Taintsville Take · Monday Afternoon, October 26, 2026

Taintsville Take: Five Things Carl Learned From The AI Company’s IPO Filing

Anthropic filed to go public this morning. Anthropic builds artificial intelligence, and it’s about as serious a company as there is. It writes long essays about the future of humanity, it hired a trust to look after the long-term good of the species, and now it wants your retirement account.

Carl printed the coverage at the hardware store and read it aloud at the counter while I was there, giggling like a schoolgirl the whole read. He’s careful to say he isn’t against the company and isn’t for it either. He’s against legions of corporate folk and legal disclosures that take themselves too seriously. By the end, three customers had pulled up stools and everyone was laughing.

Here were the ones that had everyone in stitches:

1. They named the trading ticker after a bug that bites.

The ticker is ANT. This is a company that might be worth $2 trillion, with some of the smartest people on the planet working for it, and somebody in a glass conference room looked at those three letters and said, “Perfect.”

Nobody in that room has ever had fire ants. Down here an ant is no mascot. It’s a Saturday disaster. Ants come up through the sidewalk crack, chew your tomato plants, send your nephew to urgent care and turn a backyard barbecue into a dance recital. Carl sells four kinds of ant killer, and every one of them sells out at some point in the summer.

You can tell this name was picked in a city where the only ant anybody sees is in a Pixar movie. In San Francisco an ant is “small, hardworking, collaborative.” In Taintsville an ant is the reason a kids leg swelled up like a pool noodle at the church picnic. Carl’s advice to the investor relations team: come down here in July, step on a mound once, and then tell us about branding.

2. They lost $72 billion and made it sound like a compliment.

As reported, the company’s first-half loss was $80.7 billion. Most of it, $72.4 billion, is an accounting charge on its convertible notes, and the charge grew because the company’s own value went up. The actual operating loss was $693 million, which is still a lot of deck screws.

So the headline number translates to: we lost the gross domestic product of a mid-sized country because we’re so darn popular. Carl has heard a lot of spin in his life, mostly from car dealers, but this is the first time he’s watched somebody turn a loss that size into a humblebrag.

Carl tried it on his wife. He told her the truck repair bill came out so high because the truck was so valuable. She said that was the dumbest thing she’d heard all week, but maybe that PR firm has a job for Carl.

3. You can buy a ticket, but you can’t drive the bus.

Here’s the setup, per the filing. Regular shareholders get one vote per share. The seven co-founders get one special share that carries 50.1% of the vote on most matters. Then a trust of five people, with no money in the game, picks a majority of the board. The filing itself warns this “may effectively eliminate” the ability of ordinary shareholders to influence big decisions.

Whether that’s good or bad depends on who’s asking. If you think a company building powerful AI shouldn’t be steered by whoever owns the most shares on a Tuesday, it’s a guardrail. If you’re the person who just bought the shares, it’s a suggestion box with a slot that leads to a shredder.

Carl says he’s lived under this arrangement for thirty-one years. Every Friday his wife asks where he’d like to eat. He gets one vote. She has the Class F share. They’re going to the Olive Garden.

It’s also the homeowners association, Florida’s other state sport. You pay the dues, you get a vote, and five people who don’t live on your street decide what color your mailbox, house and hedges are allowed to be. We buy a bunch of paint from Carl…

4. Your partners, your landlords and your competitors are the same three guys.

The company reportedly plans to spend upward of $450 billion on cloud computing, much of it with Amazon and Google. Amazon and Google are also big investors in the company. Google also builds its own AI that competes with it. So the investors put money in, the money goes back out as rent to the same investors, and one of the landlords is trying to open a store next door.

Nobody’s saying anything shady is going on. The problem is what happens on the day something breaks. When your banker, your landlord, your supplier and your rival are all the same people, who do you call? It’s like Thanksgiving dinner where your brother-in-law is also your mechanic, your lawyer and the guy who backed into your mailbox. Everybody’s family, everybody’s owed something, and nobody wants to be the one to bring it up over the sweet potatoes.

Carl’s had a version of this. His cousin sold him the boat, financed the boat, stores the boat and borrows the boat. When the motor died last spring, they spent three weeks figuring out whose fault it was. The boat’s still sitting there. So is the cousin.

5. The warning label reads like the back cover of a sci-fi paperback.

Every IPO has a risk section. Most of them say things like “competition may intensify” and “interest rates may rise.” This one reportedly says the models could develop capabilities the company doesn’t expect, and lays out ways they could cause real harm. Investors have read a lot of risk factors. This might be the first one that reads like a movie trailer.

Carl’s lawn mower manual warns against mowing barefoot. His chainsaw manual warns against holding the wrong end. This manual warns that the product might get ideas. That’s a whole different shelf at Barnes and Noble.

To be fair, the actual AI Carl has met so far is less Murderbot and more confused intern. His grandson asked one to plan a fishing trip, and it booked a lake two states away and recommended bringing a tennis racket. His phone’s assistant still can’t hear the difference between “call Marge” and “play Barge.” If the machines take over, Carl figures they’ll start with his autocorrect, which has been calling the shots on his text messages since 2019.

Our Verdict

Carl gets the last word of the day. “It’s a serious company doing a serious thing, and I respect that. But anybody who names themselves after a bug, loses $72 billion on purpose and puts a robot apocalypse in the fine print either has a better sense of humor than they’re letting on, or doesn’t care what anybody thinks of them. I’m not buying any until I figure out which. I’m buying ant spray instead. That’s a sure thing.”

For what it’s worth, I think the company and its technology are revolutionary. But it’s hard for me to take a San Francisco company very seriously when it wants full control and doesn’t seem to care what anybody else thinks. At least not until I see where it prices at the IPO.

Brad Hoppmann

P.S. You’ve seen how we read an IPO filing. You can see the same work on future IPOs and the whole market. A year of Sector Cycle Radar is $79, and it includes:

  • the full sector and stock analysis behind every gauge

  • a recommendation each month

  • a portfolio update each week

  • an options idea every month or two, when the market gives us one worth taking

That’s about 22 cents a day, or less than one can of the good ant spray a month. We will send out more information on the product in the next week as we set everything up for folks to upgrade their subscriptions. If you want to just keep the opening marketing overview with out the sector analysis, that’s OK too you don’t have to do anything in that case.

Keep Reading