I recently moved two sites from GoDaddy's predatory WordPress offering (they were charging $1k a year just for some security add-on) to use Hugo + DecapCMS + AWS Amplify. Decap is a fantastic "good enough" CMS to do anything clients of this size need, the only downside is it takes about 1 minute to deploy any changes. Amplify let's you lock a version of Hugo to use, or bring your own, and it will build and deploy your site on any new commit if your repo is in Github or Gitlab. Both clients are currently billed $0.51 per month, and the only reason it is that high is because Route53 costs $0.50 per month per hosted zone. So both these clients went from paying nearly $3k each year for a WordPress site to paying just over $6 a year for a site with nearly the same functionality and none of the maintenance or security concerns. And once everything is all set up, which honestly is not that hard, the only "tech" they need to know is how to sign into Gitlab, which are the credentials they use to log into their Decap admin.
I mean, yes, that sort of setup is less fragile than someone's clobbered-together homebrewed site, but it still requires a dev to maintain. What happens to your client's site if you or they can't maintain the dev-client relationship anymore, for any reason? They'd have to find another dev willing to take over that setup from you.
That sort of thing sounds great for an agency managing multiple sites running off the same template and framework, but for freelancers, it's still too bespoke to be easily portable between clients, hosts, and other freelancers. If someone came to me with a stack like that (and they have), I'd offer to help them migrate it to a more standard setup like Wordpress or Wix for a one-time cost, after which they would pay the vendor directly. But otherwise I wouldn't want to be responsible for maintaining it, especially for just one or two clients.
It's just way too much setup and maintenance. The AWS setup time would itself cost (in dev hours) a month or two's worth of hosting, and Hugo updates or DecapCMS changes would take even more time. Even if the costs to me were $6 a year, the dev hours required to keep a site like that going would far surpass what it would cost them to just pay $20 a month for a vendor-hosted + managed system.
It also introduces multiple points of failure, and if I were hit by a bus or something went wrong while I was on vacation, they'd have no idea if they need to talk to their web host, their CDN, AWS, DecapCMS, Github, or me... they probably don't need to talk to me at all (nothing I can do about any of those services if they have an outage), but they will have no support outside of me.
I don't have anything against self-hosted setups like that for the right audience — I have many such ones myself — but I think they're way more trouble than they're worth for clients who aren't already web-savvy.
I work for another headless CMS (not Decap) and I frequently have to try to help customers who inherited an old site from another agency who didn't properly explain what a headless system is, and they get really frustrated because they end up having to pay a few hundred dollars to a third-party dev just to add a new article category or whatever. It's the kind of thing that would take them ten minutes on Wix/Squarespace/Wordpress, but requires a dev for a stack like you're recommending, and it'd take anywhere from a few minutes (if it's a common stack, like Next/Astro + Vercel) to several hours/days (anything more than 2-3 years old, especially). That will far, far exceed the time and money it takes to host for several years on any of the standard consumer platforms.
For some of these sites, even the original developer who first made them for the first owner didn't want to take them on again — they knew how much work it would be to update them to a usable state again (but that's usually more Gatsby than Hugo).
I'd be very, very wary of recommending such a stack to anyone who is not working with an agency or is already themselves a developer.