Former employees who left OpenAI have already founded seven unicorn companies.
A new big - company faction has emerged in the AI startup circle, and this time it's the well - known American artificial intelligence company OpenAI, the developer of the globally popular ChatGPT.
The company was founded at the end of 2015 and is committed to the research of artificial intelligence technology. At its inception, it was positioned as a non - profit organization. It wasn't until March 1, 2019, when it established the subsidiary OpenAI LP, that the company began to transform into a for - profit organization. Later, with the support of Microsoft and many top venture capital firms, it grew into a super unicorn.
As eye - catching as their ace product ChatGPT is the upheaval in OpenAI's top - level management team and the reorganization of its technical team.
Ask ChatGPT about this story, and it can tell it very vividly. We won't go into details here.
IT Juzi has observed that currently, the core members (mainly senior executives) who have left OpenAI have founded multiple AI companies, and surprisingly, at least 7 of them have become unicorn enterprises.
Companies such as Anthropic, Perplexity, and SSI have frequently secured billions - of - dollars in financing, and their valuations have skyrocketed.
These companies focus on areas such as AI security, education, and search. Some have become direct competitors of OpenAI, while others promote the diversified application of technology, demonstrating the industry influence of the talent spill - over from OpenAI.
IT Juzi will focus on introducing the situations of 7 of these unicorn companies:
1. Thinking Machines Lab, which became a unicorn just after its establishment this year
The founder of Thinking Machines Lab is Mira Murati, the former Chief Technology Officer of OpenAI.
She was born in Albania in 1988, moved to Canada at the age of 16, and graduated from Dartmouth College in the United States with a major in mechanical engineering in 2012. She once worked as an intern analyst at Goldman Sachs, a senior concept engineer at the French aerospace company Zodiac Aerospace, and also served as a senior product manager for Tesla's Model X, leading the development of the Autopilot assisted - driving software.
After joining OpenAI in 2018, she was promoted step by step and led the research and development of signature products such as ChatGPT, DALL - E, Codex, and GPT - 4. She is known as the "Mother of ChatGPT". In 2022, she was promoted to the Chief Technology Officer of OpenAI and left to start her own business in February 2025.
Currently, her team consists of about 30 people, two - thirds of whom are from OpenAI, including co - founder John Schulman and former research director Barret Zoph.
Thinking Machines Lab is committed to building artificial intelligence models and products that support "human - AI collaboration". In June 2025, the company completed a $2 billion seed - round financing, led by a16z, with participation from NVIDIA, AMD, Accel, etc. Its valuation is about $12 billion.
2. Anthropic, the company with the highest valuation and in direct competition with OpenAI
Anthropic was founded in 2021 by a group of former OpenAI researchers and engineers. The core team left due to differences in OpenAI's approach to AI safety and ethics.
Among them, Dario Amodei serves as the CEO. He once served as the vice - president of research at OpenAI and participated in the development of GPT - 2 and GPT - 3. Dario's sister, Daniela Amodei, serves as the president. She once served as the vice - president of safety and policy at OpenAI and now focuses on aligning Anthropic's business goals with its social mission.
Anthropic is registered as a "Public Benefit Corporation (PBC)" and is committed to building reliable, interpretable, and controllable artificial intelligence systems, especially emphasizing the alignment of AI safety with human values.
Anthropic's flagship product is the Claude series of large language models (LLMs), which are designed to compete with OpenAI's ChatGPT. The company provides Claude services to enterprises through APIs and cloud platforms (such as AWS Bedrock and Google Cloud's Vertex AI), which are applied in areas such as customer support, coding tools, search, and research.
Image source: Claude client
In March 2025, after Anthropic completed a $3.5 billion Series E financing, its cumulative financing reached $18.2 billion, and its valuation was $61.5 billion . The main investors are Google and Amazon. The financing will be used to advance the development of the next - generation AI system, expand computing power, and carry out international expansion.
In July this year, it was reported that Anthropic is close to completing a new round of financing worth between $3 billion and $5 billion, led by Iconiq Capital. After this round of financing, its valuation could reach as high as $170 billion.
3. Safe Superintelligence (SSI), which has no product yet but is in the limelight
Ilya Sutskever, the former co - founder and former chief scientist of OpenAI, founded SSI in June 2024. The co - founders of the company also include Daniel Gross, the former head of AI at Apple, and Daniel Levy, a former OpenAI researcher.
Right after its establishment, SSI received $1 billion in financing, with a valuation of $5 billion. The investors are top Silicon Valley venture capital firms such as a16z and Sequoia Capital.
In April 2025, SSI completed a $2 billion financing, with a valuation of $32 billion, led by Greenoaks, with participation from Alphabet (Google's parent company), NVIDIA, Lightspeed Venture Partners, DST Global, etc.
With the rapid development of AI technology, the public and regulatory authorities are increasingly worried about the potential risks of AI (such as loss of control or conflict with human interests). SSI, with its sole goal of "safe superintelligence", meets the market's urgent need for safe and ethical AI. Coupled with the background of its founding team, it has attracted the support of top venture capital firms, making it in the limelight. Its valuation has reached an extremely high level just one year after its establishment.
Currently, there are two main points of controversy surrounding SSI:
One is the controversy over its high valuation. Although SSI has about 20 employees and no revenue or products, its valuation is as high as $32 billion, which has raised some doubts. On platforms such as Reddit and Hacker News, some users are confused about this "high - valuation without products" situation, believing that its valuation is more based on hype and Ilya Sutskever's personal reputation rather than actual achievements.
The other is the lack of transparency in its actual progress. SSI has not publicly disclosed any details about its technology or products. Its website only contains a simple statement of the company's mission, lacking a specific roadmap. This mystery has led some people to question its feasibility and transparency.
4. Elon Musk and his founded xAI, which is making waves in both the political and business worlds
In 2015, Elon Musk co - founded OpenAI with Sam Altman and others. In 2018, due to differences in concepts and disputes over control rights, Musk broke away from OpenAI. In 2023, Musk founded xAI with the goal of developing an open - source AGI "oriented towards truth", directly competing with OpenAI.
xAI has released Grok, a product competing with ChatGPT. By integrating the real - time social data stream from the X platform (formerly Twitter), Grok - 3 can instantly capture hot events on the X platform to generate answers.
Image source: Grok official website
However, Musk doesn't position xAI as a disruptor of OpenAI but rather aims to regain the original intention of "OpenAI". xAI adheres to the strategy of "maximum open - source". For example, the Grok - 1 model is open - sourced under the Apache 2.0 license, attracting global developers to participate in the ecosystem construction. This is in sharp contrast to OpenAI's recent tendency towards closed - source (such as GPT - 4 only providing API services).
5. Perplexity, the AI search unicorn
Aravind Srinivas worked at OpenAI for one year as a research scientist until he left in 2022. Later, he co - founded the AI search engine Perplexity with others.
Perplexity uses large language models (such as GPT - 4, Claude, Llama, etc.) in combination with real - time web retrieval to analyze and generate concise and accurate answers and provides the original links. In addition, Perplexity has also developed precise search functions in three specific fields: finance, travel, and academia.
Image source: Perplexity official website
Perplexity has attracted a series of well - known investors such as Jeff Bezos and NVIDIA. Its most recent financing was in early 2025, raising about $500 million, and its post - investment valuation is about $14 billion.
According to reports, Perplexity is currently in talks with mobile device manufacturers (OEMs) to pre - install its newly launched Comet AI mobile browser on smartphones. It is reported that OpenAI is also developing its own AI agent browser, which can automatically perform complex tasks such as travel booking and financial management.
6. Adept AI Labs, the Agent unicorn that experienced the departure of its founding team
David Luan, who once served as the vice - president of engineering at OpenAI, left in 2020. After a brief stint at Google, he co - founded Adept AI Labs in 2022.
Adept's core product concept is to develop an "AI teammate" that can perform tasks on existing software tools and APIs through natural language instructions, similar to enhanced robotic process automation (RPA) but more general and user - friendly.
Adept's flagship product is the AI assistant ACT - 1. It understands users' needs through natural language and can automatically perform operations in various applications such as browsers, recruitment software, and spreadsheets, such as automatically organizing data, filling out forms, and importing information, greatly improving work efficiency. ACT - 1 works in a "overlay window" mode, operating directly on existing software without users having to switch applications.
In 2023, Adept completed a $350 million financing, with a valuation of over $1 billion, led by General Catalyst and Spark Capital, with participation from strategic investors such as Microsoft and NVIDIA.
However, in June 2024, when Adept's core team (including co - founders David Luan, Ashish Vaswani, and Niki Parmar) joined Amazon's AGI department, the company's main technology also flowed into Amazon.
The remaining team of Adept consists of about 54 people, led by the new CEO Zach Brock (head of engineering) and product director Tim Weingarten, and continues to develop agent AI products.
Meanwhile, Adept has adjusted its strategy. Since developing basic models requires a large amount of capital, it has decided to focus on agent AI solutions and abandon some basic model research and development.
7. Cresta, jointly founded by Chinese entrepreneurs
Tim Shi, the co - founder of Cresta, is one of the early members of OpenAI. According to his LinkedIn profile, he worked at OpenAI for one year in 2017, focusing on the construction of safe AGI. He has also worked on deep learning and machine learning - related projects at companies such as DJI and Dropbox. In 2019, Tim Shi joined Cresta as a co - founder and CTO.
Even earlier, Tim Shi graduated from Tsinghua University's Yao Class (ranked first) and later pursued a doctorate in artificial intelligence at Stanford University, focusing on natural language processing and reinforcement learning.
Cresta focuses on AI contact center solutions and is also one of the earliest companies globally to apply large - scale generative AI models (such as GPT) to the enterprise production environment.
Similar to Adept, Cresta mainly serves the enterprise market, and its customers cover industries such as retail, telecommunications, and automotive.
It is understood that Cresta has raised more than $270 million in total from top venture capital institutions such as Sequoia Capital and Andreessen Horowitz. Its most recent financing was a $125 million Series D financing in 2022, with a valuation of over $1 billion, maintaining its status as a unicorn.
However, a class - action lawsuit in California in June 2023 accused Cresta of capturing and analyzing customer call data without consent, causing a privacy controversy that may affect its brand reputation.
Compared with established American technology companies such as NVIDIA, Microsoft, and Apple, the young OpenAI is just a startup. However, its ability to "create" unicorn companies is so powerful.
It can be said that this has created the spectacle of "one - person unicorn" - the net worth of a former core researcher at OpenAI can be approximately equal to the valuation of a startup or even a unicorn enterprise. The "gold content" is really high. Of course, behind this, capital is undoubtedly a huge driving force.
When large models are regarded as a new wave of disruptive technological revolution, no one wants to miss out. When capital rushes to board the ship, bubbles start to form.
We pay tribute to the explorers. And we also wish everyone good luck.
This article is from the WeChat official account "IT Juzi" (ID: itjuzi521), written by Wu Meimei, and is published by 36Kr with authorization.