Mira Murati, the former Chief Technology Officer of OpenAI—the powerhouse behind ChatGPT—drops a bombshell. She’s stepping down after six years, including a wild stint as interim CEO during one of the company’s most chaotic chapters. Then, poof, she disappears from the spotlight. Silicon Valley buzzed with speculation: What’s she up to? Is she taking a breather, or cooking up something big? Well, folks, the wait is over. Murati’s back, and she’s swinging for the fences with her own AI startup, Thinking Machines Lab.

The big reveal

Last month, on February 18, 2025, Murati pulled the curtain off her new venture with a casual yet intriguing post on X: “I started Thinking Machines Lab alongside a remarkable team of scientists, engineers, and builders.” No fanfare, no over-the-top hype—just a quiet confidence that’s got the tech world leaning in. What’s the mission? According to the startup’s front page, it’s all about making AI “more widely understood, customizable, and generally capable.” In other words, Murati wants to take AI out of the ivory tower and put it into the hands of everyday people—or at least that’s the vibe she’s giving off.

Unlike some of the splashy AI launches we’ve seen lately, Thinking Machines Lab isn’t spilling all the beans just yet. Product details? Still under wraps. But the company’s dropped enough hints to keep us guessing. They’re aiming to build “frontier AI models” that could shake things up in fields like programming and scientific discovery. Think less “AI that replaces you” and more “AI that works with you.” Their press release doubles down on this, promising “a future where everyone has access to the knowledge and tools to make AI work for their unique needs and goals.” It’s a lofty pitch—accessible, adaptable AI for the masses. But can they pull it off?

The dream team

If anyone’s got a shot at it, it’s Murati—and she’s not doing it alone. She’s assembled what might just be the Avengers of AI talent. First up: John Schulman, an OpenAI co-founder who bolted for rival Anthropic in August 2024 to focus on “AI alignment” (that’s tech-speak for making sure AI doesn’t go rogue). Now, he’s Thinking Machines Lab’s Chief Scientist, bringing his brainpower to Murati’s vision. Then there’s Barret Zoph, a heavy hitter from OpenAI’s research crew who left the same day as Murati—coincidence? Doubt it. He’s stepping in as CTO. Add Jonathan Lachman, ex-head of OpenAI’s special projects, to the mix, and you’ve got a leadership trio that’s already raising eyebrows.

But Murati didn’t stop there. She’s poached around 30 top-tier researchers and engineers from some of the biggest names in AI—OpenAI, Meta, Mistral, Character.AI, Google DeepMind—you name it. Reuters reports that about two-thirds of her team are OpenAI alumni, which is either a flex or a middle finger to her old employer, depending on how you look at it. Sources say more OpenAI folks might still jump ship to join her, and honestly, who could blame them? Murati’s got a track record—she was a key player in turning ChatGPT into a household name—and her pitch must be one hell of a draw.

What’s cooking at Thinking Machines Lab?

So, what’s this crew actually building? The startup’s blog lays out three big goals: helping people tweak AI to fit their specific needs, laying “strong foundations” for more capable systems, and “fostering a broader understanding” through open science. That last bit’s juicy—they’re promising to share code, datasets, and technical papers with the world. In an industry where secrecy’s often the name of the game (looking at you, OpenAI), that’s a bold move. It’s almost like Murati’s saying, “We’re not just here to compete—we’re here to change the conversation.”

The focus on “human-AI collaboration” is another clue. While rivals like Elon Musk‘s xAI are chasing cosmic-scale AI breakthroughs, Thinking Machines Lab seems to be zeroing in on practical, personalized tools. Imagine an AI that doesn’t just spit out generic answers but adapts to how you think, work, or create. Maybe it’s a scientist’s sidekick for cracking tough problems, or a coder’s wingman for debugging on the fly. The startup’s not saying much more—yet—but they’ve teased that “the most advanced models will unlock the most transformative applications and benefits.” Translation: they’re aiming high.

Oh, and one more thing: infrastructure’s a “top priority.” No cutting corners, no duct-tape solutions. They’re building this thing from the ground up to last. That’s a subtle dig at some of the rushed rollouts we’ve seen in AI land lately—cough, certain chatbot glitches, cough.

The money game

Here’s where it gets wild. Just a week after going public, word broke via Business Insider on February 25 that Thinking Machines Lab is already in talks to raise $1 billion at a jaw-dropping $9 billion valuation. For a company that’s barely out of stealth mode? That’s unicorn status on steroids. Sure, the deal’s not locked in—details could shift—but it’s a sign of how much heat Murati’s bringing. Investors are practically tripping over themselves to back AI startups with OpenAI DNA (Anthropic’s $61.5 billion valuation), and Murati’s pedigree makes her a golden ticket.

How does a newborn startup snag a $9 billion price tag? It’s not just hype. Murati’s resume—ChatGPT, OpenAI’s tech leaps, even a brief CEO gig during the 2023 Sam Altman drama—speaks for itself. Pair that with her all-star team, and you’ve got a recipe for venture capital catnip. “How the heck is it valued at $9 billion?” Fair question. The AI bubble’s real, and Thinking Machines Lab hasn’t shown its cards yet. But in this market, reputation and potential can outweigh products—at least for now.

The OpenAI shadow

You can’t talk about Thinking Machines Lab without mentioning OpenAI. Murati spent over six years there, shaping ChatGPT and steering the ship through stormy waters (remember that 2023 boardroom coup?). She was often the public face alongside CEO Sam Altman, especially when he got briefly ousted and she stepped in as interim chief. Her exit last fall was part of a wave of high-profile departures—Schulman, Zoph, others—amid governance shakeups that saw OpenAI pivot from its nonprofit roots to a $260 billion mega-corp, per recent Crunchbase reports.

Is Thinking Machines Lab a direct shot at OpenAI? Maybe not head-on—Murati’s framing it as a different beast, less about raw power and more about usability and openness. But the talent raid and her timing? That’s got to sting in San Francisco. OpenAI’s still the 800-pound gorilla, with Microsoft and SoftBank pouring in billions. Meanwhile, Murati’s playing a longer, leaner game. It’s less David vs. Goliath and more like a savvy ex-employee saying, “I’ll do it my way.”

Why it matters

Murati’s not the first OpenAI alum to spin off her own thing—Anthropic and Safe Superintelligence beat her to it—but Thinking Machines Lab feels personal. She’s betting on a future where AI isn’t just for tech giants or coders in hoodies, but for anyone with a problem to solve. If she nails it, this could shift how we think about AI tools—less black-box magic, more open-ended potential. And with heavyweights like Schulman (a guru in AI safety) on board, there’s a chance they’ll tackle the “how do we trust this stuff?” question head-on, something the industry’s still wrestling with.

Of course, it’s early days. The AI space is a circus—xAI’s chasing $75 billion, Anthropic’s at $61.5 billion, OpenAI’s a $260 billion behemoth. Thinking Machines Lab’s got the buzz, the brains, and soon, maybe the bucks. But promises are cheap, and execution’s everything. Murati’s got the chops, no question—she turned OpenAI’s tech into gold once. Now, she’s got a blank slate and a stacked deck. Will she outshine her old crew? Or is this just another shiny startup in a crowded ring?

We’ll be watching. Thinking Machines Lab might just be the next big thing—or at least the next big story.