So from what I've read, the returns are immense; the amount of learning they can do from examining their own rocket is the real initial treasure. Their rocket design should advance considerably based on this trove, and I suppose the source of many papers, publications, and public-sourced patents. Reusable rockets seems not quite a glorious enough term? There's a massive dividend any time this actually works, and imagining the future of this tech is joyous.
We can guess how much this will bring down the cost. How much will it bring down the price? Say it's just following the supply/demand curve to get more customers. If landing was reliable and the rockets were durable, what would it really take to scale? If you actually got 10 of these going full-duty. How close can you back-to-back launches? The scale of that operation in terms of engineers to manage the workload... The return-on-automation (in other words, software) is immense the first few billion dollars you spend on it. That means they need to hire a fuck-ton of very good developers. Cool!
But what if it's more of a binary market and the only new customers are only at a much lower price point? In that case it's just serving the existing market at much higher margin. I believe this will make space more accessible -- cost savings ultimately flow through.
BTW, one of the advantages of pledging your patents for free use is you can really show off the right way to write a patent. The whole point is forcing disclosure in return for a benefit, so now make that disclosure top-notch. In theory, what if patent applications were examined by the top-of-the-field peers and only real advancements in the field which were fully and properly documented by the patent would be granted? I hope not all software patents are evil. The "provisional patent" is an interesting form of self-publication at least, but it does create a pesky 1-year ticking clock.
As a talent/recruiting event this is pretty much about as good as it gets.