Like all such things, the expensive part is the number of manual steps required. So the answer is to remove the human element with
The various monitors are reasonably pricy, sure, but any large hospital has dozens of them. Your facility can certainly afford them. And investments in equipment can be amortised over time.
The other steps currently require human involvement though. Most monitoring needs a human specialist to look at scans, or take blood samples, or transfer the fertilised embryo between various solutions, and so on. Much of this is necessary because there is no integrated facility which is able to carry out all the relevant steps.
So, your lower-cost facility needs to be able to do all this in one place. And more than that, it has to be constructed such that every step in the process is capable of automation. Containers for embryos must be standardised. Equipment to take the various scans or to transfer embryos between containers must be able to do this automatically. And the process of assessing scans must be automated so that the "sunny day path" needs no human involvement. Of course people will need to get involved when things go wrong, but the more work goes into this, the less likely it is that people will be needed.
Because a company won't fully trust automation until it's been validated, your facility will end up being more reliable than "fallible" doctors. The small risk associated with amniocentesis and other procedures simply won't exist here. And developmental abnormalities linked to trauma, diet, or just random bad luck will become a thing of the past.
At this point, artificial wombs stop being a rich person's plaything, and become best practise for childbirth. Of course some people will still want natural births. This will be a lifestyle choice - some people will prefer it because they don't like the medicalisation of childbirth, in the same way as some people today choose to give birth at home. Like home birth it will be a choice with a slightly higher risk to mother and child, of course. It may cost more, in the same way that neonatal units today cost more than doing nothing, but if it's best practise then it will be adopted.
For one example of how this could be better, it is entirely possible that we could find babies which spend longer in the "womb" have better outcomes, for similar reasons to why babies which are born prematurely have worse outcomes. The impacts of premature birth are lifelong, covering their ongoing development, health and possibly even educational achievements (evidence currently disputed). If artificial wombs swing this the other way, we could see an epigenetic step-change in human health, intellect, and many other key markers.
The automated nature of this will make it a much "colder" experience though. The system would be set up very much on the lines of Amazon's warehouses, where the fertilised embryo goes in one end of a conveyor belt and a fully-developed foetus comes out the other end after 9 months. There will be monitoring, of course, but it will be a very hands-off setup. Deliberately so, because if the automation is good enough then human involvement can only lead to problems.
Of course, operators will want to avoid this appearing like The Matrix's "Fetus Fields"...