To some policymakers, everything done by a leading tech company screams “monopoly.” While no industry should be considered above fair criticism, the pattern is by now familiar. Data center outage? “Monopoly.” Data centers using electricity? “Monopoly.” Early success in a booming industry? “Monopoly.” Rather than getting to the heart of the debate over tensions in technology policy or other areas, policymakers and pundits often rush straight to the field of antitrust.
To quote Inigo Montoya, “You keep using that word. I do not think it means what you think it means.”
The discourse around data centers is the latest iteration of the “Big Tech” monopoly fallacy. Sen. Elizabeth Warren’s call once again to “break up Big Tech” rested on the claim that Amazon had broken “the entire internet,” an assertion perhaps at odds with reality, as she had made four posts over the course of that day while the internet was technically “broken.” Within the same week, a Wall Street Journal headline offered that data centers’ electricity use may become the next antitrust issue. But the use of the term monopoly rather than a deeper examination of a complicated issue is only indicative of a “big is bad” mentality that fails to engage in the actual policy issues holding back American companies from continuing to innovate and compete.
Are data centers and their failures providing evidence that Big Tech needs to be broken up? To answer that question, we first need to understand what happened with Amazon Web Services’ (AWS) outage – and whether antitrust intervention would have averted it.
The AWS outage is tied to the core software infrastructure of Amazon’s data centers in Northern Virginia, aka us-east-1. DynamoDB, AWS’s core database system for maintaining and accessing client data, bugged and briefly rendered all data stored in us-east-1 inaccessible. As a result, many online services ran into DNS resolution errors and were briefly non-functional. The us-east-1 outage was so impactful because the size, cost-efficiency, and quality advantages of AWS data centers in Northern Virginia resulted in a lot of companies choosing to rely solely on us-east-1 to host their services.
The risks of this type of singular reliance are well established. Indeed, AWS explicitly outlines several network architecture strategies for its clients to implement in the event of an outage. Many companies decided that it was more cost-effective to risk their application going down with the region than to invest in mitigating this contingency. This is a basic business trade-off, and cloud companies such as Oracle, IBM, and Microsoft offer other options for those who prefer a different choice. Companies such as Dropbox have taken these steps when they are dissatisfied with their service.
If we had simply done as Warren and others desire, and broken up Amazon through changing antitrust regulation and punitive court rulings, would things have been different? Almost certainly not.
The underlying issues are not tied to the competitive factors antitrust is designed to address. Even if breaking up Amazon would have prevented us-east-1 from going down, the ramifications would be catastrophic for the internet. Cloud computing has driven down the cost of storage overall compared to the infrastructure and other startup costs needed in server era. As a result, while the expenditures to set up cloud services are costly, the resulting decrease in costs benefits consumers as well as the companies that provide these services.
Data centers are both critical to the burgeoning artificial intelligence industry and a growing flashpoint in the debate over energy sustainability and use. Yet, like some horrible crossover episode no one asked for, antitrust appears to be joining the discussion.
Antitrust is designed for a specific purpose – preventing illegal monopolies from harming consumers’ benefits from a competitive market – and opening it up to concerns unrelated to competition, such as environmental preservation or energy usage, could result in significant consequences for the future use of this powerful tool in the broader economy. Such erroneous interpretations of antitrust law shift focus away from the objective nature of the consumer welfare standard by instead allowing policymakers to mischaracterize market conditions in an attempt to score political points. If there are actual policy concerns related to data centers’ energy use or other tech issues, the appropriate policy tools should be used for the job, not the antitrust sledgehammer.
For some lawmakers, “successful firm” and “monopoly” have become synonyms, but this is a dangerous policy recipe. Shifting away from our existing standards to follow political motivations against successful companies ultimately harms the consumer, as well. Policymakers would do better to remember more precisely what “monopoly” means before rewriting the rules that have long powered American innovation.
