“Multi-cloud for resilience” is one of the most expensive default decisions a young company can make. It can be the right call — but only for specific, nameable reasons.
What single-cloud actually buys you
One cloud means one identity model, one networking model, one set of managed services your team has to master, and deep discounts as you commit spend. Velocity is high because there is less surface area to learn and operate.
For most companies before serious scale, this is the correct choice, and the resilience story is handled within one provider — multiple regions, multiple availability zones.
Reasons multi-cloud is worth the tax
It earns its cost when there is a concrete driver: a customer or regulator contractually requires provider independence; a specific managed service exists only on one cloud; you need genuine negotiating leverage on a very large bill; or a real regulatory-jurisdiction constraint forces data into a particular provider's region.
“What if AWS goes down” is usually not, on its own, a sufficient reason — multi-region on one provider covers the vast majority of that risk at a fraction of the complexity.
A middle path
You can stay single-cloud for compute and data while keeping DNS, object storage backups, and your CI outside it. That gives you a recovery path and avoids total lock-in without paying the full multi-cloud operational cost. Decide this on purpose, and write down the trigger that would move you.