Edit: see 'Theory of the Firm'.
In the same way that a data center has full redundancy for failure of an individual machine, but limited redundancy for failure of the city's power grid, all we _really_ need is enough redundancy to give us time to solve the problem.
Large organisations move too slowly for this to be an option, so they need the much more expensive full redundancy.
After almost two decades of professional experience (including a few years of consulting), I'm becoming more convinced that employees often make a similar 'lazy' decision when they choose full time employment. That decision is only optimal for under performers (and only for as long as they can hold on to the gig).
Selling labor in units of time offers no possibility for economic profit. The value of any productivity gains are realized by the employer. This is not the case for a business (or a consultant), where productivity gains yield profit.
EDIT: Consultants sell time too, but are free to adjust their rates as they see fit. Employees have significant barriers to changing their 'price'; infact it often involves a job change. Another challenge for the employee is demonstrating their value created due to the extremely complex structuring of most companies (this makes it hard to objectively justify a raise).