Lots of monopoly news, as usual. Paramount had another setback in its merger plans, there was some comical villainy involving Mark Zuckerberg’s yacht, and the state of New Jersey just filed a ground-breaking monopoly lawsuit against Amazon over driver wages.
Before getting to the full round-up, I want to discuss some events this week in the courts that demonstrate the increasing vulnerability of big tech and AI firms. There are three separate ways the position of these firms is weakening. Their legal protections are fraying, their financial position is more fragile, and the politics for them are turning downright toxic.
We’ll start with the legal piece, because that’s the part mostly not being covered by the press. For years, the main thrust of the assault on big tech came via antitrust law, the argument being these companies are monopolies thwarting competitors and extorting customers. But over the past few years, it seemed like the antitrust campaign against big tech ended in failure. One judge ruled that Meta is not a monopolist, and a different judge ruled that Google is a monopolist, but then mandated irrelevant and minor penalties. And there are rumors the DOJ case against Apple may settle.
But it increasingly looks like these legal setbacks were a pause before a much larger assault on their power. It’s just that it may not come through the antitrust channel, but by taking on the tyranny of unregulated algorithms. Here’s what I mean.
Much of the consumer internet-economy operates on top of feeds that draw one’s attention. Meta, TikTok, and YouTube all have high profit margins because they use feeds that require little human curation, and thus have low operating costs. But just because outlays are low for the companies themselves doesn’t mean there aren’t costs, it just means that they are borne by users and communities, in the form of bullying, sex trafficking, addiction, polarization, shorter attention spans, et al. It’s like pollution, where a company pours chemicals into a river, which doesn’t cost the company anything but does poison entire communities downstream.
For decades, tech firms have lobbied to stop states and the Federal government from regulating them. At the same time, they have argued in courts that algorithms are protected by Section 230 of the Communications Decency Act, which prohibits companies from liability for third party behavior. Algorithms are also a form of speech, and thus are shielded from state action by the First Amendment. In cases such as Netchoice vs Moody, tech firms have claimed there is a constitutional prohibition against regulating them.
Two legal decisions this week suggest big tech will soon lose its legal protections. First, in California, Judge James Donato ruled that engagement-based algorithms are not protected by the First Amendment. Second, in New Mexico, Judge Bryan Biedscheid held Meta liable for being a public nuisance in causing a youth mental health crisis through its engagement-based algorithm, including significant sexual exploitation, “sextortion,” and sex trafficking. Together, these cases suggest the high-margin big tech business model of social pollution is ending.
We’ll start with California. On Wednesday, Judge Donato ruled that a state law regulating addictive social media feeds, the Protecting Our Kids from Social Media Addiction Act, can be enforced. Google, Meta, and TikTok sued, arguing that the law is unconstitutional and violates the First Amendment. Their feeds offered to users, they argued, were “expressive.”
Judge Donato disagreed. In his order, he wrote that relying on “algorithms’ number-crunching capabilities” to serve engagement-bait is not an attempt to express an idea, nor is it speech. It is akin to a “dumb pipe,” like a telephone wire, and thus it can be regulated. Content moderation, by contrast, is expressive and thus protected, but content moderation and engagement algorithms can and should be distinguished. Donato is offering one of the more sophisticated opinions fleshing out how the Supreme Court might handle algorithms and the First Amendment, and one that looks quite reasonable.
If regulation of engagement algorithms is constitutional, then what might that look like? In New Mexico, we got a partial answer. Judge Biedscheid has been overseeing a case about the threats of social media to children. It was partly a jury trial, to decide monetary damages on violations of the New Mexico Unfair Practices Act. That was decided months ago, with the jury holding Meta liable. On the second set of claims, whether Meta is a public nuisance, Judge Biedscheid himself had to decide, because he would have to mandate product changes, which juries don’t do. He ruled on Thursday against the company.
During the trial, it came out that Meta executives knew that features such as infinite scrolls, auto play, and algorithmic feeds are addictive and particularly difficult for children to avoid. The company’s services have worsened bullying, eating disorders, suicide, depression, anxiety, sleep deprivation, you name it, all of which Meta executives knew as they were deploying their products.
School officials, law enforcement, social workers, technology experts, even former Meta executives testified about the harm that came to children from these products. The most lurid was that its algorithm facilitated tens of thousands of cases of “technology-facilitated sexual abuse.” As one former Meta executive, noted, “The product is very good at connecting people with interests, and if your interest is little girls, it will be really good at connecting you with little girls.”
Some of the problems involved facilitating trafficking and exploitation, but there was also the notable phenomenon of “sextortion,” where a predator induces a child to send nude pictures and then engages in blackmail. The judge wrote that “Meta’s engagement-optimized platform features recommend the accounts of vulnerable children to adult sexual predators, and vice versa.”
And it’s not just social networking products, it is also the new AI-powered tools. Experts “testified to how Meta’s chatbots can be misused by adults and adolescents to create sexualized interactions, and the dangers of doing so to users under the age of 18.” These claims are not new, indeed many of them came out during the Meta antitrust trial, but Meta’s public relations team and a snotty liberal judge named James Boasberg were able to prevent them from being widely disseminated.
This time, however, Meta had to deal with some real penalties. Judge Biedscheid ruled the “weight of evidence demonstrates that Meta’s platforms are a cause of and substantial contributing factor to the youth mental health crisis in New Mexico.” He even accepted the metaphor of pollution, ruling that “The Court considers Meta’s platforms to be analogous to the factory, the advertising and other content displayed on those platforms to be what is produced by the factory, and the psychological harm to and sexual exploitation of children to be the pollution that must be abated.”
The judge mandated the company pay $567 million into an abatement fund to help victims, adding to the earlier $375 million penalty the jury levied. More importantly, he also issued an injunction forcing a host of changes to the site, such adding humans to deal with complaints and law enforcement, cracking down on adult users who engage in child sexual exploitation, restricting push notifications, and saying Meta cannot allow its chatbots to talk to kids about sexualized materials. That’s a lot of money and effort just for one single state. And it’s likely to be repeated elsewhere, not just in the U.S., but globally, and not just for Meta but TikTok and YouTube.
Basically, the high-margin business model of engagement-bait is genuinely threatened, as legislators and judges are figuring out how to allow some sort of governance of big tech platforms, including forcing them to accept liability for the consequences of AI chatbots.
And that’s not all for the business model of dominant tech firms who thrive on data-intensive methods of advertising. On Wednesday, the Senate held its first ever hearing on surveillance pricing, where companies set prices based on personal data they hold about individual consumers. Republican Senator Josh Hawley oversaw a bipartisan approach titled “Your Data, Their Profit: The Consumer Cost of AI Surveillance Pricing.”
In the hearing, Robert Hedges, Former Chief Data Officer at Visa, described how companies calculate consumer “willingness to pay” scores based on all sorts of personalized factors, and then offer them prices based on those scores. These include when you booked the ticket, whether you’ve already booked a hotel, if you have friends or relatives in the area, how long ago the trip was organized, et al. I went to the hearing, and every Senator who attended, on the right and left, was hostile to surveillance pricing. Multiple states are on the verge of passing laws to ban the practice, as are cities.
Lee Hepner@LeeHepner
Great hearing on a Seattle bill to ban surveillance pricing and to protect bona fide discounts & loyalty programs, with more engaged public comment than I’ve seen in any jurisdiction.
Great work by @MayorofSeattle and Councilmembers @Alexis4Seattle, @CMRobSaka and Dionne Foster!
6:49 PM · Aug 7, 2026 · 939 Views
1 Reply · 3 Reposts · 21 Likes
And then there’s Google. Appeals came in on the Google monopolization case, which concluded late last year with a remedy basically letting the company off scot-free. Judge Amit Mehta decided that, yes, Google is an illegal monopolist in search, but no, he won’t do anything about it because generative artificial intelligence innovations will take care of the problem.
Multiple briefs filed by third party groups (including my organization) pointed out that the judge’s remedy decision has already shown to have failed, with Google’s Gemini and AI Now overviews increasingly dominating the consumer facing AI chatbot market. Even Microsoft weighed in, arguing that one reason its Copilot product is so bad is because of Google’s monopoly.
I am not optimistic about the D.C. Court of Appeals, but the case is continuing to move. If it dies, or Google otherwise escapes, I do think we’ll find other paths to address the problems of big tech.
So that’s the legal position. Then there’s financial fragility. This one’s pretty easy to understand. Big tech firms are spending all of the cash they are generating on AI chips and putting them into data centers. Instead of buying back stock, they are issuing stock and borrowing money. There are plenty of charts showing the phenomenon, here’s one from Yahoo Finance. Bank of America Global Research calls it a “generational transfer in free cash flow” from big tech to chipmakers.
Microsoft, Amazon, Google, and Meta all have monopolies, in business software, cloud computing, various advertising markets, and online commerce, and they are leveraging their cash and data to pivot into an attempt to sell intelligence on demand. Will selling AI services gain back the large investments these companies are making? That’s not clear. But when you have no free cash flow, your financial position is inherently more fragile.
Finally, there’s the toxic politics of big tech. AI is extremely unpopular, with 70% of Americans opposing building data centers in their local areas. Roughly three quarters of Americans think that surveillance pricing should be banned. And data centers are taking center stage in political races across the country, with candidates who promise to halt construction winning everywhere from Wisconsin to Michigan to Kansas.
Americans have decided that technology is now being deployed against their interests. This shift is hard to fathom, so let me try to offer some context of how beloved these giants used to be, so we can understand how much has changed. Just fifteen years ago, when Steve Jobs died in 2011, Occupy Wall Street encampments set up memorials to the billionaire founder of Apple. Yes, the left of the left were strongly pro-big tech. In part, this dynamic hearkens back to a 1970s era political love affair with high-tech firms, but it had a specific flavor in the 2000s and 2010s.
In 2008, Obama used Facebook as a road to a youth-centric progressive future, with story after story about how young people were flocking to Obama on “the internet” and this new-fangled social networking site. A few years later, activists celebrated Twitter and Facebook as unleashing a wave of protests during the Arab Spring, as well as offering tools to fight back against corrupt governments.
Wall Street lost credibility during the financial crisis, but Silicon Valley gained prestige, such that in 2016, it was simply understood that Facebook COO Sheryl Sandberg would be in President Hillary Clinton’s cabinet. Companies like Uber and AirBNB seemed to offer useful services in a “sharing economy,” these were inclusive and progressive institutions that didn’t pollute, while being pro-business enough for the GOP. Or so went the narrative.
That era of good feelings ended with Trump’s election. In 2019-2020, anti-monopolists forced Congressional investigations and antitrust suits. But Americans still broadly had warm feelings about Google, Amazon, and Apple. During the Biden administration, most Democrats tended to be neutral to positive towards big tech, though skeptical enough to allow antitrust action.
On the right, George W. Bush was strongly pro-Google, and during his first term, Trump celebrated big tech’s stock market gains. In his second, Trump’s policy framework is organized entirely around promoting U.S. AI firms, with massive favors in trade deals but concessions across the board.
But after 2024, as Silicon Valley titans turned towards Trump with displays of obsequiousness indicating they are political figures, this dynamic shifted. Moreover, the construction of data centers started fostering local political fights over non-disclosure agreements, water rights, and electricity.
In 2026, candidates are for the first time running and winning with explicit anti-big tech and anti-data center platforms.
So do these changes in politics matter? Yes. It takes time for public opinion to be reflected in governance. Elections are in two year cycles, and incumbents, who carry with them old attitudes, still get reelected. The Senate is on staggered six year terms, so it’s impossible to change more than one third of that body at a time. Even after elections happen, there’s crafting bills and compromising on language, passing them into law, issuing regulations, bringing lawsuits to give them meaning, and appealing to define how the courts will organize things. It’s a multi-year process, with lots of moving parts.
But I think we’re starting to see the signs that the pieces are snapping into place to really roll back the dominance of big tech. I started this essay by looking at some legal developments. But the big shifts are really the changing financial situation of big tech firms, as well as the politics, as voters express anger and revulsion towards high tech interests. Without that, there are limits to what kinds of changes can occur. But with the public on your side, well, there really aren’t.
We’re in a strange moment right now, where the stock market is at record highs and the Silicon Valley ecosystem of affluent AI product managers and titans operate oblivious to the increasing political rage towards them. But that won’t last forever. And if there’s an economic downturn, well, these guys are living on borrowed time.
And now, the rest of the monopoly news. There are some fun stories in here. Private equity may actually destroy big law, in an Alien vs Predator style fight, Mark Zuckerberg was booed as his yacht refused to help someone in distress in a comical display of villainy, a judge smacked down the fancy corporate lawyers trying to orchestrate a local TV broadcaster roll-up, and the first serious cracks in the electric utility return on investment scam are appearing.
Also, the new Trump Antitrust chief nominee endorsed price-fixing by corporate landlords, Spotify may have just wrecked podcasting, and it looks like the crypto lobby just failed to achieve its major legislative goal, due to opposition from banks.
You can read all that, and more, after the paywall.
