Unveiled: Mira Murati's Crazy AI Lab That Could Change Everything—Are You Ready for This?

In the bustling world of artificial intelligence startups, few have created as much buzz as Thinking Machines Lab. Founded by former OpenAI CTO Mira Murati along with a cadre of alumni from the AI powerhouse, Thinking Machines has rapidly positioned itself as a significant player in the tech landscape. Within just five months of its inception in February 2025, the startup raised an impressive $2 billion seed round, achieving a valuation of $12 billion—one of the largest seed rounds in Silicon Valley history.

Unveiling Thinking Machines Lab

Thinking Machines Lab distinguishes itself by focusing on smarter, more efficient AI models rather than competing for dominance with larger models that require extensive data and computing power. The company recently launched its first product, Tinker, designed to help developers fine-tune AI models with greater ease. This innovative approach aims to simplify the complexities and costs associated with distributed computing, making AI more accessible to a broader audience.

Murati's leadership style has garnered attention not only for her technical acumen but also for her management decisions. After a high-profile internal shake-up at OpenAI that saw CEO Sam Altman briefly ousted in November 2023, Murati found herself at the center of controversy. She initially served as interim CEO and expressed concerns over Altman's leadership before ultimately signing a petition to reinstate him. Shortly thereafter, she departed OpenAI to pursue new opportunities, leading to the formation of Thinking Machines Lab.

With a growing team of 30 talented individuals, including top developers and researchers from Meta, Mistral, and other AI labs, the company has quickly established itself as an enigmatic force in the tech world. Murati's significant voting power on the board positions her uniquely to steer the company's direction, a trait that may prove pivotal as the AI landscape continues to evolve.

The competitive dynamics surrounding Thinking Machines intensified when Meta CEO Mark Zuckerberg attempted to acquire the startup. After Murati declined his offer, Zuckerberg sought to entice her employees, particularly targeting Andrew Tulloch, a pivotal figure in the company's formation. Despite receiving a lucrative six-year contract offer worth up to $1.5 billion, Tulloch ultimately decided to join Meta in October 2025. Conversely, Soumith Chintala, co-creator of the popular open-source AI framework PyTorch, left Meta to join Thinking Machines Lab, underscoring the fierce competition for talent in the industry.

Research Innovations at Thinking Machines Lab

Thinking Machines Lab is committed to developing multimodal AI systems designed for collaborative work across various fields. Their mission emphasizes the need to democratize AI knowledge and tools, which is often concentrated within elite research labs. The company has pledged to share its research openly, contributing to a collaborative AI community.

Among their research endeavors is an exploration of Low-Rank Adaptation (LoRA), a fine-tuning technique that has gained traction since its introduction in 2021. This approach enables developers to adapt existing models efficiently without the resource-intensive process of full retraining. Thinking Machines has found that when configured correctly, LoRA can match the performance of traditional methods using smaller datasets, a finding integral to the functionality of Tinker.

Tinker, launched in October 2025, streamlines the customization of AI models for specific tasks, allowing users to modify a range of open-source models, including Meta’s Llama, Alibaba’s Qwen, and OpenAI’s gpt-oss models. By simplifying the process, Tinker enables researchers and developers—whether in medical, legal, or other specialized fields—to create custom-trained models tailored to their specific needs without the burden of extensive computing infrastructure.

While Tinker enhances model fine-tuning, it still necessitates a degree of machine learning expertise. Users can select a base model and utilize Tinker’s flexible API for supervised learning or reinforcement learning. The tool has already been adopted by researchers at prestigious institutions like Princeton University, Stanford University, and the University of California, Berkeley.

As Thinking Machines Lab approaches its one-year mark, the company is eyeing ambitious plans for 2026, including a new funding round seeking an additional $5 billion at a valuation of $50 billion. Chief Scientist John Schulman hinted at plans to release proprietary models and expand Tinker’s capabilities to serve a wider audience with limited technical knowledge. Should they succeed, it could revolutionize how AI is applied across various sectors.

The impact of Thinking Machines Lab will be closely watched as the AI industry continues to evolve. With its emphasis on customization, accessibility, and efficiency, the company has positioned itself at the forefront of a rapidly changing landscape, making its next moves crucial not just for its own trajectory, but for the industry as a whole.