The highest seed round in history goes to AI: Leading by a huge margin of $2 billion, even tech giants like Apple and Meta can't get a piece of the action, and Mark Zuckerberg's attempt to poach a co-founder has also been rebuffed.
The AI industry is in a complete mess with financing, acquisitions, and talent poaching!
According to the latest news, Thinking Machines, founded by the former CTO of OpenAI, has completed a $2 billion seed round of financing. This not only breaks the financing record in the AI field but also becomes the largest seed - round financing in history.
Initially, market rumors suggested that the company's financing target was about $1 billion, but the target amount doubled within just a few months.
Even before this, both Apple and Meta had attempted to invest in or acquire Thinking Machines, but both were rejected by the founder, Mira Murati.
Among them, the negotiation between Apple and Thinking Machines was put on hold temporarily before it entered the in - depth stage.
But Mark Zuckerberg of Meta didn't give up. After the acquisition attempt failed, he turned to poaching talent, making the recruitment of AI talents like the NBA transfer mode.
After all, the core competitiveness of Thinking Machines, which has obtained more than expected financing, lies in its team. About two - thirds of its members are from OpenAI. Key figures include:
John Schulman, co - founder and former head of the reinforcement learning team at OpenAI
Li Weng, an alumnus of Peking University, former head of the security team at OpenAI
Barret Zoph, former vice - president of research at OpenAI
In addition, Thinking Machines has also hired two veteran figures from OpenAI as advisors, namely Alec Radford, the first author of GPT - 1, and Bob McGrew, the former head of research at OpenAI.
It is reported that Zuckerberg tried to poach John Schulman but failed. He is now the co - founder and CTO of Thinking Machines.
In addition, during this period, Zuckerberg also tried to poach Bill Peebles, the author of Sora at OpenAI, and attempted to acquire Perplexity, but both attempts failed. Eventually, he acquired Scale AI at a sky - high price.
Now, Thinking Machines has officially announced the completion of the financing, putting an end to this chaotic acquisition and talent - poaching war.
Breaking the historical record of seed - round financing
Before Thinking Machines, the largest seed - round financings in history were between $200 million and $450 million, and there was no one close to it. It leads by a large margin.
The second - largest was Yuga Labs, which raised $450 million in seed financing in 2022. This company mainly focuses on digital collectibles and NFTs. In addition, several other companies that received $200 million in seed - round financing are all related to blockchain.
The largest seed - round financing in the AI industry before was Lila Sciences, which raised $200 million in March this year. It is committed to building a fully autonomous laboratory for life, chemistry, and materials science.
Among the founding team of Lila Sciences, the chief scientist George Church is a pioneer of CRISPR technology.
Let's get back to today's protagonist, Thinking Machines.
This $2 billion seed - round financing was led by Andreessen Horowitz, with other investors such as Accel and Conviction Partners following. The valuation has reached $10 billion.
After the completion of this round of financing, Murati's voting rights on the board of directors will exceed the sum of all other directors, ensuring that she has the final say on all key decisions of the company.
However, there is still little information about the research and products the company is currently working on.
When it was announced in February this year, it stated that it would break the current "one - size - fits - all" model of AI and focus on developing a highly customizable, multi - modal AI system that supports human - machine collaboration.
Different from pursuing fully autonomous AI agents, Thinking Machines Lab focuses on developing multi - modal systems that can work in collaboration with human experts. They believe that the potential of AI lies in enhancing rather than replacing humans. Their models will be able to adapt to a wider range of human professional fields from science to programming, rather than being limited to the narrow tasks that current models are good at.
The company positions itself as an advocate of open science, which is in sharp contrast to the closed ecosystems of many top AI laboratories. They plan to promote cooperation and progress in the entire research community and build public trust in AI technology by frequently publishing technical papers, open - sourcing code, sharing data sets, and model specifications.
The only available information so far:
Murati told investors that they plan to develop customized AI models based on enterprise business KPIs and plan to launch consumer - oriented products. Investors who have talked with Murati described their technology as "reinforcement learning for enterprises".
Reference links:
[1]https://news.crunchbase.com/venture/biggest-seed-round-ai-thinking-machines-mira-murati/
[2]https://www.bloomberg.com/news/newsletters/2025-06-22/will-apple-buy-an-ai-company-apple-takes-a-close-look-at-acquiring-perplexity-mc7m98q9
[3]https://www.wsj.com/tech/ai/meta-ai-recruiting-mark-zuckerberg-5c231f75
This article is from the WeChat official account "QbitAI". Author: Focusing on cutting - edge technology. Republished by 36Kr with authorization.