Errr, RS-68 shared a lot of infrastructure with RS-25. That went away. Then the contract negotiated right after the dissolution of the URSS needed to be extended and Energomash was in a much better position to negotiate. That explains most of first stage propulsion.
In SpaceX case, what makes sense for building in house is that they have core competencies in doing space related manufacturing. They have the experience, expertise, software and process control for the whole life cycle of space qualified products. In many cases, implementing that system is more expensive than the design and tooling of the part. Thus,while there isn't a big market for space qualified products, it will make sense for SpaceX to do a lot in house.
SpaceX also has the benefit of designing both a vehicle and a plant from a clean sheet. The rocket and its parts could be designed for modern manufacturing since there were no designs already existing. Existing designs can complete development faster (and therefore at lower development cost) but possibly at the expense of recurring cost. Having no build tools means SOMETHING has to be bought, so the company can optimize for tools that can fabricate a lot of different things. Then again, ULA Decatur wasn't there 18 years ago.
You are forgetting SpaceX "sister" companies.
Tesla - Electric vehicle, electro-motive, battery and very high-tech mass production skills it can share with SpaceX
SolarCity - experience in solar power generation, storage and distribution, sourcing and experimenting with cells from a wide variety of commercial sources. Lots of this applies to rockets, sats and bases.
Finance - Elon is not a "banker" anymore but he definitely knows the movers and shakers and ins and outs after creating and owning
Pay Pal. - a good knowledge of multiple sources and sinks of capital never hurts
Zip2 - This product let Elon explore using computers to map complex systems in unique ways (in Zip2's case cities)
Then there are his Hi-Tech investors and what they may or may not be bringing to the table. (Not just VCs)
I would consider it more appropriate to consider a profile of Musk's firms to be better matched to a Japanese Keiretsu than typical Western firms. The nearest US match would have been the Motorola combine under Galvin Sr in the 80s and 90s.
http://en.wikipedia.org/wiki/Keiretsu