14 billion. The largest seed round this year has emerged.
Author I Wang Lu
Report I PEdaily of the Investment Community
How crazy is the AI financing?
The Investment Community learned that the AI startup Thinking Machines Lab has completed a $2 billion seed round of financing, with a valuation of $10 billion (approximately RMB 71.7 billion), led by a16z.
What has this company done? Currently, almost nothing is known. It has been established for less than half a year, has no products, and has not disclosed its specific technological direction, yet it has created the largest seed - round financing in history.
What really attracted investors to flock in is the founder "the mother of ChatGPT", Mira Murati, and a luxurious team she assembled - the core members from OpenAI, including Weng Li, a female academic superstar from Xi'an and an alumnus of Peking University, who is well - known in China.
Investment is about investing in people. Looking at this, the AI competition has fully entered the era of "scrambling for talents".
She Leads the Team to Create the Largest Seed Round in History
Let's first look at this latest financing.
According to the disclosure of the Financial Times, Thinking Machines Lab has completed a $2 billion seed round of financing, led by a16z, followed by institutions such as Conviction Partners of Chinese - American investor Sarah Guo, with a valuation of $10 billion.
It is certain that this is currently the largest - scale seed - round financing in history.
You know, Thinking Machines Lab has only been established for half a year, has no products or users, and there is no need to mention revenue. The most eye - catching thing is its extremely luxurious founding team, which can be called the AI "dream team".
Let's go back to last September when Mira Murati, the former CTO of OpenAI, officially left. Even before the new company was established, investors were lining up to visit. Then in February this year, Thinking Machine Labs was officially announced. More than half of the initial team members are from OpenAI -
Weng Li, an alumnus of Peking University, the former head of the security team at OpenAI, and now a co - founder;
John Schulman, a co - founder and the former head of the reinforcement learning team at OpenAI, and now the chief scientist;
Barret Zoph, the former vice - president of research at OpenAI, and now the chief technology officer;
Andrew Tulloch, an expert in machine - learning system engineering, who has worked at OpenAI and Meta, and now the chief architect...
"We are the scientists, engineers, and builders behind the most widely used AI products." The official website of Thinking Machines Lab is extremely concise. It claims to be an artificial - intelligence research and product company, aiming to make artificial - intelligence systems more widely understood, customizable, and have general capabilities.
According to Murati, the company is doing three things: first, helping people adapt large AI models to meet their specific needs; second, building a solid foundation for building more powerful large models; third, cultivating an open - science cultural atmosphere to help the entire industry understand and continue to improve. In addition, she also previewed that new products will be announced in the next few weeks.
However, except for the vision, the company still remains mysterious and has not publicly disclosed any technological roadmap for discussion. According to insiders, some venture - capital institutions decided to give up investing because the company did not provide product information and financial plans during the roadshow.
Even so, the AI "celebrities" are still popular. At the beginning of the year, market news said that the company's financing amount was about $1 billion, and within just a few months, the financing amount has doubled. Simply put, existing investment institutions such as a16z are willing to place a big bet on the founder's past resume. This round of financing also gives Murati extraordinary control over the company. Her voting rights on the board exceed the sum of all other directors, ensuring the final decision - making power on all key company decisions.
This is a typical Silicon Valley bet: in the eyes of investors, in the AI battlefield, top - level talents are scarce resources. Those who have created large AI models may create miracles again, or even do better.
They Join Hands to Start a Business, and a Female Academic Superstar from Xi'an is the Co - founder
Mira Murati, known as "the mother of ChatGPT", is not yet 37 years old this year.
Data shows that she was born in the Eastern European country Albania. She has been interested in mathematics and physics since childhood and has participated in the Olympic competitions many times. At the age of 16, she won an academic scholarship from the United World Colleges (UWC) and went to Pearson College in Canada to study. After graduating in 2005, she studied in the United States and successively obtained a Bachelor of Arts degree from Colby College and a Bachelor of Engineering degree from Dartmouth College. She once interned at Goldman Sachs.
After graduation, Murati first joined Tesla as a product manager for the Model X and then worked at Leap Motion. During this period, she deeply realized that VR had not been accepted by the public. In order to find technologies to solve the world's biggest challenges, she joined OpenAI in 2018 and was promoted to CTO four years later. It was under her promotion that ChatGPT left the laboratory and began to be open to the public.
In a previous interview, Murati believed that the ultimate goal of AI technology is to serve humanity, so it should be centered around human interests and needs to solve the practical problems faced by humans. Time magazine called her "the creator of ChatGPT".
She became well - known in November 2023. In the globally - shocking "power - struggle" incident, Sam Altman was removed from his position, and Murati briefly served as the CEO of OpenAI.
Less than a year later, Murati announced her departure, saying that "she would conduct her own exploration", and took away more than 20 former OpenAI employees, including Weng Li. Thus, the current story unfolds.
Here, we have to mention Weng Li. She grew up in Xi'an, Shaanxi, and attended the famous High School Affiliated to Northwestern Polytechnical University. In 2005, Weng Li was admitted to Peking University, majoring in Information Systems and Computer Science. A few years later, she went to Indiana University Bloomington to pursue a doctorate in computer science.
Around 2017, Weng Li joined OpenAI. Starting as a founding - level employee, she has managed the security team and has become one of the highest - ranking Chinese scientists. During this period, she was responsible for the pre - training, reinforcement learning, alignment, and model security work of the GPT - 4 project, and led the development of the o1 - preview model, effectively resisting adversarial attacks.
Moreover, Weng Li often appears as a researcher. There are many in - depth industry analysis articles on her personal website, covering aspects such as external hallucinations of large models and hacker attacks, which are widely circulated in the industry. Many of her blogs have become popular on GitHub, and the citation count on Google Scholar exceeds 13,000.
What the outside world may not know is that Weng Li has also become a VC. At the end of last year, Weng Li announced that she had joined the well - known Silicon Valley investment institution Fellows Fund as a Distinguished Fellow, providing support for the new - generation AI founders.
Now in the new company, as a co - founder, Weng Li is fighting side by side with her former boss Murati again. She will be committed to promoting the transparency, customizability, and multimodal capabilities of AI systems, focusing on AI security and multimodal research.
There is also an episode: Apple and Meta had acquisition negotiations with Thinking Machines Lab, but were rejected by Murati. Not giving up, Meta then tried to poach the key members of the new company, but also failed.
In Murati's view, everything is just beginning. "We have the same vision. What the future world will look like depends on how this generation promotes the development of AI."
The Global AI Talent War An unprecedented battle for talents is in full swing.
The latest scene occurred this month. Meta invested approximately $15 billion (approximately RMB 107.8 billion) in Scale AI, precisely to recruit its founder Alexandr Wang.
In order to gain the upper hand, Meta also offered a $100 million bonus to attract AI talents from other companies. Sam Altman, the CEO of OpenAI, publicly confirmed this: "To my relief, at least so far, none of our core members have accepted these terms." However, as soon as he finished speaking, four former OpenAI researchers, including Zhai Xiaohua, collectively jumped to Meta. Mark Zuckerberg's personal "poaching" efforts have achieved initial results.
OpenAI is also sparing no effort in talent recruitment. In order to retain researchers who want to jump to companies such as SSI, OpenAI has offered a $2 million retention bonus and more than $20 million in equity incentives.
VC/PEs are also betting crazily. SSI, founded by Ilya Sutskever, a co - founder and former chief scientist of OpenAI, has no products and no users, but has become a unicorn worth tens of billions.
In April this year, Safe Superintelligence (SSI) announced that it had completed a $2 billion financing, with a valuation reaching $32 billion. It has been established for less than a year and has not yet announced its technological details. Relying solely on the reputation of its founder and the company's mission, the investors include well - known institutions such as Google, Nvidia, a16z, Sequoia, and DST Global.
Top - level talents are always in high demand. In fact, there may be only a few hundred top - level talents globally who can truly promote the development of AI. Altman once sighed, "A 10 - times engineer is cool, but those 10,000 - times engineers and researchers are really amazing."
"A year ago, the mainstream narrative in AI was that huge computing power was needed to drive the industry's development. Now, the focus has shifted to the talent advantage." In the view of David Kahn, a partner at Sequoia Capital, when computing power becomes abundant and cheap, relying solely on large - scale computing clusters can no longer support leap - forward breakthroughs. The key to victory lies in whether one can attract and unleash the potential of talents and find new breakthrough points.
Looking back, this change starts with DeepSeek. At the beginning of the year, the open - source large model of DeepSeek emerged suddenly. With the technological breakthrough of "low cost, high performance", it has "skillfully" changed the established rules for the development of AI companies - reconstructing the AI training paradigm with underlying algorithms and reducing the inference cost of large models to one - tenth of the industry level.
It can be said that the previous approach of the giants, which was to "create miracles with brute force" by piling up computing power, has failed to some extent. Remarkably, nearly half of the talents at DeepSeek are fresh graduates from top - level domestic universities, yet they have achieved efficient training by optimizing the underlying code of chips. What really makes the difference is not the original computing - power scale, but the talents.
In this competition for AI talents, the presence of young Chinese faces is becoming more and more prominent. Looking at China, whether it is Yang Zhilin of Kimi, Xiao Hong of Manus, Peng Zhihui of Zhipu Robotics, or Luo Fuli, who is rumored to have been poached by Lei Jun with an annual salary of tens of millions, the promising post - 90s generation is in the spotlight.
Jensen Huang once said bluntly that 50% of the global AI researchers are from China, which is currently the largest single talent group, far ahead of other countries. "Therefore, China's influence in AI research is destined to be profound."
The competition in artificial intelligence will ultimately come back to people themselves. It is those who are at the forefront that lead the direction of the trend - what they are doing, what will explode in the future.
This article is from the WeChat official account "The Investment Community" (ID: pedaily2012), author: Wang Lu, published by 36Kr with authorization.