Thinking Machines Lab: Building the Future of Collaborative AI

Thinking Machines Lab is an emerging artificial intelligence research and product company founded by Mira Murati – known for her tenure as OpenAI’s Chief Technology Officer and briefly its interim CEO. Established in early 2025 in San Francisco, the lab has quickly become one of the most watched new entrants in the AI industry. Backed by unprecedented funding and staffed with top talent from leading AI organizations, Thinking Machines Lab is on a mission to democratize AI by making advanced AI systems more understandable, customizable, and collaborative for everyone. The tone set by Murati and her team is overwhelmingly positive and ambitious – they envision AI as a tool to empower humanity and serve as “an extension of individual agency” distributed as widely and equitably as possible. This article provides a comprehensive look at the lab’s origins, key milestones, current endeavors, and future aspirations, painting a picture of a startup aiming to reshape the AI landscape in a bold, inclusive direction.

Origins and Founding Vision

Mira Murati founded Thinking Machines Lab in February 2025, following her departure from OpenAI in late 2024. Murati had been a pivotal figure at OpenAI – helping lead the development of breakthrough products like ChatGPT – and even stepped in as interim CEO during the turbulent days when OpenAI’s board briefly ousted Sam Altman in 2023. By September 2024, however, Murati made an abrupt exit from OpenAI “to create the time and space to do my own exploration,” as she later explained. That exploration took shape as Thinking Machines Lab, born from Murati’s conviction that the next generation of AI should be developed in a more open, human-centric manner.

From the outset, Murati’s vision for Thinking Machines Lab has been clear and compelling. “We’re building a future where everyone has access to the knowledge and tools to make AI work for their unique needs and goals,” the lab’s website proclaims. This vision addresses several gaps Murati perceived in the current AI landscape: The understanding of frontier AI systems still lags behind their rapidly advancing capabilities; expertise in training such systems is concentrated in only a few big labs; and AI models are often difficult to adapt or customize to specific user values and domains. Murati launched Thinking Machines Lab specifically to bridge these gaps. Her founding premise is that artificial intelligence should not be the guarded domain of tech giants, but a broadly accessible technology that works for everyone.

Crucially, Murati chose to incorporate the company as a Public Benefit Corporation (PBC), signaling a commitment to societal good alongside profit. This means the lab’s charter includes public interest objectives – like advancing safe and beneficial AI – in addition to traditional business goals. In practice, this ethos shines through in the lab’s emphasis on transparency, safety, and collaboration. Murati even structured the corporate governance to safeguard the mission: she retains a deciding vote on board matters and the founding team’s shares carry super-voting power, ensuring that short-term commercial pressures won’t derail the lab’s long-term vision. From day one, Thinking Machines Lab was set up to pursue “collaborative general intelligence” as a grand goal – pushing toward AI that can reason generally like humans, but in a collaborative, human-aligned way.

Assembling a World-Class Team

A key factor setting Thinking Machines Lab apart is the all-star team that Murati assembled even before the lab’s public debut. By the official launch in February 2025, the startup had quietly hired around 30 seasoned AI researchers and engineers poached from top institutions such as OpenAI, Meta AI, Google, and the startup Mistral AI. Remarkably, nearly two-thirds of the initial team were former OpenAI employees. This brain trust of talent included several notable figures in the AI field:

• John Schulman – A co-founder of OpenAI and co-author of the seminal Reinforcement Learning from Human Feedback (RLHF) approach, Schulman left OpenAI (after a stint at competitor Anthropic) to join Murati as Thinking Machines Lab’s Chief Scientist. His expertise in AI alignment and deep reinforcement learning is foundational for the lab’s research direction.
• Barret Zoph – A prominent AI researcher known for his work at Google Brain (AutoML) and later OpenAI, Zoph departed OpenAI on the same day as Murati in late September 2024. He now serves as Thinking Machines Lab’s Chief Technology Officer, bringing prowess in cutting-edge model development and architecture search.
• Luke Metz – An expert in machine learning and former researcher at OpenAI, Metz joined the lab early on, contributing to its research strength.
• Lilian Weng – Former head of safety research at OpenAI, Weng is another high-profile addition to the team. Her background in AI safety and alignment research underscores the lab’s commitment to developing AI that is safe and aligned with human values.
• Jonathan Lachman and Alec Radford (Advisor) – Among other OpenAI alumni on board, Lachman (a technical lead from OpenAI) joined the lab, and Alec Radford – co-creator of the generative GPT models at OpenAI – signed on as an advisor. Their involvement adds deep generative modeling and strategy experience.
• Bob McGrew (Advisor) – Formerly OpenAI’s Chief Research Officer, McGrew joined as an advisor as well, lending the lab guidance from one of the minds who helped steer OpenAI’s research.

Together, Murati and this core team bring an unparalleled pedigree; they have helped build some of the world’s most impactful AI systems (from ChatGPT to Tesla’s Autopilot to OpenAI Gym). The excitement around Thinking Machines Lab is largely because such a concentration of AI talent is now working under one roof on a shared mission. “We are scientists, engineers, and builders who’ve created some of the most widely used AI products… and popular open-source projects,” the company notes, emphasizing the team’s collective experience. By uniting experts in large-scale model training, AI safety, and product engineering, the lab is well-positioned to tackle ambitious problems. This world-class team not only validates Murati’s drawing power as a leader but also signals that the startup can hit the ground running in the race to advance AI.

Major Milestones and Achievements

Though Thinking Machines Lab is still in its infancy, it has already notched several key milestones in its journey. Below is a timeline of the most significant events and achievements shaping the organization so far:

1. September 2024 – A New Venture Takes Shape: Mira Murati and a handful of colleagues, including Barret Zoph, depart OpenAI to form a stealth AI initiative. Over the next few months, Murati quietly recruits top talent from AI labs like Meta and Google. By year’s end, the venture has a core team in place and early backing from prominent investors (with rumors swirling that venture capitalists were eager to fund Murati’s project).
2. February 18, 2025 – Official Launch and Team Reveal: Thinking Machines Lab emerges from stealth, announcing its name, mission, and initial team to the world. In a blog post and press interviews, the lab reveals it has hired about 30 leading researchers and engineers, roughly “two-thirds of which [are] former OpenAI employees”. Murati is CEO, John Schulman is named Chief Scientist, and Barret Zoph is CTO. The launch communications highlight the lab’s goal of bridging the gap between AI research and real-world applications – making AI more understandable, customizable, and collaborative for diverse use cases. Media outlets immediately label Thinking Machines Lab a potential “OpenAI competitor,” given its star leadership and ambitious vision.
3. March–April 2025 – Rapid Growth and High-Profile Advisors: In the weeks following launch, the lab continues to attract talent. Notably, two respected ex-OpenAI researchers – Bob McGrew and Alec Radford – join as advisors in early April, bringing deep expertise in AI strategy and research oversight. Meanwhile, reports surface that Murati’s startup is seeking a record-breaking funding round: by April 11, Reuters insiders report that Andreessen Horowitz (a16z) is in talks to lead an approximately $2 billion seed investment in Thinking Machines Lab. If completed, it would be one of the largest seed rounds ever in Silicon Valley, reflecting extraordinary investor confidence.
4. July 15, 2025 – Record Seed Funding of $2 Billion: Thinking Machines Lab closes on a massive early-stage funding round of $2 billion, led by venture firm Andreessen Horowitz, with participation from tech heavyweights including Nvidia, AMD, Cisco, ServiceNow, Accel, and Jane Street. The investment, confirmed by the company in mid-July, values the startup at around $12 billion – an astonishing valuation for a company barely half a year old with no product on the market. This milestone funding underscores the faith that investors have in Murati’s vision and the team’s capabilities. Even the government of Albania (Murati’s country of origin) joined the round with a $10 million investment, a point of pride that required a special amendment to Albania’s national budget. The seed round is among the largest in history, and it gives Thinking Machines Lab a formidable “war chest” to develop advanced AI models and infrastructure going forward.
5. Mid-2025 – Strategic Partnerships and Industry Buzz: Alongside the funding news, Murati discloses that Thinking Machines Lab has struck a deal with Google Cloud to provide the computing infrastructure for its AI development. This partnership ensures the lab has the necessary cloud resources and hardware (augmenting support from GPU maker Nvidia and chipmaker AMD) to train frontier models. Around the same time, industry chatter reveals that Meta (Facebook’s parent company) had held talks about acquiring Thinking Machines Lab outright, aiming to bolster its own AI efforts. No formal offer materialized – a sign that Murati intends to keep the company independent – but the mere fact that Meta showed interest speaks to how significant the lab’s potential appears within the tech community. Observers note that Thinking Machines Lab has quickly joined the ranks of top-tier AI startups considered credible challengers to incumbents like OpenAI, Anthropic, and Google DeepMind.
6. August 2025 – Poised for First Product Launch: By late summer 2025, Thinking Machines Lab remains in development mode with no public product yet – but that is about to change. Murati has hinted on social media and in interviews that the first product will be unveiled in a matter of months, saying “in the next couple [of] months we will be able to share our first product”, which will include a “significant open source component”. As anticipation builds, the team has grown to around 50 employees and continues to hire aggressively, particularly seeking engineers experienced in building AI-driven products and robust AI infrastructure. This positions the lab to enter its next phase – delivering tangible AI tools – before the end of 2025.

Through each of these milestones, Thinking Machines Lab has maintained steady momentum and kept true to its core philosophy. The speed at which it secured funding and talent is matched only by the deliberation with which it advances its research goals (so far opting to perfect its work behind closed doors rather than rush out a demo). The milestones also illustrate how the lab has been shaped by both internal decisions (hiring, corporate structure, research focus) and external forces (investor enthusiasm, industry dynamics). In just a short time, Murati’s venture has evolved from an idea into a well-funded organization that peers already consider a major new player in AI.

Mission and Core Initiatives

At the heart of Thinking Machines Lab lies a commitment to fundamentally improve how AI is developed and applied. The lab’s mission can be distilled into a few core initiatives and principles that guide its research and product strategy:

• Open Science and Transparency: “Scientific progress is a collective effort,” Murati’s team declares, emphasizing that sharing knowledge benefits everyone. In contrast to many AI companies that operate behind closed doors, Thinking Machines Lab has pledged to frequently publish technical blog posts, research papers, and open-source code. This open-science ethos is intended to accelerate understanding of advanced AI systems across the entire field. By making its models, training methodologies, and findings transparent, the lab hopes to empower other researchers and invite feedback. As noted in its launch materials, the company believes collaboration with the wider community will “improve our own research culture” while advancing the state of the art for all.
• Human-AI Collaboration (AI that Works With People): A distinguishing focus of Thinking Machines Lab is on AI systems that augment and partner with humans, rather than just autonomous agents. “Instead of focusing solely on making fully autonomous AI, we are excited to build multimodal systems that work with people collaboratively,” Murati says. This means developing AI that can engage through natural human modes of interaction – conversing in language, perceiving the visual world, and responding to context – in order to truly assist in human endeavors. The lab has coined the term “collaborative general intelligence” to describe AI that is both broadly capable and inherently designed to cooperate with human input. This could manifest as AI assistants that join experts in problem-solving, creative tools that work hand-in-hand with users, or AI systems embedded in robotics that learn from and adapt to human coworkers. Thinking Machines Lab sees multimodality (integrating text, images, perhaps audio and other inputs) as crucial for more natural and effective human-AI partnerships.
• Customization and Modular AI Architectures: Another pillar of the lab’s approach is making AI flexible and adaptable to different needs. Today’s leading AI models (like GPT-4) are powerful but largely one-size-fits-all; by contrast, Murati’s team aims to create AI that can be tailored to specific industries, tasks, and user preferences. In practice, this involves research into modular architectures – AI systems composed of interchangeable components or skills that can be configured for particular domains. “We see enormous potential for AI to help in every field of work,” the company notes, “while current systems excel at programming and mathematics, we’re building AI that can adapt to the full spectrum of human expertise and enable a broader spectrum of applications.” From precision manufacturing to life sciences, Thinking Machines Lab intends its AI to be customizable, so that a scientist, a factory engineer, or a healthcare professional could each have AI systems tuned to their unique requirements. This modular, LEGO-block style design also aligns with the lab’s collaborative philosophy – allowing external contributors to add or improve components that benefit the whole.
• Safety and Alignment as a Foundation: Given the backgrounds of Murati (who oversaw safety at OpenAI) and others on the team, it’s no surprise that AI safety is a core emphasis. Thinking Machines Lab is deeply invested in AI alignment – ensuring AI systems behave in accordance with human values and do not pose unintended harms. The lab’s strategy involves both proactive research and practical measures: training AI with human feedback, rigorous “red-teaming” tests to probe for weaknesses, and post-deployment monitoring to catch issues in real-world use. Uniquely, the lab has stated it will share its alignment methods openly (such as releasing datasets, model specifications, and best-practice “recipes” for safe AI) to help the broader industry tackle AI safety challenges. By accelerating external research on alignment, Murati hopes to contribute to making future AI systems (even outside her lab) more reliable and trustworthy. Indeed, the presence of alignment researchers like John Schulman and Lillian Weng on the team reinforces that safe AI development isn’t an afterthought—it’s baked into the lab’s R&D from the ground up.
• Research–Product Co-Design: In line with Murati’s experience delivering real-world AI products, Thinking Machines Lab champions an iterative cycle between fundamental research and product development. The team believes that deploying products provides valuable feedback to guide research toward truly impactful problems. Conversely, cutting-edge research can unlock new product capabilities. This “learning by doing” philosophy means the lab, despite its research orientation, is building with an eventual end-user in mind. Murati has explicitly sought team members who are excited to wear multiple hats – from crafting user interfaces to scaling backend systems – to translate AI research breakthroughs into “useful things” people can actually use. By co-designing research and product in tandem, the lab aims to avoid the common pitfall of AI projects that look impressive in the lab but fail to solve real-world needs. In essence, practicality and scientific experimentation go hand-in-hand at Thinking Machines Lab.

These core initiatives form a cohesive strategy: the lab is pushing the envelope on AI capabilities (advanced general AI research) and doing so in a way that remains open, safety-conscious, and attuned to human collaboration. The tone of all their communications has been optimistic and inclusive – rather than framing AI as a mysterious superintelligence to be feared or locked down, Thinking Machines Lab talks about demystifying AI and making it a readily available tool. “AI that works for everyone” isn’t just a slogan; it underpins choices from sharing code to designing AI that individuals can personalize. By setting these principles early, the lab is deliberately carving out a niche as the responsible and community-driven contender in an industry often criticized for secrecy and competitive friction.

Funding, Structure, and Industry Impact

Thinking Machines Lab’s positive vision has resonated not only with researchers but also with investors and industry leaders, resulting in extraordinary support for the young company. The lab’s financial backing and organizational structure are worth examining, as they provide the foundation upon which future ambitions will be built.

The $2 billion seed funding raised in mid-2025 stands out as a landmark event. This was no ordinary startup fundraising – it was one of the largest seed rounds in Silicon Valley history. The round was led by Andreessen Horowitz (a16z), a top venture capital firm known for big bets on AI, and included an all-star roster of co-investors: Nvidia (the GPU manufacturer supplying the computing power behind modern AI), AMD (another chip giant), enterprise tech leaders like Cisco and ServiceNow, global investment firm Jane Street, and more. Such a broad coalition of backers signals immense confidence in Thinking Machines Lab’s prospects. As one journalist noted, this “massive funding round for a company launched only in February, with no revenue or products yet, underscores Murati’s ability to attract investors” in a hotly competitive sector. Indeed, investors appear to be betting on the personnel and vision above all – essentially, faith that Murati’s leadership and her team’s expertise will yield breakthrough innovations (even if details are under wraps for now).

With a post-money valuation of around $12 billion, the lab is valued more highly than many companies with years of product development. This puts Thinking Machines Lab in rare company alongside other high-profile AI startups founded by ex-OpenAI leaders (for example, Dario Amodei’s Anthropic and Ilya Sutskever’s Safe Superintelligence) which have also quickly reached valuations in the tens of billions. The message is clear: there is intense market appetite to fund new AI labs that have the right mix of talent and vision, even ahead of concrete deliverables. For Thinking Machines Lab, the immediate practical impact of the $2B infusion is access to resources—state-of-the-art computing infrastructure, top-tier hires, and the ability to experiment freely—on a scale that few startups ever attain. It gives the lab a fighting chance to train “frontier” AI models (those at the cutting edge of size and capability) and compete with tech giants in the AI research arms race.

Murati has leveraged this position of strength to shape the company’s structure and external partnerships strategically. We’ve noted the choice to register as a Public Benefit Corporation, aligning the legal framework with the lab’s mission-driven ethos. Additionally, the governance is set up to keep decision-making power in mission-aligned hands: Murati herself holds a special majority on the board, and early shareholders (presumably including the core team) have super-voting shares. These measures, unusual for a startup taking on so much venture capital, suggest that Murati negotiated to ensure long-term autonomy and fidelity to the lab’s principles, even as investors come on board. In effect, supporters like a16z and others are entrusting her with significant capital and control, a testament to their belief in her leadership. For Murati, this autonomy is likely critical to pursuing high-risk, high-reward research (which might not pay off immediately) and enforcing strict safety standards without external pressure to rush products to market.

The lab’s swelling war chest has also enabled key partnerships. One significant alliance is with Google Cloud, which is providing cloud computing resources to power Thinking Machines Lab’s experiments and model training. This partnership, alongside backing from hardware makers, ensures the lab can access vast computing scale (potentially tens of thousands of GPU/TPU chips) necessary to train large-scale AI models and handle complex multimodal data. Such infrastructure is often a bottleneck for new AI ventures, but Thinking Machines Lab appears to have secured what it needs to operate at the frontier. Another ripple effect of the lab’s rapid rise is its impact on the talent market in AI. Rival companies have taken notice: in fact, as mentioned, Meta reportedly explored acquiring the lab outright in 2025 to bolster its own superintelligence initiatives. While no deal transpired, the very notion of an acquisition so early hints at how valuable Thinking Machines Lab’s talent and ideas are considered. The startup has essentially become a kingmaker in the AI space overnight – its team and tech are highly coveted, and its moves are closely watched.

All of this has positioned Thinking Machines Lab as a potential major player in the AI industry much faster than normal. Media coverage frequently cites the lab as one of a handful of startups that could meaningfully challenge the established AI giants in the coming years. Its combination of factors – exceptional team, huge funding, open and ethical stance – makes it stand out. While it’s still early days, the lab’s influence is already being felt: it has intensified the competition for AI research talent (hiring away experts from larger labs), contributed to the narrative of an ongoing “AI startup boom,” and raised expectations for what new ventures in this space can achieve with the right backing. If Murati’s venture succeeds, it could validate an alternative model to Big Tech’s dominance in AI, one where a lean independent lab can innovate faster and more responsibly.

Current State and Future Outlook

As of mid-2025, Thinking Machines Lab is in a strong position, though much of its work remains behind closed doors. The current state of the lab is characterized by intense research and development activity as the team prepares for its first public releases. Internally, the company has grown to around 50 employees, including not just researchers but also engineers and product designers – a relatively small but elite group. These individuals are busy building the lab’s core AI systems, training large models, and iterating on prototypes. Thanks to the ample funding and partnerships, they have access to cutting-edge hardware and tools, meaning experimentation can happen at a pace similar to that in far larger organizations.

Morale and sense of purpose at the lab are reportedly high. Mira Murati’s leadership style – described as collaborative and forward-looking – has fostered an environment where top minds are “passionate about translating research into useful things”. The team’s daily work likely involves pushing the boundaries of multimodal AI (AI that can understand and generate various types of data, such as text and images together), as well as refining the modular AI platform that will allow customization for different tasks. Safety checkpoints are built into this process: the presence of alignment experts suggests ongoing audits of the models’ behavior and incorporation of alignment techniques. While exact details are sparse (the lab has been deliberately secretive about technical specifics so far), one can imagine the current projects range from training a general language-vision model to developing an interface or API through which outside developers might leverage the lab’s AI.

The big question on everyone’s mind is: What will Thinking Machines Lab release first? Murati has tantalized the public with hints that an initial product is imminent. Based on her comments, this first product will likely be some form of AI platform or model that others can build on – essentially a demonstration of the lab’s collaborative AI philosophy. She noted that it will “be useful for researchers and startups developing custom models”, implying a tool or service rather than a consumer-facing app. Importantly, it will have a significant open-source component. This could mean the model weights might be released openly or a portion of the system (for example, a model foundation or a framework) will be made free for the community. An open-source release would align with Thinking Machines Lab’s mission to broaden access to AI, and it would immediately differentiate their approach from more closed competitors. It might mirror how companies like Meta have open-sourced models (e.g., LLaMA) to catalyze innovation, but with the twist that Murati’s team is doing so from the outset as part of their strategy.

Beyond the first product, the future goals and aspirations of Thinking Machines Lab are lofty and far-reaching. Murati has articulated a vision of developing collaborative general intelligence – AI that not only performs a wide array of intellectual tasks at a human level (general intelligence), but does so alongside humans, enhancing human capabilities. This contrasts with the usual notion of AI superseding human work; instead, the lab imagines AI amplifying human creativity, knowledge, and productivity in a harmonious loop. Achieving this will likely require incremental advances: first mastering multimodal understanding, then real-time human-AI interaction, then embedding alignment deeply so that the AI’s actions remain trustworthy in complex scenarios. The team’s focus on “the full spectrum of human expertise” suggests that in the long run, they aim for AI that can assist in nearly any domain – from scientific research and engineering design to artistic endeavors and yes, even robotics and physical-world tasks. In fact, Thinking Machines Lab’s mention of fields like precision manufacturing hints that their AI could be integrated into robotics and automation, enabling smarter factory machines or collaborative robots that work safely with people on the assembly line. Similarly, in life sciences, one could foresee their AI helping researchers sift through data or design experiments. The applications are vast, and the lab appears intent on laying a foundation that is adaptable to all these scenarios.

In the coming years, one benchmark for the lab’s success will be how it contributes to the AI community at large. Murati has pledged to share not just products but also “our best science to help the research community better understand frontier AI systems.”. We can expect white papers detailing their model architectures or findings on AI behavior, open datasets that others can use for alignment research, and perhaps new benchmarks for evaluating collaborative AI. If Thinking Machines Lab fulfills this promise, it could become a hub for open innovation much like OpenAI was in its early days – but with an even more openly collaborative bent. The lab’s work on alignment might also set standards in the industry, especially if they demonstrate practical techniques to keep powerful AI systems safe and under human control. Given how critical AI safety is for the future, the broader impact of such contributions could be highly significant.

Another aspiration implicitly held by Murati and her colleagues is to shape the narrative around AI’s development in a positive way. Thinking Machines Lab consistently frames AI as a beneficial force – one that, if guided correctly, can “empower humanity” rather than threaten it. By building socially beneficial applications and making their tools accessible, the lab could help shift public perception of AI toward optimism and away from fear. This advocacy for responsible AI optimism might involve engaging with policymakers, participating in industry-wide safety collaborations, or setting an example of ethical startup culture. Murati, as one of the few female CEOs in AI and someone with a platform, has already become an influential voice. Her startup’s journey will be closely watched not just for technical achievements but also for how it navigates doing AI differently – more openly, ethically, and inclusively.

In summary, the future for Thinking Machines Lab is bright and filled with hard work. In the near term, all eyes are on the lab’s upcoming product reveal, which will provide the first concrete look at what this talented team has been creating. If they deliver a versatile AI platform with open-source elements as promised, it could rapidly become a foundational tool for researchers and developers worldwide, instantly raising the lab’s profile. Longer term, as the lab iterates and improves its systems, we may witness the emergence of AI that feels like a true collaborator – an AI that can see, converse, and reason within the messy context of human endeavors, helping us solve problems together. That is the exciting endgame that Murati and her colleagues are working toward. They are keenly aware of the challenges ahead (from stiff competition to the scientific hurdles of alignment and generalization), yet they face them with a sense of optimism and determination that permeates everything the lab communicates.

Thinking Machines Lab, in just its first chapters, has already begun to leave an imprint on the AI world. Its story is one of bold departures – a break from the closed, big-player mold – and bold aspirations to make AI better for everyone. As the Universal Robot Consortium Advocates and technology enthusiasts observe this venture’s progress, there is cause for genuine excitement. With its positive, human-centric approach, Thinking Machines Lab embodies the hope that the next era of AI will not only be defined by greater intelligence, but by greater wisdom in how that intelligence is shared and employed. In the quest to build AI that truly serves humanity, Murati’s lab is charting a new path, and it’s one that many are eager to follow in the months and years ahead.

---

References

1. Buntz, Brian. “Ex-OpenAI CTO launches startup to bridge the gap between AI research and real-world applications.” R&D World, 19 Feb. 2025.
2. Reuters. “Mira Murati debuts her AI startup, Thinking Machines Lab.” Fast Company, 18 Feb. 2025.
3. Hu, Krystal, and Aditya Soni. “Mira Murati’s AI startup Thinking Machines valued at $12 billion in early-stage funding.” Reuters (via Yahoo! Finance), 15 July 2025.
4. Zeff, Maxwell. “Mira Murati’s Thinking Machines Lab is worth $12B in seed round.” TechCrunch, 15 July 2025.
5. Tremayne-Pengelly, Alexandra. “Former OpenAI CTO Mira Murati’s A.I. Startup Hits $12B Valuation in Just 5 Months.” Observer, 16 July 2025.
6. Capoot, Ashley. “Former OpenAI CTO Mira Murati raises $2 billion for new AI startup Thinking Machines Lab.” CNBC, 15 July 2025.