Is Shopify harming the partner ecosystem?

The question going around partner Slack channels and the Shopify developer forums is blunt: is Shopify harming the partner ecosystem?

I think the honest answer is yes, if you mean the ecosystem as a services market. Not because merchants should stay dependent on agencies. Because Shopify is now shipping the tools that used to be the invoice.

Canvas, rolling out from the start of October 2026, is a design surface where merchants build a store with Sidekick. Shopify’s own numbers make the direction hard to miss. Sidekick has been generating apps, building custom theme sections, and making more than twenty-five million theme edits in the first half of 2026. Ben Sehl, Director of Product for Online Store, put the pitch in one line: an entrepreneur can build a fully custom store in twenty minutes that used to take someone who could code two weeks.

That is a gift to a founder with a clear idea and a thin budget. It is also a direct substitute for a large slice of theme work, section work, and “can you just build me a small tool” work that kept freelancers, agencies, and early app developers in business.

What Shopify is actually taking

Partners did not get paid, in the main, for inventing commerce. They got paid for closing the gap between what Shopify shipped and what a merchant could operate.

That gap used to look like this. A theme from the Theme Store, then a few weeks of custom sections, landing pages, and merchandising logic. A small embedded app when the admin could not do the workflow. A developer on retainer because Liquid, metafields, and app conflicts were too expensive to learn in the middle of a launch.

Canvas and Sidekick collapse that gap on purpose. Sidekick now works on the theme’s files, not only on isolated editor knobs. Theme architecture has been simplified so an agent can change structure, logic, and design across templates instead of one section at a time. Merchants can pan across every page, describe what they want, and get coordinated code back. Shopify has also been clear that themes are becoming a starting point, not the finished store, and that it is “deeply considering” the role of the Theme Store in that world.

App developers are not spectators. Sidekick already generates apps. At launch, Canvas does not support theme app extensions, and themes created or duplicated in Canvas will not expose app blocks and embeds. Shopify says that support is coming and that apps still work in the existing editor. The forum thread is already full of partners saying the quiet part: if a merchant can design the store without me, why would they keep recommending the platform I build on?

That reaction is not nostalgia. It is a pricing signal. When the platform absorbs the task, the rate for the task goes to zero.

Empowering merchants is the product strategy, not a side effect

I do not think this is an accident, or a temporary overreach while the models are new. Shopify has been walking this path in public for a while.

Readable Liquid, typed blocks, and agent-friendly theme contracts, the direction in Liquid: Back to the future, exist so a person or an agent can change a store without a specialist in the room. Sidekick app extensions, which we have shipped ourselves in RWA: Waitlist Flow, put app actions inside the same assistant the merchant already talks to. The mid-September platform updates kept moving admin, metafields, catalog, and dev-store tooling toward something an agent can operate.

Each release is easy to applaud on its own. Together they move the default builder from “hire a partner” to “ask Sidekick.”

Merchants should have that. A store that only exists if an agency is on the clock is a fragile store. The problem for partners is pretending the old menu still has the same prices. Custom theme sections, one-off landing pages, and thin admin utilities were the bread and butter. Those are exactly the jobs Sidekick is already counting in the millions.

The money comes off the table, then the work changes shape

From where I sit, building Shopify apps and doing store work at Reiwa Dev, this feels less like “learn a new API” and more like a pay cut for a whole category of labour.

Shopify is taking a lot of implementation money off the table. Not by banning partners. By making the merchant good enough at the work partners used to monopolise. Agencies that still lead with “we will build your theme” are going to spend the next year discounting against a free collaborator that remembers the merchant’s taste and edits every template in one pass.

The partners who stay billable are being pushed into one of two places.

Master the AI layer and sell that. Not a vague “we use AI” line on the homepage. Actual services: setting up agent workflows, reviewing what Sidekick generated, wiring apps so Sidekick and other tools can query them, and stopping a store from shipping a pretty theme that cannot take payments, track inventory, or survive a sale. Merchants will generate more code than they can judge. Someone still has to say whether it should go live.

Or join the business as a fractional developer or marketer. Embedded, ongoing, close to revenue. Not a six-week theme project. The gap is where the model is confident and wrong: architecture, refactoring a pile of generated Liquid and app code into something you can own, and the broader strategy of what to build at all. AI is genuinely not there yet on those. It will draft the section. It will not decide whether the brand should be a content engine, a wholesale motion, or a waitlist-led drop. It will not unwind three conflicting apps and a checkout rule that only fails on subscription renewals.

That is a smaller market than “every store needs a developer.” It is also the market that still has margin, because the cost of being wrong is a broken catalog or a quiet revenue leak, not an ugly homepage.

What I would stop selling

I would stop selling the illusion that a partner’s edge is faster clicking in the theme editor. Sidekick is already faster at that, and it will get less wrong.

I would also stop telling junior developers that a portfolio of custom sections is a career. The platform just productised that portfolio. The durable skills are the ones the agent still hands back to a human: data model, failure modes, billing and auth, performance under a promotion, and whether the generated thing matches how the merchant actually makes money.

We have written before about agentic coding brain fog: speed without comprehension. Canvas scales that risk from one engineer’s repo to every merchant who can type a prompt. The mess will not look like a blank store. It will look like a finished store that cannot be maintained, cannot take an app, and cannot be explained when something breaks on a Friday. That cleanup is real work. It is not the same work as the original build, and it will not be sold at the original volume.

So is Shopify helping or hindering?

Helping merchants, clearly. Hindering the partner ecosystem as it has been structured, also clearly.

Shopify still needs partners who ship real apps, who run migrations that touch money, and who will sit inside a team when the agent runs out of context. It does not need a large bench of people whose main offer is “I will implement what you can now describe to Sidekick.” If that was the business, the recent changelog is not a feature list. It is a demand forecast.

The move is not to argue the tools back into the box. The move is to stop charging for the layer Shopify just gave away, and to charge for judgment, architecture, and the parts of a Shopify business an agent still cannot be trusted to own.

If you are a merchant staring at a pile of generated theme code, or an agency replanning what you sell, book a consultation. The useful conversation now is which work is still yours, and which work Sidekick already took.