Nvidia is reportedly in talks to invest about $2.5 billion in Thinking Machines Lab, the AI startup founded by former OpenAI CTO Mira Murati, as part of a larger financing round that could total $5 billion to $6 billion and carry a pre-money valuation of at least $40 billion. The deal remains under negotiation and final terms have not been determined.
Nvidia’s expanding model‑layer investments
According to market sources, the proposed Nvidia stake would follow a string of recent capital deployments by the chip maker into the model layer. A version of the financing materials shared by Andreessen Horowitz with potential investors reportedly envisions a $5 billion to $6 billion round with Nvidia contributing roughly $2.5 billion, positioning it as one of the round’s largest investors. An alternative account says Thinking Machines Lab is in talks with existing investor Accel to lead a new round of at least $1 billion at a valuation near $40 billion, with Nvidia still weighing participation. Both scenarios were described as possible and the final size and structure remain unresolved.
Public reports place Nvidia’s recent model‑layer moves alongside other large commitments: an approximately $5 billion investment in SSI on July 28, an association with Perplexity in late August where reports cited a negotiated valuation north of $30 billion (Nvidia’s actual investment amount in Perplexity has not been disclosed), and the announced $12.9303 billion acquisition of Hugging Face on September 3. Those deals were cited as spanning different segments of the AI stack, from safety research to search and developer platforms.
Thinking Machines Lab: rapid re‑rating and outstanding questions
Thinking Machines Lab completed a $2 billion seed round last July at a $12 billion post‑money valuation, led by Andreessen Horowitz with participation from Nvidia, GV, Lightspeed and others. If the new round closes at a pre‑money valuation of at least $40 billion, the company’s valuation would have grown roughly threefold in just over a year. The startup, founded in 2025, launched the open‑weight model Inkling and the Tinker customization platform in July and was reported to have annualized revenue already exceeding $100 million.
The company previously attempted to raise capital at a $50 billion to $60 billion valuation earlier this year but that effort failed, and three co‑founders subsequently returned to OpenAI. Market commentary in the reports highlights governance structure and talent attrition as key concerns for investors.
Hardware ties and strategic partnership
Nvidia and Thinking Machines Lab have an existing strategic relationship: in March they announced a long‑term “gigawatt‑scale” partnership under which Nvidia committed to supply at least 1 gigawatt of next‑generation Vera Rubin computing systems, with initial deployment targeted for early 2027. The reports emphasize that capital investment and large‑scale hardware supply are intertwined in the relationship, with compute commitments seen as critical to Thinking Machines Lab’s model training and enterprise customization services while offering Nvidia potential upside beyond hardware sales.
The negotiations are ongoing and the final terms, including the ultimate size of the round and Nvidia’s final participation, have yet to be settled.
