Every few years, a CTO at a growing company looks at the hosting bills, sees four different vendors (AWS for some things, Vercel for the website, Cloudflare for DNS, a managed Postgres provider for the database) and proposes consolidation: pick one cloud, move everything to it, simplify the operations. The arguments are plausible — one bill, one set of credentials, one set of skills on the team — and the savings sound real until you actually do the math.
Three things usually go wrong. First, the savings are smaller than the migration cost, often by a lot. A six-month engineering project to migrate from Vendor A to Vendor B can cost more than a decade of the price difference between the two. Second, the new vendor's services are not actually as good as the specialized vendors you're consolidating away from. AWS's CDN is not Cloudflare. AWS's serverless edge is not Vercel's. The compromise tax of using AWS for everything is real — small in some places, large in others — and shows up in the experience long after the migration is over.
Third, and most importantly: you've now correlated your failure modes. When your DNS, your CDN, your database, and your application server were on four different vendors, an outage at any one of them was bounded. When they're all on AWS, an AWS regional outage takes your entire stack down at once. The argument "AWS doesn't go down" is one we've heard, and it's worth remembering that we've heard it for fifteen years and AWS has had four or five region-wide outages in that span. The bet isn't that AWS goes down often. The bet is that when it does, your business doesn't.
The case for consolidation is real in two scenarios: when you're so small that the operational overhead of multiple vendors is genuinely painful, or when you're using truly redundant services (say, two CDNs that serve the same purpose). In every other case — most cases — the multi-vendor architecture is doing useful work. It separates failure domains. It uses each vendor for what they're best at. It gives you negotiating leverage. The bill is slightly bigger; the risk is meaningfully smaller. That's usually the right trade.