In 2014, Australian entrepreneur Tuhin Srivastava had scored a meeting with Sarah Guo, then the youngest partner at storied VC firm Greylock. He was pitching her on a healthcare startup that used machine learning to analyze a person’s medical history. Guo was impressed — not by the idea, which was “generic,” she says, but by him and his cofounder. Five years later, he tried again, this time with something far more promising: tools that make it easier to build and run AI applications.

It was years before the stunning launch of ChatGPT mainstreamed artificial intelligence, but Guo was already confident that more businesses would soon turn to AI and need cheap and efficient ways to use it. She wrote a $1.5 million check into what became Baseten, co-leading the startup’s $3 million seed round in 2019. “All we had was some idea of a company on scratch paper,” Srivastava says.

For the first four years, the company made no money. AI tools weren’t being rapidly adopted back then. The best thing to do was to wait for the market to come around. Almost overnight, it did.

Today, Baseten is valued at $5 billion, with revenue growing over 10 times in the past year. (It’s now reportedly in talks to raise at an $11 billion valuation.) Guo invested in every round, first from Greylock and then from her own VC firm Conviction, which she launched in October 2022. Today, she says her stake is worth 10 times its initial value. “We own the most from day zero and it's clearly going to be a winner company,” says Guo, 36.

Her stakes have grown more than tenfold in other portfolio crown jewels: legal AI company Harvey, ChatGPT for doctors OpenEvidence, AI-powered customer service firm Sierra, AI coding startup Cognition and open source AI developer Mistral. Combined, these six startups alone are worth $62 billion.

Of all of Guo’s prescient bets, her most notable one is on artificial intelligence itself. She realized early that the rise of AI would upend beliefs long held by large VC firms—that robotics would never work, that it’s difficult to sell software to lawyers and doctors, that you shouldn’t do scientific research in venture-backed companies, Guo says.

“There are a lot of priors that come from traditional venture about both markets and how you build companies that we thought would be challenged,” Guo tells Forbes.

In 2022, she wagered her entire career on that thesis, leaving her cushy job at Greylock to launch AI-focused early stage VC firm Conviction with a $101 million inaugural fund — one month before ChatGPT made AI’s potential obvious to the masses. Now, that gamble has clearly paid off. This year, Guo makes her debut at No. 56 on the Midas List of the top 100 venture capital investors, and returns to the Midas Seed list for the second time.

In an era of spray-and-pray investing, Guo decided to go the other way: making fewer, more targeted bets. At Conviction, Guo has invested in just 27 startups and taken board positions at only six of those in the last three years. “When you work on early stage, you make a concentrated commitment to a person and an overall idea and then you suspend disbelief and work on the company for a long time,” she says. “It's satisfying to see the people grow and be right.”

That’s allowed her to be more involved with the startups she backs. That doesn’t just mean being available for strategy phone calls or instantly replying on email. In one instance, Guo flew across the country to help keep one of Baseten’s customers from ditching them for a rival. In another, she convinced Nvidia to pour $150 million into the company as part of its $300 million Series E round. No wonder Baseten CEO Srivastava fondly calls Guo a cofounder of the company.

Harvey CEO and cofounder Winston Weinberg says after he’d sent thousands of cold messages to law firms on LinkedIn, it was Guo who introduced him to a lawyer at A&O Shearman, which became the startup’s first client. “I don't know if we even would've gotten that customer without her,” Weinberg says. When the company decided to “rip apart” its entire product to build agents, Guo helped connect him to experts who could show them how to do it, Weinberg says.

“Because she's so plugged in, she's just very, very early with this stuff and she has personal relationships with all these founders,” he says. “And it's not arm's length. She's actually friends with them.”

Every two months, Guo hosts a networking and recruiting event for startups and young professionals called “Mixture of Experts.” She also runs a podcast called No Priors with solo venture capitalist Elad Gil, ranked at No. 6 on this year’s Midas List, whose high-profile guests range from pioneers in the field like Dr. Fei Fei Li to heavyweights like Jensen Huang. Discussing ideas with the brightest minds in AI helps inform her understanding of the industry. The challenge, however, is not being swayed by whoever’s got the strongest opinion that week.

“There's narrative and then you're trying to figure out what of that is real,” Guo says. “You have to know when to insulate yourself from the opinions of the broader venture market and the public market because it's a very herd mentality.”

As a kid growing up in Wisconsin, Guo fell in love with computers. At age 10, she built a robot dog because her mother wouldn’t let her have a real one. Her parents, immigrants from China, worked as engineers at research and development company Bell Labs and eventually started their own networking infrastructure startup Casa Systems. It went public in 2017 at about $1 billion valuation (and filed for bankruptcy in 2024). Guo helped build its first website at 14 and began pitching investors on its behalf when she was 19.

Her own entrepreneurial ideas didn’t go very far. While at the University of Pennsylvania, she tried (and failed) to build an app to use GPS to find friends nearby (just like Apple’s). Over five years, she graduated with four degrees: a B.A. and M.A. in Chinese history and literature, a B.S. in economics and an MBA. (Guo also met her husband, Midas List investor Pat Grady of Sequoia, through a friend from Penn.)

Guo worked at Goldman Sachs for a year, helping giants like Netflix, Nvidia, Workday and Twitter go public, then joined Greylock in 2013. “They basically didn't like my company ideas, but they thought that I should come hang out for a year or so,” she says. Despite not having any formal education in computer science, she regularly blogged about deeply technical concepts like long short-term memory networks that got the attention of deep learning legend Andrew Ng.

Her technical prowess shows up today in meetings with startup founders, says Conviction general partner Mike Vernal, who joined the firm in 2025 from Sequoia. During a pitch meeting with chip startup Fractile in April, Guo asked specific questions about the chip’s packaging and how memory would be programmed into the chip, Vernal says. “The team was blown away by the depth of her questions,” he adds. (She invested in their $220 million Series A round.)

To keep pace with the AI market, Guo says she spends a quarter of her time reading research papers and speaking to researchers at OpenAI and Anthropic (both of which were already too large for her to invest in). Every time a company pitches her, Guo asks herself, “Is this company going to be generational in the next five to ten years?” She isn’t always right, but thinking long-term grounds her. “This is a very heady time,” she says. “I don't want to make predictions in a vacuum.”

More from Forbes

ForbesInside The Murky Market Selling Pre-IPO SpaceX And OpenAI SharesForbesThe Midas All-StarsBy Sofia ChierchioForbesInvesting Superstar Yasmin Razavi Turned A $75 Million Check Into A $3 Billion AI WindfallBy Iain Martin