Ok, in any standard supply-demand analysis, you will find that if a company is attaining non-zero profits, in a competitive environment, another company WILL sprout up out of nowhere to snatch a percentage of that, possibly by selling the same products at a lower cost. This reduces the sum total profits attainable by both companies. The reduced value goes into the pool of money that customers get to KEEP.
However! There is a catch to that. It's called barriers to entry. Examples include: patents that the first company holds that are required to produce the product, infrastructure required to deliver the product (think ISPs' fiber, cell towers), and marketability (if you wanted to sell an OS in the mid 1990s, your company name better be Microsoft). There is a huge (debatable in terms of fairness) advantage the market gives to the FIRST guy to succeed in a certain line of business.
So in your examples, it's not in 100% of the cases in small businesses, they give up profit out of a sense of misplaced altruism. It's simply that they don't have anything that can command a higher profit. For companies that rise up again and again in the news (Microsoft, Apple, AT&T, Verizon, the list goes on), they DO have the chips to play. They understand that barriers to entry prevent Joe Shmoe from writing his own OS or making his own smart phone or building his own cell network and compete successfully with them. They understand they are not in a perfectly competitive environment and can thus act monopolistically or oligopolistically.
This is all not to say that analysis involving assumptions of a competitive environment is purely academic and impractical. In this case, the best action for any governing body to do is to try to increase the competitive potential of the market. This is why they allow certain mergers. The standard case is when two small companies merge to better compete against a bigger one (see Thompson-Reuter vis-a-vis Bloomberg). They will/should never allow a merger that is purely or mostly for eliminating competition (which is the case that was leaked here).
Machines certainly can solve problems, store information, correlate, and play games -- but not with pleasure. -- Leo Rosten