In the 80's, IBM outsourced the OS and the microprocessor of the PC and lost control of this industry to the Wintel alliance. In the 90's, Dell grew to dominate as a PC manufacturer with a strategy defining outsourcing as a competitive advantage.
The "make or buy" decision is a corollary of market failure when customer's economic benefit is positive against the purchase price but turns negative when transaction cost is added. It is usually seen in economic terms focusing upon price, quality, volumes and time saving.
Vertical integration (the buy-side) is beneficial when it increases market power or efficiency. For instance, Schneider Electric and Legrand are able to resist pressure from Rexel, their biggest European distributor, to market own-brand circuit breakers although they have practiced cross branding between each other for a long-time. Rexel is effectively kept out the market except if it builds its own manufacturing operations.
Revealingly, GE entered the European market by acquiring small manufacturers but resisted Rexel's proposal too (despite the much needed boost in volume).
Another way is to gain efficiency by guaranteeing supply, gaining information, acquiring specialized assets or eliminating layered mark-ups (that distorts marginal costs and misplace the supply & demand equilibrium).
Similarly, Air Liquide, a leader in industrial gases, is also France's largest logistics operator. Logistics is kept in-house because it requires specialized equipment and knowledge that Air Liquide cannot outsource at an affordable cost.
Hamel and Prahalad equate strategy with the competition for core competencies: the skills necessary to create or dominate emerging opportunities and shape the structure of future industries. Therefore it is "strategically important" to keep an activity in-house if its outsourcing affects the acquisition of core competencies.
Obviously, a mature technology with many suppliers is not strategic. Strategic subsystems are those for which few suppliers are available...