Thinking Machines Lab, the AI research startup founded by former OpenAI co-founder Mira Murati, has just secured a landmark multi-year strategic partnership with semiconductor powerhouse Nvidia. The deal, announced on Tuesday, positions the two-year-old company as one of the most ambitious players in the race to build reliable, reproducible artificial intelligence — and it comes with a jaw-dropping commitment: at least one gigawatt of Nvidia's next-generation Vera Rubin computing systems, set to come online starting in 2027.
 |
| Credit: Patrick T. Fallon / AFP / Getty Images |
What Is the Thinking Machines Lab and Nvidia Deal, Exactly?
The partnership goes far beyond a simple hardware purchase agreement. Nvidia is not just supplying chips — it is also making a direct strategic investment in Thinking Machines Lab as part of the arrangement. The financial terms of the overall deal remain undisclosed, but the scale of what is being committed tells its own story. Deploying one gigawatt of compute infrastructure is not a modest ambition; it signals a company preparing to train and run AI models at a scale comparable to the world's top AI labs.
Thinking Machines Lab was founded in February 2025 and has moved with remarkable speed since its launch. The company has already raised more than $2 billion from an impressive roster of backers, including Andreessen Horowitz, Accel, Nvidia, and even the venture arm of AMD — a rival chipmaker. That unusual mix of investors reflects both the confidence the industry has in Murati's vision and the enormous strategic value of being close to this company as it grows.
Who Is Mira Murati, and Why Does This Matter?
Mira Murati spent years at OpenAI, eventually serving as Chief Technology Officer before departing to launch Thinking Machines Lab. She was one of the architects behind some of the most transformative AI products the world has seen — meaning her new venture carries serious intellectual credibility from day one. That pedigree is a large part of why investors were willing to pour billions into a company that, at the time of its first major funding rounds, had not yet shipped a single product.
In October 2025, the company released its debut product: an API called Tinker. The release marked the first tangible output from a lab that had, until that point, operated largely behind closed doors. Tinker represents the company's early commitment to building AI systems that deliver not just impressive outputs, but consistent, reproducible results — a problem that has quietly plagued AI development across the industry.
The Vera Rubin Bet: Why Nvidia's Latest Systems Are Central to This Story
The compute infrastructure at the heart of this deal is built around Nvidia's Vera Rubin systems, a platform released earlier in 2026 that represents the company's most advanced architecture to date. Named after the pioneering astronomer, Vera Rubin systems are designed to handle the enormous computational demands of training and running next-generation AI models. Committing to at least one gigawatt of this infrastructure — starting in 2027 — puts Thinking Machines Lab in the same league as the largest hyperscale AI operators in the world.
To put that number in perspective: one gigawatt of compute power is roughly equivalent to the output of a medium-sized power plant. It is an extraordinary resource commitment for a company that is still, technically, at the seed stage. Yet that contrast — seed-stage company, gigawatt-scale ambitions — perfectly captures the pace at which the AI industry is moving right now.
A $12 Billion Valuation With a Clear Mission
Thinking Machines Lab is currently valued at more than $12 billion, making it one of the most valuable seed-stage companies in the history of technology. That valuation reflects not just the talent behind the lab, but the clarity of its mission. While many AI companies are racing to produce the most capable model possible — often at the expense of predictability — Thinking Machines Lab is specifically focused on building AI that creates reproducible results.
Reproducibility may sound like a technical footnote, but it is actually one of the most important unsolved problems in modern AI. When an AI system gives different answers to the same question at different times, it undermines trust and limits how reliably it can be deployed in real-world applications. A lab that cracks this problem will have a significant edge in enterprise, medical, legal, and scientific use cases — all areas where consistency is not optional.
What This Partnership Means for the Broader AI Compute Race
The deal between Thinking Machines Lab and Nvidia is the latest signal that the competition for AI supremacy is increasingly being decided at the infrastructure layer. Building a brilliant model is no longer enough on its own — you need the hardware to train it, the power to run it, and the strategic relationships to secure access to the best chips before your competitors do. Nvidia has become the central node in that ecosystem, and its decision to invest directly in Thinking Machines Lab is a bet that Murati's approach will be among the most impactful of the decade.
It also reflects a broader shift in how AI companies are being built. The old model — raise a modest seed round, build a prototype, prove product-market fit, then scale — has been replaced by something far more aggressive. Today's leading AI labs are raising billions before they have a single customer, securing compute years in advance, and operating at a scale that would have seemed impossible just five years ago.
What Comes Next for Thinking Machines Lab?
With the Nvidia partnership announced and more than $2 billion in funding secured, all eyes will now turn to what Thinking Machines Lab actually builds with these resources. The Tinker API is a starting point, but the company's long-term ambitions are clearly much larger. The infrastructure being put in place — Vera Rubin systems, gigawatt-scale compute, a war chest of investor capital — is designed to support the training of foundation models that can compete with, or surpass, anything currently on the market.
For Mira Murati, the Nvidia deal is more than a press release. It is a declaration of intent. Two years after leaving one of the world's most famous AI companies, she is building something that, by every measure, has the resources and the talent to challenge it. The AI race in 2026 has never been more intense — and Thinking Machines Lab has just made clear that it intends to be at the front of the pack.
