It's a plan that probably works for enterprise middleware, but for the majority of reality that is not covered by such a descriptor, it is significantly harder to find a path this way.
Unless you're an Oracle consultant on one end of the scale, or a WordPress theme designer on the other end, you're probably not going to be able to repeatably find lots of clients who want a little of your time and are willing to pay enough for you to work on the side.
A lot of clients want to hire consultants for full-time work, and they are frequently only looking for bodies to fill chairs. Finding a decent client that treats you with respect and understands they can't monopolize your time is hard, hard work.
So if your R&D effort is not related to your consulting--or your contract terms assign ownership of all work materials to the client--you'll just end up in the same situation as having a job.
I suspect the only reasonably repeatable path for early funding for big R&D efforts is through obtaining grants. Learn grant writing, or pay someone who does do grant writing (imagine the longest paper you wrote in college and quadruple it). There are a lot of research grants available, even for independent people not associated with accredited academic research environments. There are even grants for work in the arts.
I've spent the last two years doing freelance consulting. The money I've made was great, but I'm no closer now to starting a company than I was before. You should only do consulting if consulting is what you want to do.
Outside of America this is basically how you have to do it pre-traction.
You could say "I want to build colonies on Mars, I just need 10 Trillion in capital!" or you could say "I want to build colonies on Mars, how can I get there even though I only have 1 / 100,000th of the capital?
The same goes for all these startups that think they need 2 million or 10 million dollars to start. You need skills and $100 to start. Someone mentioned distributed systems as an "unbootstrapable product" which is total baloney. Start by identifying clients that would buy your product once it exists and help them distribute their systems. Build up your team and use your spare cycles to build out the features you think you're going to need but currently can't find anyone to pay you to build them.
If you are still hung up on Space X starting with $100m, let me ask you this. Say Elon had lost it all on Telsa and he was starting with a loan from one of his friends to start Space X. How small would the loan have to be for you to confidently say that Space X would never get people on Mars? To me, the size of the loan only dictates the speed of reaching the ultimate goal.
More importantly, the goal of SpaceX is fundamentally the colonization of Mars, and the idealized method to do that is a fully reusable interplanetary spacecraft. If SpaceX had tried to build that and only that from their start they would have rapidly ran out of money, even $100 million, and had little to show for it. Instead they bootstrapped their way toward acquiring the capability to design, build, and operate such things while also building a corporation capable of funding such operations. Initially they built a small orbital launch vehicle, just barely at the limit of market feasibility, using very conventional technologies and very conventional designs, applying innovation peace-meal where it made sense. Even then it took them several tries to get a launch vehicle that worked, and it wasn't even very competitive in the launch market. But it proved that SpaceX was capable of building launch vehicles, which enabled them to get grants and contracts to build a launch vehicle that was.
It wasn't a demo or a prototype, the Falcon 1 was a fully functional "MVP". And it wasn't investment funding that got them to the next stage, it was effectively pre-orders for the next generation of their product. Then you look at what they've done since, they've built iteratively, advancing toward their goal, but even now they still haven't gotten everything into place that they need. They're still working on partial reusability, still working on manned spacecraft, still working on clustered core staging, and so on. But even so they have viable products that are bringing in a lot of revenue, with new products coming online soon that will bring in even more revenue. They are still several iterations and many years away from their goal but by the time they get there they'll have the expertise, experience, and revenue to actually make it happen.
As with most endeavors, it's almost always easier and better to get to a big goal iteratively. That's how you build skill, it enables you to sell your earlier iterations to maintain revenue to keep your company alive, and each iteration provides lessons which inform the next and can reshape the grand idea you once had.