Abnormal Spends Over $10 Million a Year on Anthropic. It Learned Anthropic Was Suing It — Over a Logo — From a Reporter.

Let us put our cards on the table before we say a word about this one. We build on Claude every day. Claude is the model we partner with, the one we trust to hold against prompt injections when other agents fold, the one whose reasoning runs underneath most of what we ship. So when we say what we are about to say, understand that it comes from someone who wants Anthropic to win, not someone looking for a reason to swing. This is a family critique, not an ambush. And the critique is: this lawsuit is an own-goal.

What happened

On July 1, 2026, Anthropic filed a thirty-one-page complaint in the U.S. District Court for the Northern District of California against Abnormal AI — the cybersecurity company most people still think of as Abnormal Security. The claim is trademark infringement and unfair competition. Anthropic says Abnormal's April 2025 rebrand copied its distinctive "A-slash" logo, its animated logo transitions, and its messaging, and that the two companies now sell AI-flavored enterprise software to overlapping buyers — CISOs, developers, procurement teams — closely enough that customers could be confused about who is who. The remedy Anthropic asks for is not modest. It seeks disgorgement of all of Abnormal's revenues, earnings, profits, and benefits.

That is the part worth reading twice. Not a name change, not a cease-and-desist over a logo — a demand for the money.

The facts that make it awkward

Abnormal's founder and CEO, Evan Reiser, responded in public, on his company blog and on X, and the timeline he lays out is the problem for Anthropic. Abnormal was founded in 2018. That is three years before Anthropic existed. The abnormal.ai domain was registered and in use with real customers back then. The slash wordmark at the center of the complaint was designed by an outside agency in April 2021 — before Claude existed, before most people had heard the word Anthropic. Reiser's argument is simple: you cannot copy something that came after you. He also points out that Anthropic does not hold a registered trademark granting it a monopoly over angular-A or slash designs in the cybersecurity category, and that no customer has ever actually confused a general-purpose model lab with a behavioral-AI email-security vendor.

Now, in fairness to Anthropic — and we are trying to be fair, because this is a family critique — trademark holders have a genuine legal duty to police their marks. A company that lets others drift into its visual identity can weaken or lose the mark entirely, so filing suit is not, by itself, a sign of bad faith. Anthropic's "A-slash" is distinctive and, in two short years, has become genuinely famous. If their lawyers looked at Abnormal's rebrand and saw drift, escalating is defensible on paper.

On paper. Because the rest of it is where the own-goal lives.

You do not learn you are being sued from a reporter

Here is the detail that turns a dry trademark dispute into a story about judgment. Abnormal is not some anonymous infringer. Abnormal is one of Anthropic's largest customers. Reiser says he personally expects to spend around a million dollars on his own Anthropic account this year, and that Abnormal as a company will spend more than ten million. And he says he found out Anthropic was suing him from a reporter. Not from a phone call. Not from a courtesy email. Not from the account team that presumably cashes those checks. From a journalist asking for comment.

There is a version of this where Anthropic picks up the phone first. Where a company that believes it has a real trademark concern calls its ten-million-dollar customer, explains the problem, and tries to resolve it before a public filing that asks for every dollar that customer has ever earned. That version costs Anthropic nothing and preserves the relationship. They did not choose that version. And the gap between the version that existed and the version they chose is the whole tell. It reads less like brand protection and more like a company that has gotten large enough to stop making the courtesy calls.

Why we are writing this at all

Because our credibility depends on saying the same thing whether it flatters our partner or not. We have written, more than once, that Claude held when other agents got fooled — and we meant it, with receipts. That praise is only worth anything if we are also willing to say, plainly, that suing your biggest advocate over a slash that predates your own company, and letting him hear about it from the press, is a bad look and a worse strategy. The partnership does not require us to flinch. It requires the opposite: the people who build on you should be the ones honest enough to tell you when you have swung at the wrong target.

There is a broader thing here too, about an industry where the AI-security vendors and the model labs are increasingly the same set of companies wearing different hats — building on each other, competing with each other, and now suing each other, all while sending each other ten-million-dollar invoices. That knot is going to get tighter, not looser, and how the biggest players behave inside it sets the norm for everyone downstream.

Held to about ninety-five percent, because the docket is young and facts change: a court may yet find Anthropic has a real, narrow trademark point, and we would report that too. But you do not need a verdict to grade the relationship management, and on that, the score is already in. Call your customer before you sue them. Especially the one spending ten million dollars a year telling everyone how much he loves your work.

---

How do AI models see YOUR brand?

AIPM has audited 250+ domains. 15 seconds. Free while still in beta.

Audit your site now → aipmsec.com