In a move that has left investors scratching their heads and AI developers nervously checking their firewalls, Irregular, the AI security startup, has somehow convinced a group of very rich people to part with $80 million. That's right, folks—$80 million to secure "frontier AI models," which, for those not in the know, is tech-speak for "AI that's so advanced, it might accidentally solve world hunger or turn your toaster into a sentient being that demands voting rights."

According to a source close to the deal who probably signed an NDA so tight they can't even whisper in their sleep, Irregular's valuation has now skyrocketed to a cool $450 million. That's almost half a billion dollars for a company whose main job is to make sure the robots don't get too big for their britches. Talk about a growth spurt! It's like watching a toddler with a trust fund—adorable, but also slightly terrifying.

So, what exactly does Irregular do with all that cash? Well, they're focusing on securing AI models that are on the "frontier," which apparently means they're so new and untested that they might as well be labeled "handle with care: may contain sentience." Using cutting-edge techniques like adversarial training and red-teaming (no, not the kind where you wear camo and play paintball), Irregular promises to keep these AIs from going rogue. Because nothing says "responsible innovation" like needing a security detail for your algorithms.

But let's be real for a second. The irony here is thicker than a bowl of oatmeal left out overnight. We're throwing millions at a startup to protect us from the very technology we're racing to build. It's like buying a fire extinguisher after you've already set the house on fire for "research purposes." And the best part? This funding round was led by investors who probably can't even spell "AI" without autocorrect, but hey, when there's money to be made, who needs common sense?

In the world of tech, where hype often outpaces reality, Irregular's raise is a masterpiece of timing. They've tapped into the collective fear that our creations might one day outsmart us—a fear that's been fueled by decades of sci-fi movies and one too many late-night coding sessions. With this funding, they plan to hire more experts, develop better tools, and maybe even throw a party where the only rule is "no robots allowed." Priorities, people!

As we march boldly into a future filled with AI that can write poetry, drive cars, and probably file our taxes better than we can, let's all take a moment to appreciate the absurdity of it all. Because if there's one thing we've learned from history, it's that throwing money at a problem usually makes it go away—or at least makes for a great headline. So here's to Irregular: may your security be tight, your investors be patient, and your AI models remain blissfully unaware of how much you're worth.

Sequioa

Irregular Raises $80 Million to Secure AI That May or May Not Start a Robot Uprising

The future is glitched.