Energy costs are a win for well-run datacenters. It’s quite right to point out, with varying degrees of accuracy, that datacenters use a lot of electricity, or that energy dominates run-time datacenter costs. It’s entirely different, and in fact incorrect, to conclude that rising energy costs should or will dampen datacenter growth and consolidation.

At least two distinct arguments make this clear. First, centralized, top-of-the-line datacenter calculation costs less in energy inputs than the obvious alternatives of casual on-premises computing in closeted servers or workstations. Serious computational loads–pharmaceutical proteomic data reduction, bigdata-scale optimization of retail pricing, complex budgeting, and so on–cost only a small fraction as much energy in a modern datacenter when compared to the same processing on a desktop or conventional “back office”. While datacenter operation is secretive and rapidly-changing, and therefore difficult to quantify objectively, all indications are that higher energy costs increase datacenters’ advantages, compared to on-site processing.

Even more than that, rising energy costs make digital processing relatively more valuable. When crude oil’s spot price ascends, organizations have more incentive to:

• host videoconferences rather than fly teams together for meetings;
• further automate supply chains, to wring out inefficiencies and reduce transport costs;
• make more information available to customers remotely, rather than requiring they visit storefronts;
• computerize manufacturing processes to reduce losses;
• and so on.

A fundamental principle of economics is that absolute prices or values are of secondary importance: what determines behavior are relations and alternatives. Yes, energy price rises can wreak havoc in the short term on departmental budgets; as Jeffrey R. Clark concludes in the second article cited above, “[e]nergy is the largest and most universal concern for companies with data centers.” For a rational organization, though, those same boosts in energy prices only raise the long-term comparative advantage of datacenter computing.

While I’ve yet to find a satisfying unified treatment of datacenter finance and strategy, James Hamilton writes sensibly about datacenter designs and costs in analyses such as this one on overall datacenter costs.