It’s not that cost will significantly increase or decrease whether you use a monolithic or composable architecture. There are too many factors that affect cost.
What is more predictable is that costs will change. Here are a few scenarios to expect.
Developers will be more effective at solving business problems. Thus, you'll be able to do more with your engineering staff, as you'll lessen the time spent on low-level infrastructure.
But the types of engineers you'll need will change. You’ll need fewer DevOps roles, as you’ll outsource most of that work to the various services within your ecosystem. But you’ll also want more support in wiring up services, which means more frontend engineers. (A “frontend engineer” is a more advanced engineer in the composable world, often requiring working with server-side code.)
The cost of services will increase significantly. You’ll likely end up using a wide array of services, and these vendors will engage with enterprise-level contracts.
This may evolve as companies merge and are acquired in the composable space, but today it means holding several contracts with hefty price tags.
In most cases, services that serve enterprises are value positive, as long as they are solving a problem you are experiencing. The cost of low-level engineering is so expensive that even the hefty price tags of an increased number of services is going to mean higher value results.
Depending how you handle engineering adjustments, the cost of services may be offset by engineers alone. But it also affects your current infrastructure costs.
They will still be there and will still be expensive. But, for example, by using a headless CMS, you can completely remove the database for your CMS, the CMS service, and any servers running the CMS or its database. It may not be a one-to-one tradeoff, but it will likely be a significant reduction.