[rsstack]

First, I appreciate your response (and overall your life work - we're Svelte.js users, about 60% of our source code is in Svelte, and we have a couple of non-critical-path SvelteKit apps).

While Vercel's intentions are good, and your personal intentions are beyond reproach, this strategy all but ensures eventual vendor lock-in without explicitly saying so. Vercel's compatibility will increase over time and other vendors will have best-effort implementations of certain features, but never the complete feature matrix. The industry (esp. medium/large companies) will quickly pick on the relation of "Vercel is needed for ease of mind when using SvelteKit in production", as I've heard from many companies considering/using Next.js. Do people use Next.js without Vercel? Yes. Do companies evaluating Next.js consider it vendor-locked? If they're experienced, yes.

[whatshisface]

The way adapters work in SvelteKit doesn't change very fast, in fact it's stable enough that community-created adapters for things like Google Cloud can keep up just with volunteer effort. I do not think what you're describing will happen.