuse to become a destination — to aggregate its merchants’ demand into one marketplace and start collecting the customer relationship, which is the single most valuable asset in retail and the one Amazon had spent two decades assembling.
Lütke declined. His formulation of the strategy, quoted by Forbes, is that “Amazon is trying to build an empire. Shopify is trying to arm the rebels.”
It is a good line, and it is also a real cost. Shopify’s merchants keep their own brands, their own domains and their own buyers. Shopify takes a slice and stays invisible. That is a smaller near-term business than a marketplace and a far larger long-term one, because a merchant will leave a marketplace that competes with them and will not leave the machinery that runs their store.
Two subtractions that were not foresight
The pattern is not a story of a man who is always right. In July 2022 Shopify cut roughly 10% of its staff, about 1,000 people, after Lütke told employees the pandemic bet had not paid off and that on the call to expand, “I got this wrong.”
Ten months later he went further, cutting another 20% of the workforce and selling Shopify Logistics to Flexport in exchange for a 13% stake in Flexport. Internally he framed it through a distinction between the company’s main quest and its side quests: warehouses were a worthwhile side quest and not the reason the company exists.
Both moves were corrections of his own overreach, not prophecy. But the speed matters. Most founders defend an expensive mistake for years because unwinding it is an admission. Lütke unwound a multi-billion-dollar logistics ambition in under two years and took equity instead of cash, keeping optionality while shedding the operation.
Subtracting the headcount
In April 2025 he published an internal memo after it began leaking, under the title “Reflexive AI usage is now a baseline expectation at Shopify.” Its sharpest clause required teams to demonstrate why AI could not do a job before requesting a person to do it.
The memo was widely read as a threat. Read alongside the rest of his record, it looks more like the same move: remove the visible resource, keep the output. Shopify ended 2025 with roughly 7,600 employees — fewer than before the 2022 cuts — while full-year revenue grew about 30% to $11.56 billion on $378 billion of merchant GMV.
Now he is giving away the storefront
The most recent subtraction is the largest. On 11 January 2026, Shopify and Google announced the Universal Commerce Protocol, an open standard letting AI agents discover products and complete purchases on a merchant’s behalf. Shopify’s own announcement described merchants selling natively inside Google’s AI Mode, Gemini and Microsoft Copilot.
Follow that to its conclusion and the online store — the thing Shopify sells — stops being where the transaction happens. Lütke is helping standardise the protocol that makes his product’s front end optional, and making it open rather than proprietary, which gives away the lock-in too.
The quarter after suggests the calculation is not naive. On 5 August 2026 Shopify reported revenue of $3,583 million, up 34%, on GMV of $115,567 million, up 32%, with $654 million of free cash flow. Orders arriving through AI channels tripled year over year. Lütke’s own stake, about 6% of the company, is valued by Forbes at $13.1 billion.
The part that transfers
Subtraction is not a strategy on its own. Plenty of founders shed the wrong thing, keep the commodity layer and call it focus.
What makes Lütke’s version work is that he has been unusually consistent about which layer he thinks is load-bearing: never the surface a customer sees, always the machinery that has to keep working when the surface changes. Snowboards, marketplace, warehouses, headcount, storefront — all surfaces. All given away.
The question worth asking of your own company is not whether you would have the nerve to do that. It is whether you could name, right now, the one layer you would refuse to sell.
This is an editorial profile assembled from public reporting and Shopify’s own disclosures. Sources: Forbes, Shopify Newsroom, Shopify Q2 2026 results, CNBC, TechCrunch, Digital Commerce 360, Shopify Engineering, Wikipedia — Shopify, Wikipedia — Tobias Lütke.
