Andrej Karpathy, an OpenAI co-founder and research scientist who has quit the A.I. powerhouse twice, has been making waves in the startup world since his last exit a year ago. Earlier this month, Karpathy participated in a $480 million round in Lambda, an A.I. infrastructure startup backed by Nvidia and Cathie Wood’s Ark Invest, among others. Lambda is the latest addition to Karpathy’s growing portfolio of A.I. startups, many of which focus on lucrative subsets of the technology, such as compute infrastructure and agent A.I. Since leaving OpenAI, Karpathy has also founded an A.I. education platform that offers university-level courses on A.I. topics.
Karpathy is one of the leading voices in A.I. education, known for blending cutting-edge research with practical insights through his YouTube channel courses (701,000 subscribers). He also has a strong presence on GitHub, with nearly 100,000 followers.
Karpathy’s interest in A.I. began at the University of Toronto, where he studied computer science and physics and attended Geoffrey Hinton’s deep learning classes, later earning a Ph.D. at Stanford under Fei-Fei Li. As one of OpenAI’s 11 co-founders, he played a key role in developing the GPT model before leaving in 2017 to lead Tesla’s Autopilot effort. He returned to OpenAI in February 2023 as a research scientist, building a small team to enhance GPT-4’s capabilities, but departed again in February 2024.
Here are six startups Karpathy either co-founded or invested in:
Eureka Labs
Karpathy founded Eureka Labs last summer after leaving OpenAI with the goal to make A.I. education accessible for all. The platform’s flagship course, LLM101n, offers undergraduate-level technical education to students for training and developing their own A.I. models. LLM101n is currently available for free on GitHub as a repository for educational purposes.
Lambda
Earlier this month, Karpathy participated in a $480 million Series D round in Lambda, a startup that provides high-performance computing services, such as Nvidia’s GPU clusters, for A.I. developers, researchers and Fortune 500 companies. Its clients include Apple, Intel, Microsoft, and China’s Tencent.
Experts say the onset of companies like Lambda reflects an industrywide shift from model providers to cloud compute platforms as key value drivers in the A.I. industry. “API prices from OpenAI and Anthropic continue to decline, approaching ‘at-cost’ pricing. Unlike model providers facing downward pricing pressure, cloud compute platforms remain closer to price setters, benefiting from rising A.I. adoption,” Kang said. “Karpathy’s investments reflect this shift—betting on infrastructure and applications positioned to capture value as A.I. scales.”
/dev/agents
In December 2024, Karpathy invested in /dev/agents, a cloud-based development platform for A.I. agents to streamline coding and software engineering. The startup raised $56 million with a valuation of $500 million. The company was founded by former Google and Meta executives, including David Singleton and Nicholas Jitkoff, who have deep expertise in building large-scale technology platforms.
Lamini
In May 2024, Karpathy participated in a $25 million series A round in Lamini, a startup that helps enterprise clients create custom LLMs. It was founded and led by Sharon Zhou, a former machine learning product manager at Google Cloud. The startup, valued at $325 million, is backed by investors including Dropbox, LVMH, and Google Brain co-founder Andrew Ng., under whom Zhou studied for her Ph.D. in A.I. at Stanford.
Perplexity AI
In April 2024, Karpathy joined Perplexity AI’s $63 million Series B funding round alongside Jeff Bezos, Nvidia and Stanley Druckenmiller. Founded by former OpenAI researcher Aravind Srinivas, Perplexity AI is known for its “answer engine,” which aims at answering user questions by summarizing information on the internet, instead of just providing links in response to a few key words, making it a popular tool for research and fact-checking. The company has raised $915 million across seven funding rounds and is valued at $9 billion.
Adept
Adept, founded by former OpenAI researcher David Luan in 2022, develops A.I. assistants that can understand and execute complex tasks with high precision. Its enterprise A.I. tools can automate repetitive workflows across software applications. In April 2022, Karpathy joined Adept’s $65 million round as an angel investor. The startup has raised $415 million in total funding and is valued at over $1 billion.
<