AI changed the build-versus-buy maths
Buying software used to win the argument before anyone finished making it. Building meant a quote, a timeline nobody believed, and a real chance the thing never shipped. Webflow or Bubble put something in front of customers inside a week. Two years on, both sides of that comparison have moved, and they did not move in the same direction.
What buying was actually buying
The pitch was never the feature list. Any competent team could build a booking form. What the platform sold was certainty: a working version existed the afternoon you signed up, and you could click it. A custom build sold you a PDF and a start date.
That certainty was worth the most when you did not yet know whether the business worked. Burning four months of engineering on a product nobody wanted was the expensive mistake, and no-code was insurance against exactly that. Anyone who took the deal in 2021 made the right call with the information available.
The premium on that insurance was deferred, not waived. You pay it later, in the pricing tier you graduate into and in the parts of the product you cannot change once customers start asking.
Both prices moved
Webflow's pricing page today lists Basic at $15 a month and Premium at $25, billed yearly. The next thing up is a Team platform plan at $2,500 a month on an annual contract, which is where publishing workflows, the site activity log and API, and localization live. That is not a slope. It is a step, and it is a hundred times the Premium price. Nobody migrates over $25 a month. They migrate at the step.
Bubble reads the same way one layer down. Growth costs $209 a month for 250,000 workload units, Team costs $549 for 500,000. You pay 2.6 times as much for twice the capacity, and workload is a metric your users generate rather than one you choose. Both numbers come from the vendors' own pages, checked today.
I do not think either figure makes the platform a bad deal. Twenty-five dollars a month for a marketing site is a bargain and I would tell anyone in that position to stay put and stop reading. The question worth asking is what your bill and your options look like on the far side of the step.
The build side, honestly
The claim that AI made building ten times faster does not survive contact with the evidence. METR ran a randomized controlled trial in 2025 with 16 experienced open source developers across 246 tasks. Allowing AI tools made them 19% slower, while the same developers believed they had been sped up by 20%. That result was real and it embarrassed a lot of confident predictions.
METR rerun the experiment and published an update in February 2026. Among developers who had taken part the first time, the estimate flipped to an 18% speedup, with a confidence interval running from 38% faster to 9% slower. METR is careful to say the signal is unreliable. Google's 2025 DORA report, meanwhile, found 90% of surveyed practitioners now use AI in their work and concluded that it amplifies whatever a team already is, good or bad.
So the honest version is narrower and more useful: for a team that knows its stack, the gap between a decision and a running screen collapsed. We build the riskiest screen first now, in the real framework, against real migrated data, before anyone signs an estimate. When we moved Stay World Class off Webflow and Xano to Next.js, NestJS and Supabase, the Lighthouse score went from 55.91 to 91 out of 100 and largest contentful paint dropped 77%. Those numbers existed as a running staging site well before the invoice did. You are comparing a working thing to a working thing, which is the comparison no-code used to win by default.
When buying still wins
Stay if your site fits inside Premium's limits and always will. A brochure site with a CMS is a solved problem and paying a team to rebuild it is a bad trade at any speed. Stay if the workload curve is flat, because $209 a month costs less than a single contractor month, every month, forever.
Stay, above all, if nobody on your side can own a repository. A migration hands you code, deploys, dependency updates and an on-call question you did not have before. Handing that to a company with no one to catch it swaps a ceiling for an outage.
The signal to move is narrow. A customer asks for something, you know exactly how to build it, and the platform will not let you. When that happens twice in a quarter, run the numbers. Until then the old answer is still the right one.
