While working two jobs in her twenties as an engineer at Google and part-time investor at Bessemer Venture Partners, Casey Caruso started to develop an investment thesis: “Tech is changing so fast that if you don’t fully understand it, you kind of just lose the plot.”
Based on this conviction—that investors who better understand the increasing complexity of technology are poised for success—Caruso, 31, has raised $75 million for a new fund called Topology Ventures. It will invest in frontier tech, which she defines as startups that can “make science fiction reality,” including AI, decentralized networks, neurotechnology, aerospace, and robotics. Cendana Capital and Accolade Partners anchor the fund, which is three times oversubscribed. Other limited partners include Marc Andreessen of a16z; Bob Goodman, a managing partner of Bessemer; Ethan Kurzweil, a Bessemer alum who now has the fund Chemistry; a cofounder of OpenAI; a coauthor of GPT-4; and Paris Hilton. (Caruso declined to make public which cofounder of OpenAI invested, but says it was his first fund investment and he is still at the company.)
Joelle Kayden, Accolade’s founder and managing partner, calls Caruso a “unique talent” with “technical acumen, grit, and determination to build a generational firm.”
Topology has made six investments so far, including a startup also backed by a16z that is still in stealth mode and addresses the inference time compute or reasoning ability of AI models, a subject raised by OpenAI in its latest model. “A lot of nontechnical investors would maybe overlook this, but because we’re so close to the tech, we really saw this as an opportunity,” Caruso says. “There’s no way we would have made the investment unless we were reading AI research papers.”
In other ways, too, Caruso as an investor stays close to her engineer roots; her specialty at Google was as an engineer on research systems (UX and machine learning) within cloud. She spends a lot of time in hacker houses and often reaches out to founders through Discord or GitHub rather than email. She first began making angel investments in companies including Databricks and did a stint at the crypto investment firm Paradigm. “I think we pick better investments because we can call technical BS and support founders better,” she says.
Amid the AI boom, Caruso has seen the flip side of her thesis proven out—investments that appear to lack technical acumen. “I do see a lot of undisciplined investing in these AI products,” she notes, adding that some investors misinterpreted last year’s GPU shortage, while those who were closer to the technology and its supply chain had a better sense of when the shortage might come to an end—and how to plan and budget accordingly.
Caruso is looking to avoid investments that are only about technology, without a compelling business. “A lot of money is being lit on fire right now into these R&D labs, and we’re interested in finding companies that are leveraging AI to make real businesses,” she says. For the same reason, she has so far avoided investments in the neuro category, even though she’s bullish on it long term.
“I can only build so much with my own one mind and two hands as an engineer,” she says. “Investing was the perfect answer, because you can really scale your impact.”
