Although frontier Artificial Intelligence systems are advancing at unprecedented speed, the capacity to interpret, adapt and reliably integrate them into practical settings remains limited. Thinking Machines Lab, a San Francisco-based company is emerging as a notable player in the global AI market, developing models that are transparent, collaborative, and applicable across diverse domains.
Founded in 2025 by former OpenAI leaders – Mira Murati, John Schulman, Lilian Weng, Andrew Tulloch, and Barrett Zoph, the company is building a new generation of AI systems designed to be widely understandable, customisable, and integrated into real-world applications.
Thinking Machines Lab was officially unveiled in February 2025. With Mira Murati at the helm as CEO and co-founder, the company’s team is building a venture grounded in scientific openness and practical deployment.
It is one of the prominent players in the global AI market and among the most generously funded AI companies in the industry’s history. The company has expressed its intentions to collaborate with the wider community of researchers and builders for its system, and a focus on their product’s reproducibility.
Thinking Machines Lab’s team includes experts who have previously contributed to widely used AI tools such as ChatGPT, PyTorch, and Segment Anything. Their collective experience in developing these high-impact systems brings deep technical expertise that has been central to the company’s growth and its ability to deliver advanced AI solutions.
Thinking Machines Lab’s vision
Mira Murati captured the ethos of their company in a social media post saying: “Thinking Machines Lab exists to empower humanity through advancing collaborative general intelligence.” She also wrote: “We believe AI should serve as an extension of individual agency and, in the spirit of freedom, be distributed as widely and equitably as possible,” reflecting the company’s mission catering to a growing movement in the AI world that wants social responsibility and transparency alongside capability.
Funding and economic journey
In July 2025, the company was valued at approximately $12 billion following a $2 billion seed funding round led by Andreessen Horowitz. Thinking Machines Lab confirmed the investment, which also included participation from Nvidia, Accel, AMD, Cisco, Jane Street and ServiceNow. Investors described the move as a bet on a future where open access AI infrastructure sits alongside advanced safety research. By attracting investors from both the chip and enterprise software industries, the company is positioned to connect advanced AI computing capabilities with practical, real-world applications.
This early capital injection has given Thinking Machines Lab the freedom to operate independently, build its own training clusters and attract senior researchers who may otherwise have joined larger incumbents.
Tinker Launch
The company’s first public product arrived in October 2025, called Tinker. It provides developers and researchers with a managed API to fine-tune open-weight language models on Thinking Machines Lab’s own infrastructure. Users can run experiments on models of varying sizes without needing to invest in expensive hardware or manage complex computing resources. By simplifying access to high-performance training, Tinker makes it easier for teams to customise AI models for practical applications and research projects. According to sources, the tool supports multiple models and is designed for researchers, startups and advanced technical teams who want to create tailored AI without heavy engineering overhead.
Customisable AI Systems
This launch reflects the company’s focus on practical, accessible AI. The product is designed to be adaptable, allowing users to customise it for their own needs, aligning with Murati’s vision of AI as collaborative, transparent, and widely available.
As Murati wrote in a post on the social network X: “Tinker brings frontier tools to researchers, offering clean abstractions for writing experiments and training pipelines while handling distributed training complexity. It enables novel research, custom models, and solid baselines.”
While Thinking Machines Lab does intend to build frontier systems for science and programming applications, the narrative is about enabling people to customise and understand those systems. Reports indicated that by launch, the company had hired around 30 researchers and engineers from OpenAI, Meta AI, and Mistral AI, bringing together a team with deep expertise in AI research and deployment. The founders argue that the world benefits when more minds experiment, audit, and refine AI. In practice, that has translated to plans for frequent technical blog posts, open research papers, and code releases. The company has said its next releases will include open-source components to support model building and evaluation, alongside tools that help users test behaviour and monitor safety.
Future Focus: Multimodality and Safety
Future products are expected to lean heavily into multimodality. The team has spoken about systems that respond to natural conversation, visual input and practical tasks in messy environments, such as creative workflows and scientific exploration. There is also a practical stance on safety. The group promotes an empirical approach that uses real-world testing, red-teaming and post-deployment monitoring. By committing to publish best practices and model specifications, the company hopes to encourage a culture where safe development is informed by shared evidence. This pragmatic approach has been welcomed by developers who often struggle with abstract or opaque safety standards.
Market reception and competitive dynamics
Thinking Machines Lab has captured strong attention across the AI industry. Startups and researchers expressed enthusiasm for a lab led by individuals who helped build some of the most recognisable AI systems yet want a more open culture. Investors described it as a long term bet on distributed compute and collaborative intelligence.
Conclusion
Still early in its development, Thinking Machines Lab has blended scientific ambition with a clear commitment to openness. If the company delivers on its promise of accessible tools and rigorous research, it could become one of the most influential AI organisations of the next decade.
