It was an eventful—and decidedly negative—summer for Silicon Valley. In August the Federal Trade Commission (FTC) approved Amazon’s $13.7 billion acquisition of Whole Foods, not only sending shockwaves through the grocery business but adding to growing anxieties about Amazon’s dominance of the larger retail industry.

Around the same time, New America, a Washington D.C.-based think tank, was engulfed in controversy after an overly public parting of ways with one of its Fellows, Barry Lynn, who had been a leading advocate for expanded antitrust enforcement against tech platforms such as Google, one of New America’s funders. Whether or not Google played a direct role in Lynn’s dismissal, the incident nevertheless raised questions about the political influence of leading Silicon Valley firms.

Meanwhile, the growing trickle of stories about Facebook’s and Twitter’s potential implication in foreign political ad campaigns—and the growing controversies over the proliferation of “fake news”—has erupted into a full-blown crisis of legitimacy. As New York magazine asked last week, “Does Even Mark Zuckerberg Know What Facebook Is?”

Technology platforms are no ordinary corporations. They are privately controlled infrastructure, the underlying backbone for much of our economic, social, and political life.

Amazon. Google. Facebook. Twitter. These are the most powerful and influential tech platforms of the modern economy, and the headlines over the last few weeks underscore the degree to which these firms have accumulated an outsized influence on our economic, political, and social life. To many, including acting FTC Chair Maureen Ohlhausen, the status quo is great: the benefits to consumers—from cheap prices to easy access to information to rapid delivery of goods and services—outweigh greater regulation, lest policymakers undermine Silicon Valley innovation.

But the recent controversies suggest a very different perspective—that private power is increasingly concentrated among a handful of tech platforms, representing a major challenge to the survival of our democracy and the potential for a more dynamic and inclusive economic order. A growing clamor from both the left and right has created a sense of “blood in the water,” and suggests that Silicon Valley’s long honeymoon may finally be over.

Americans have always been distrustful of corporate power, and never more so than where concentrated wealth and economic power threaten to spill over into our political process. This specter of oligarchic influence—whether through outright corruption or through influence-peddling with campaign or financial donations—lies behind the continued calls for campaign finance reform as well as the more recent anxieties about Google’s political influence in Washington. But the power of today’s tech platforms is about more than their lobbying acumen.

The danger of the “platform power” accumulated by Amazon, Google, Facebook, and Twitter arises from their ability to control the foundational infrastructure of our economic, informational, and political life. Even if they didn’t spend a dime on lobbying or influencing elected officials, this power would still pose a grave threat to democracy and economic opportunity. The fact that these companies provide enormously popular and useful goods and services is indisputable—but also beside the point. The central issue here is not simply the value for the consumer. Instead it is vast, unaccountable private power over the foundations of contemporary society and politics. In a word, the central issue is democracy.

What then should we do about twenty-first century platform power? While the technological realities of Amazon, Google, Facebook, and Twitter are distinctly modern, the problems they pose for our economy and our politics are in many ways deeply familiar. They recall the threat that earlier generations faced in the height of the industrial revolution with the rise of corporate giants, which suggests that today’s reformers could learn a lot from the anti-monopoly movement.

•••

In the late nineteenth century, a common concern among both rural Populists and urban Progressives was the rise of concentrated corporate power. The mega-conglomerates in finance (J.P. Morgan), infrastructure (the Vanderbilts’ railroads), and power (John D. Rockefeller’s Standard Oil) represented private control over the foundational goods and services upon which the rest of the economy depended. These firms could not only extract higher prices, they could also employ control over national economic systems to influence commerce to their own advantage.

Amazon, Google, Facebook, and Twitter are the new ‘utilities’ of the modern era.

In the same way, Amazon, Google, Facebook, and Twitter are private entities that have cornered control over the infrastructure of the modern economy—the channels of retail commerce, the systems of information flow, and the avenues of access. As the new railroads and power of the digital economy, it matters little whether these firms live up to their public-minded rhetoric. Google can claim “don’t be evil” as its slogan, but as Louis Brandeis, one of the leading intellectual figures of last century’s anti-monopoly movement, argued:

benevolent absolutism . . . is an absolutism all the same; and it is that which makes the great corporation so dangerous. There develops within the State a state so powerful that the ordinary social and industrial forces existing are insufficient to cope with it.

It was this deeper problem of power—not merely the impacts on prices or the consumer experience—that motivated reformers such as Brandeis to develop whole new institutions and legal regimes: antitrust laws to break up monopolies, public utility regulation to assure fair prices and nondiscrimination on “common carriers” such as railroads, the creation of the FTC itself, and much of President Franklin Roosevelt’s early New Deal push to establish governmental regulatory agencies charged with overseeing finance, market competition, and labor.

But the late twentieth century saw a widespread shift away from the New Deal ethos. Starting in the 1970s, intellectual critiques of economic regulation highlighted the likelihood of corruption, capture, and inefficiency, while scholars in economics espoused the virtues of self-regulation, growth-optimization, and efficient markets. In these intellectual constructs big business and the conservative right found support for their attacks on the New Deal edifice, and in the 1980s and 1990s, we saw the bipartisan adoption of a deregulatory ethic—including in market competition policy.

These cultural currents—the skepticism of government as corrupt at worst and inefficient at best, the belief in private enterprise and the virtues of “free markets,” and a commitment to delivering for consumers above the broader social and political repercussions—suffuses our current political economic discourse. The Brandeis-ian critique of private power has been wholly absent in recent decades and nowhere is this absence more pronounced than in the worldview of Silicon Valley.

•••

In our current moment, it is as if technological innovation has been divorced from the corporations that profit from it. Through these rose-colored glasses, technology is seen as a good in itself, promising efficiency, delivering new wonders to consumers, running laps around otherwise stale and plodding government institutions. Amazon, Google, Facebook, and Twitter have been able to resist corporate criticism (until recently, that is) by emphasizing their cultural and ideational commitment to the consumer and to innovation. They have casted themselves as the vanguards of social progress, the future’s cavalry who should not be constrained by government regulation because they offer a better mode of social order than the government itself.

But as the anxieties of the last few months indicate, this image does not capture reality. Indeed, these technology platforms are not just “innovators,” nor are they ordinary corporations anymore. They are better seen and understood as privately controlled infrastructure, the underlying backbone for much of our economic, social, and political life. Such control and influence brings with it the ability to skew, rig, or otherwise manage these systems—all outside the kinds of checks and balances we would expect to accompany such power.

Democracy cannot persist in an information environment that is designed to maximize advertisers’ reach without regard for the larger political and social consequences.

In terms of shaping our economic life, Amazon is in many ways the modern railroad, the shipping and logistics system that makes our larger retail market function. Like the railroads of yesteryear, Amazon has the power to make or break whole towns and whole industries through its control of this basic infrastructure. The Whole Foods merger must thus be seen in light of this larger role. It does not matter if Amazon is a relatively new player in the grocery market or that Whole Foods is a relatively niche company. Amazon still accounts for a dominant share of the retail and consumer goods sector, and its continued expansion into neighboring goods and services like groceries helps further consolidate and entrench this dominance.

In a similar vein, Google is a vital informational platform whose search algorithm exercises life or death control over the fates of businesses. Yelp, a rival search platform that provides user-based reviews for restaurants and other businesses, has documented how Google’s algorithm deliberately downplays results that direct users to Yelp in favor of Google’s own ratings and business listings. Yelp is a relatively big player among Internet companies, but it has seen first hand the threat of Google’s completely arbitrary, opaque, and unchecked whims. Not surprisingly it has become one of the biggest advocates for antitrust enforcement on Google.

This kind of infrastructural power also explains the myriad concerns about how platforms might taint, skew, or undermine our political system itself—concerns that extend well beyond the ability of these firms to lobby inside the Beltway. Even before the 2016 election, a number of studies and scholars raised the concern that Facebook and Google could swing elections if they wanted to by manipulating their search and feed algorithms. Through subtle and unnoticeable tweaks, these companies could place search results for some political candidates or viewpoints above others, impacting the flow of information enough to influence voters.

Moreover, as Facebook’s growing implication in the Russia scandal indicates, the threat to democracy is compounded by these algorithms’ ability to precisely target readers. Russia’s activities around the election reveal the ways in which any actor—foreign or domestic—can strategically shift political views by deploying sophisticated, subtle, and therefore effective forms of targeted information, misinformation, or obfuscation to influence and motivate sympathetic audiences. As others have noted, there is already widespread use of these techniques among white supremacist or hate groups. And recent reports detail how supporters of President Trump have deployed targeted Facebook ads to assure supporters that the “border wall” is in fact being built, even though it remains in limbo in Congressional budget negotiations. While not on the same scale as Facebook in terms of users, Twitter has faced the same troubling dynamics in its role as a major mediator of information flows.

Despite Facebook’s or Twitter’s public statements these problems are not just the result of bad actors operating on the margins of the platforms. Rather, as Zeynep Tufecki has written, these problems are baked into the business model of online platforms, whose whole economic wealth and value stems from their ability to enable effective, precise targeting by advertisers to influence users. Until recently, the platforms have tended to resist calls for more responsible management of these systems—a form of “absentee ownership”—but it is important to note that they are already in the business of governing speech. In the past, they have issued ham-handed and poorly thought out guidelines that fail to adequately balance the competing values of free speech and preventing incitement, hate speech, or misinformation.

This degree of power belies our conventional tropes about avoiding governmental oversight in favor of the “free market.” The reality is that our economy and politics are already governed and already regulated. They are governed by the opaque judgments of Amazon, Google, Facebook, and Twitter—judgments that are not subject to the mechanisms for representation, participation, or accountability that we would expect of similarly empowered governmental bodies.

•••

Given our reality, it would be helpful to think of Amazon, Google, Facebook, and Twitter as the new “utilities” of the modern era. Today the idea of public utility” conjures images of rate regulation and electric utility bureaucracies. But for Progressive Era reformers, public utility was a broad concept that, at its heart, was about creating regulations to ensure adequate checks and balances on private actors who had come to control the basic necessities of life, from telecommunications to transit to water. This historical tradition helps us identify what kinds of private power are especially troubling. The problem, ultimately, is not just raw “bigness,” or market capitalization. Rather, the central concern is about private control over infrastructure.

In today’s world we must start thinking about infrastructure as more than roads and bridges; instead, infrastructure are those goods and services that enable a wide range of downstream uses. The more critical the good or service might be to individual and community well-being, the more vulnerable end users are to potentially arbitrary limitations, cut-offs, or discrimination in access to those goods. Thus, like the railroads or the telegraph of the nineteenth century, these firms present not just the specter of private power, but the specter of private power over the foundational infrastructure of modern society.

The central issue is vast private power over the foundations of contemporary society and politics. In a word, the central issue is democracy.

As I’ve written elsewhere, updating historical public utility concepts can help illuminate our current dilemma and offer a roadmap for reigning in platforms and their control over twenty-first century infrastructure.

First, there is the antitrust solution, wherein monopoly concentrations are simply broken up into smaller competing firms—such as the breakup of AT&T under the Sherman Antitrust Act in 1982. Using antitrust laws the government could subject mergers and acquisitions involving the dominant platforms to greater scrutiny—deals like the Amazon–Whole Foods merger. But it is still not entirely clear what it might mean to “break up” Google or Facebook.

One option would be to impose firewalls between different lines of business to prevent self-dealing. In this view, if Amazon controls the retail and shipping infrastructure for most consumer goods, it ought not also be able to sell its own goods on that same platform. Indeed, one of the more troubling expressions of Amazon’s power stems from how Amazon is able to undercut third-party sellers by introducing its own variations of goods that are doing well on the platform.

A second solution could involve expanded federal regulatory oversight to assure nondiscrimination, fair and equal access, consumer protection, and basic minimum standards of quality. Net neutrality is perhaps the best current example of this kind of public oversight since it seeks to prevent internet service providers from slowing or blocking access to some kinds of content and speeding up access to other content in exchange for payments from content providers. For information platforms, we can imagine a regime of “search neutrality,” which would emphasize standards for delisting or de-indexing search results and would regulate “stealth marketing” wherein seemingly neutral search results are in fact prioritizing companies that have paid.

This kind of standard-setting could help shape the emerging efforts to govern fake news. As Google, Facebook, and Twitter play a more explicit editorial and curatorial role, real questions arise about collateral damage to speech, governance, and democracy itself. The stakes are too high to leave these questions solely in the hands of private actors. Instead, these challenges call for a “pluralist” model of regulating speech online—one where a complex interplay between government regulation, platforms, and end users works together to find consensus.

Lastly we might create “public options,” publicly run alternatives to private infrastructure. Creating a publicly run alternative is not always feasible or desirable, but like these other strategies, a public option can help ensure the public platform focuses on the core values and goods being delivered—and, by competing alongside private actors, it can help set new standards.

None of these solutions are mutually exclusive, nor are our challenges limited to Amazon, Google, Facebook, or Twitter. The massive data breach at the credit reporting company Equifax raises a compelling example for this way of thinking about platform power. For better or worse, credit reporting is now an infrastructural good in the sense that individual credit scores have come to play a massively determinative role in one’s economic prospects. Yet credit scores in the United States are provided by an oligopoly of companies—companies that have developed primarily to serve not the public but the firms who pay for information on who to sell or lend to. Thus we are left with the perfect storm: an infrastructural power with misaligned incentives and a lack of accountability.

At a minimum Equifax’s data breach suggests a need for regulatory oversight imposing public obligations of data security, safety, and consumer protection on these firms. Some commentators have suggested an antitrust-style breaking up of credit reporting agencies while others have called for replacing the oligopoly altogether with public databases.

A century ago, the threat of the concentrated private power of railroads, financiers, telecommunications magnates, and other monopolists spurred a wave of innovation that gave rise to antitrust laws, public utility regulations, and many of the proposals that would become the foundation of Roosevelt’s New Deal. We face a similar set of challenges today, and like Brandeis, we must not lose sight of the stakes. This is not just another conventional story of left–right battles between governmental regulation and corporate power, after all.

At stake are much larger values of free and fair political speech, free and fair markets, and free and fair democracy. Businesses, towns, communities, and individuals cannot thrive in an economy where access to the market is contingent on the whims and control of a few private actors. Nor can our democracy persist in an information environment that is designed to maximize advertisers’ reach without regard for the larger political and social consequences. The future of our politics and our economy rests on our ability to engage with these foundational challenges.