Startups rarely compete in the same markets. Take an Airline doing 10 billion in revenue that ends up as 250 million a year in profit. Suppose someone discovers an inefficiency say in how they adjust routing based on maintenance scheduling that could save them 1/10th of 1 percent of their operating expenses. Well 10,000 * 1/1000th = 10 million a year in savings based on some investment. Sure, you could try and outsource the whole thing to India, but when every month the system is not up and running they are out close to a million dollars speed becomes important.
Of course that's the upside, the there are a wold of risks and many companies burn millions chasing after penny's.
How about SpaceX? They've already accomplished what no government in the world has been able to do. How about startups emerging in the energy, scientific, and education spaces? It's not all photo sharing, geo-tagging, social engagement rah rah. The discontent is obvious on HN over real problems not being addressed and companies are now beginning to emerge. Not only is it socially pragmatic, but as you so clearly detailed, it's mind-numbingly lucrative.
If you perform a technical service encumbered with vast inefficiencies, your lunch will be eaten in time. It's not a matter of if but when. Just because there are all sorts of barriers to entry right now because of government contracts, regulations, etc, that allow you to offer a solution right now that is cost effective because it's less messed up than the system you address, does not guarantee those walls will not be chipped away, bit by bit, by more innovative, smarter, cost effective ways of doing things.
SpaceX is creating a new system from scratch. But, at the end of the day they are not a software company so in 10 or 20 years you can expect them to be hiring the same consulting company's for internal projects that Lockheed Martin uses.
In the consulting world product is irrelevant you serve large organizations and the problems they develop over time. While they are young Google, Facebook, and SpaceX don't have a lot of the internal cruft that consultants exist to deal with, but I can guarantee they are creating it right now.
Amazon EC2 and GMail are probably the best examples of direct completion with traditional consultants by start-ups. But again it's all about the organization and and not what the organization produces and non of them started by focusing on that stuff. All large companies need email and don't necessarily know how to make is safe, secure, and salable.
PS: By software company, I am talking about mindset. GM creates a lot of software, but it's not what upper management focuses on day to day.
Of course that's the upside, the there are a wold of risks and many companies burn millions chasing after penny's.