The subtitle is a little dramatic, I know. Bear with me…
Part two of two. The first half asked why no one in the media (or culture) has taken up the position, and the culprit we found guilty was booming marketing budgets. This half looks at politics, a different animal. I recommend you read part one, though it is by no means required…

On a calm Thursday morning in 2010, Anthony Kennedy read a decision from the bench that a majority of the country would grow a distaste for, a loathing stronger than what they hold for a lot of what the Court has done in living memory. Citizens United vs Federal Election Commission. It was first a small and rather silly dispute about whether a conservative nonprofit was allowed by law to air a documentary attacking Hillary Clinton inside the window before a primary. The Court had heard arguments, then ordered the case reargued on broader grounds. Or in English, they just realized this is a pretty serious question and something larger needed to be must be contemplated. The precedent-setting case ended 5-4, the Court voted in favor of not limiting independent political spending by corporations, unions and associations, because such spending is speech and the identity of the speaker does not strip it of protection. Six days later, and the president at the time (Obama) openly criticized the ruling in the State of the Union while the justices sat straight in front of him, and Samuel Alito was infamously caught was caught by broadcasters shaking his head and mouthing the words “not true.”
I find three primary arguments to be the most compelling. Buckley v. Valeo in 1976 holds that spending money to disseminate political speech is itself speech, since the speech cannot reach anyone without the expenditure. The next reason holds that a corporation is an association of citizens and that the government cannot ban the speech of citizens on the grounds that they have organized themselves into a company. The other significant reason was that expenditures made independently of a campaign cannot give rise to corruption, because corruption in the constitutional sense means a quid pro quo, and money the candidate does not control and cannot direct cannot purchase a quid.
Alright, whether you are for Citizens United or against it, it is here, and that is the fact of the matter. I won’t directly state whether I am for or against it, that is for you to make out (it is also, incidentally, one of those issues where public opinion is actually quite emphatic and bipartisan in their disapproval. 80% of Americans opposed the decision when it landed, with 85% of Democrats, 81% of independents and 75% of Republicans, and a more recent survey found three quarters of respondents, including two thirds of Republicans, supporting a constitutional amendment to undo it. I mean, 22 states have formally called for an amendment for Christ’s sake. I’m beginning to write another article here, I know. I’ll leave that one for another day).
What followed that decision was a funny little word that apparently was so great we put the word “super” before it. Kidding. A traditional PAC collects limited contributions and hands them to candidates. A super PAC may accept unlimited money from anyone -- individuals, corporations, unions, other PACs -- and may spend it on anything except giving it to a candidate or coordinating with one. Super duper!
Television, digital, mail, texts, polling, and the increasingly baroque business of attacking someone on behalf of an opponent are all responsibilities of super PACs these days. Between 2010 and 2022, super PACs spent approximately $6.4 billion on federal elections. $6.4 billion that would not have existed before that morning in January.
Housewarming gifts
We often forget how recently this new system of political expression came into being. In fact, the tech industry’s housewarming gift to this new paradigm began most notably with crypto, a young industry. Though this was no simple housewarmer. It was like the first guest arriving with a $400 bottle, setting the tone. The 2024 cycle proved that a single issue industry PAC could in fact remove a sitting senator who served as chairman on the committee that had jurisdiction over that same industry. Sherrod Brown, a veteran who had held his seat for eighteen years and chaired the Senate Banking Committee, lost it to one of the most crypto-forward senators ever, Bernie Moreno. Andreessen Horowitz gave more than $23 million into Fairshake and its affiliates. Fairshake in turn spent more than $40 million to remove Sherrod Brown. $40 million is, of course, all but a rounding error against the market cap of the industry that spent it. And their return was massive: both a Senate seat and the sexy attention of all who watched it unfold. In a business sense, it was a no-brainer of an investment. The AI industry is now following suit.
A new independent-expenditure committee registered with the Federal Election Commission on August 15, 2025. Just 10 days later, it named itself Leading the Future (super PACs love their generic names), with more than $100 million already committed. Andreessen Horowitz (hello again) put in $25 million; Greg and Anna Brockman put in a combined $25 million; Perplexity contributed $100,000; Ron Conway sent $500,000 to the affiliated Think Big PAC, and Joe Lonsdale sent $250,000 to American Mission. Think Big PAC and American Mission, two entities being the Democratic facing and Republican facing halves of the exact same operation -- what a creepy idea, isn’t it? Feels almost conspiratorial, like the picture below:

Anyway, the stated purpose was to push for a single national framework for AI and the defeat of anyone in their path who was legislating otherwise. $125 million raised by the end of 25’ and roughly $70 million in cash entering the election year. By spring it was being described as having double that amount in operation, and it has since spent more than $50 million.
Want to see another illusion of free choice? Let’s take a look at its counterweight. In February 2026, Anthropic gave $20 million to Public First Action, a bipartisan nonprofit advocating AI safeguards and opposing preemption, and followed it with another $20 million in July. Dario Amodei gave a million dollars personally, and Anthropic employees gave a combined $2.15 million in a single quarter. The group reports roughly $80 million raised, and it runs a Democratic affiliate and a Republican one, which by now you should recognize as the house style.
Feels a bit weird that the largest donors to the supposed pro-AI regulation PAC are not only an AI company, not only its CEO, but its employees too. Whether we use the term “opposition” or not, let’s be clear about the dynamics here. Anthropic trains frontier models, builds data centers, and it is not hampering the speed of the expansion of its tech out of the goodness of their heart. What it wants is disclosure requirements (users or developers interacting with or deploying their AI models must clearly communicate that users are interacting with artificial intelligence rather than a human), a federal standard with some teeth, and no preemption of state law until that standard exists. These are real disagreements with a16z and OpenAI, sure, but they are disagreements about which rules govern an industry both sides intend to keep rapidly expanding. After all, mandatory safety obligations are cheapest for the company that already built its entire brand on them -- arguably a better return on investment for them than falling in line with GPT & Co.
Is either side of this “battle” truly representing the 70% of voters who do not want the big nasty data center built near their house? You tell me. Seen this way, at least it’s clear that the sector has armed itself heavily, and is playing it very strategically too. But surprisingly, it has remarkably little to show for it.
“What do I exist” — AI Money
Here, AI money, let me help you. You exist to buy one thing: federal preemption of state AI law. The “supreme law of the land,” or a federal statute that overrides the states, so that a company no longer has to answer to 50 different rulemakers, and just needs to follow one. Every nascent industry would kill for federal preemption mainly because it provides some clarity to investors about regulatory horizons.
AI Money, you’re not doing so great. Three swings and three misses so far. Ted Cruz tried to attach a 10 year ban on state AI enforcement to the reconciliation bill in July 2025 and the Senate smashed it to the ground, 99 to 1 -- brutal. Congress then kept it out of the defense authorization. States meanwhile enacted 109 AI laws and 28 data center laws in the first half of 2026, roughly keeping the pace of the year before. Brace yourselves, a lot more are coming.
One race this year (that has been reported on a fair amount) gave us a clear battle between what many call oppositions but what smart PTT readers would call two sides of roughly the same coin. Leading the Future (pro-AI) and Public First (still pro-AI, but with a nice spin to it) spent $27 million dollars on… a democratic primary! Alex Bores, the New York assemblyman who co-wrote the RAISE (Responsible AI Safety and Education) Act, ran for Jerry Nadler’s seat. Bores ended up losing to Micah Lasher, 39 to 35. A funny little detail from this race: the millions spent against the author of one of the most consequential state AI safety laws in the country made virtually no mention of AI regulation.
Correction: a funny telling (not little) detail. Fox found 70% of voters opposed to a data center in their area; Gallup chimed in and put it at 71%, 48% of them strongly opposed, which, funnily enough, is worse than nuclear plants do.

$31 million has gone into political ads mentioning data centers this year, briefly reaching 8% of all broadcast campaign spending against under 1% a year earlier, and more than 99% of it was critical. There is functionally no market for the pro data center ads, and yet the one of the largest and fastest growing concentration of political capital is in that industry exactly. They seriously cannot find one competitive district where their own subject matter is an asset…
I hate “conclusions”
One of my favorite history teachers ever hated traditional conclusions more than anything in the world (shoutout Mr. Herzman). At my school, conclusions were understood as a way to “wrap everything up” and to “restate your thesis again in new words.” That teacher actually made us exercise not writing a conclusion paragraph at all in our essays. Most students (including myself) found this infuriating as we grew up treating the conclusion as the last sprinkle of words that let an essay end poetically and happily ever after. Instead, he told us to write a volta. A volta is a turn in a sonnet that complicates a claim rather than tucking it into bed. Or an envoi, the send off stanza where a poet turns from the subject to address the reader directly. Or even the excursus, a digression traveling somewhere the argument had no room to go. Adorno argued in “The Essay as Form” that an essay should stop when it is finished thinking, and a conclusion that “wraps everything up” is by definition the point at which thinking has stopped.
So here is my attempt at a volta:
The title of this essay carries a misnomer. I have treated the absence of such a leader as failure of sorts, one that our screen addicted generation cannot and will not overcome. But that absence may be why the AI money, both sides of that coin, has not yet gotten its federal preemption. Leaders transform a diffuse grievance into a concrete political demand. But once grievances turn into a centralized concrete political demand, they can be negotiated, funded, diluted, or otherwise incorporated into existing political structures. 833 county-level opposition groups lack a centralized authority through which their collective position can be negotiated or co-opted in any way. We mustn’t forget that fragmentation can be a source of political durability. No single organization possesses the authority to speak or bargain on behalf of the movement as a whole.
Hope you all enjoyed our article. If you’d like to read up on what’s going on in the markets this morning, check out Opening Bell Daily, written by a friend of the newsletter, Phil Rosen: read here.
Hope you enjoyed and see ya later, folks.
—J&E