Is Mira Murati’s next act the rise of a new AI empire?
Subscribe to our free newsletter today to keep up to date with the latest business news.
When Mira Murati left her position as OpenAI’s Chief Technology Officer in September 2024, the tech world speculated about her next move. Now, less than a year later, she’s unveiled Thinking Machines Lab — a new artificial intelligence startup with bold ambitions. With an elite team, deep funding, and a philosophy that diverges sharply from her former employer’s, Murati seems poised to challenge the very systems she helped shape.
A Strategic Departure at the Right Time
Murati’s departure from OpenAI came during a period of reflection within the industry. With public trust in large AI labs wavering, debates around openness and governance were intensifying. In interviews since, Murati has hinted at her discomfort with the increasing secrecy in AI development, and her desire to build tools that are more transparent and user-friendly.
Thinking Machines Lab positions itself as a response to that moment. A lab where the next wave of innovation is rooted in customizability, accessibility, and open science. The startup aims to bring control back to users by offering tools that can interpret and generate not only text, but also images, video, and audio all through one multimodal interface.
Building a World-Class Team
Thinking Machines Lab is not a solo venture. Murati has attracted some of the industry’s most respected figures. John Schulman, a key researcher in reinforcement learning from OpenAI, has joined as Chief Scientist. Barrett Zoph, who previously led large-scale model research, is on board as Chief Technology Officer. According to company insiders, nearly two-thirds of the team have previously worked at OpenAI or other leading AI labs.
This depth of expertise adds credibility and momentum, and the startup has already closed a $2 billion seed round. Its valuation now sits between $10 and $12 billion, which is an extraordinary figure for a company not yet out of stealth mode.
A Vision for Open, Multimodal AI
What sets Thinking Machines apart is not just who is involved, but what they plan to build. Murati describes the lab’s approach as “human-centered AI.” Rather than training ever-larger models in secret, the company intends to make its research, code, and insights publicly available. The first model, currently in private testing, is designed to understand and generate language, images, sound, and video all within one system.
Where other labs are scaling up proprietary systems behind closed doors, Thinking Machines aims to open up the design process. It wants users to customize how their AI behaves and interacts, offering new levels of transparency and control. This is not just a technical strategy, but a philosophical one, driven by concerns about centralized AI power and the need for public accountability.
Competing with the Giants
While OpenAI, DeepMind, and Anthropic continue to dominate the frontier, Murati is betting on agility and openness. By sidestepping the bureaucratic and ethical challenges that slow down incumbents, she believes Thinking Machines Lab can outpace them in research and real-world application.
Still, challenges loom. Maintaining this level of funding will require delivering commercially viable products. Developing new model architectures that are truly multimodal is technically demanding. And aligning a business built on transparency with investor expectations won’t be easy.
Yet Murati’s track record gives her a solid foundation. At OpenAI, she played a key role in shaping ChatGPT and building the partnerships that helped define the company’s trajectory. Now, she’s free to pursue a vision without compromise.
Whether Thinking Machines Lab becomes the next OpenAI or forges its own path entirely, it is already setting new expectations for what an AI lab can be. In a field often driven by hype and secrecy, Murati’s next act is beginning with something rare — clarity.
Source: