Thinking Machines in Talks to Raise $6 Billion, The Information Reports
Mira Murati’s Thinking Machines Lab is in discussions to secure funding between $5 billion and $6 billion, according to an updated report by The Information. This significant raise would value the company at a pre-money valuation of at least $40 billion. Accel, an existing investor, is set to lead the investment round.
Nvidia, the chipmaker, will contribute approximately half of the capital, amounting to roughly $2.5 billion to $3 billion. The announcement, sourced from Andreessen Horowitz, reveals a private briefing for limited partners.
Initial Misreporting and Correction
The initial report by The Information on September 3rd mentioned a funding target of over $1 billion, which was later corrected to the more substantial figure. Several tech publications, including TechCrunch, RuntimeWire, and Tech Funding News, echoed the initial report before the update.
The Significance of the Funding
At $1 billion, the funding would not have been sufficient for Thinking Machines’ infrastructure goals. As pointed out by RuntimeWire, a $1 billion raise would not cover even a gigawatt of Nvidia’s Vera Rubin systems, which is part of their partnership.
The updated figure of $5-6 billion changes the dynamics. With Nvidia contributing half, this funding enables Thinking Machines to compete with more established rivals in terms of chips, power, and data center space.
Business Model and Nvidia’s Involvement
Thinking Machines provides open-source AI models, such as Inkling, a 975-billion-parameter system for text, images, and audio processing. They monetize through Tinker, a platform that offers access to GPU clusters for training and fine-tuning these models. Customers retain data and algorithm control via an API, while Thinking Machines manages the hardware and billing.
Nvidia’s backing is presented as support for American open-source AI developers but also has a commercial advantage. Every job running on Tinker utilizes Nvidia-sold hardware.