Scale EconomicsWide moat

Underneath the cross-sell sits the brute economics of scale, which in cloud computing is a genuine and widening moat. Only a small handful of companies on earth can afford to build data centers on the required scale, to buy computer chips and electricity in the quantities that drive unit costs down, and to run the whole apparatus reliably across the globe. A new entrant cannot cheaply replicate either the physical plant or the years of accumulated operational know-how.

Scale buys more than lower prices. It buys the ability to place data centers in every region a global customer needs, to offer a breadth of services no small provider can match, and to absorb the cost of the security and compliance certifications that regulated buyers demand. Each of these is a fixed cost spread over an enormous base, and the bigger the base, the thinner the cost per customer.

The capital requirement is itself the barrier. Building a competitive cloud now costs tens of billions of dollars a year in data centers, chips, and power — a sum that narrows the field, by the sheer weight of money, to a few titans, of which Microsoft is firmly one. This is a moat dug with capital that very few possess.

For the owner, scale economics is the reassuring kind of advantage that compounds with success: the more Microsoft sells, the lower its unit costs, the better it can price or reinvest, and the more it sells. It is the same low-cost-producer logic that protects the best businesses in any industry, applied here to the plumbing of the digital economy — plumbing on which Microsoft spent $115.9 billion last year, against $44.5 billion two years earlier¹.