Leonis Capital Closes $25M Fund II to Back AI Builders

A fully subscribed fund backed by institutions and AI researchers, built to invest early in founders with deep technical conviction.

SAN FRANCISCO, CALIFORNIA / ACCESS Newswire / September 16, 2025 / Leonis Capital announced its fully subscribed Fund II of $25M. The fund is backed by institutional LPs and a diverse mix of operators and founders in the AI industry that includes FoFs, foundations, as well as researchers and executives from Nvidia, OpenAI, and Anthropic. This comes three years after raising Fund I, reflecting the firm's steady momentum and strong performance. "Fund II represents a meaningful milestone in our evolution from a specialized boutique to building an institutional-grade investment platform investing in one of the most important industries (in AI) of our lifetime," says Jay Zhao, Managing Partner at Leonis Capital.

Leonis Capital was founded by Jay Zhao, who spent over a decade as a venture investor, investing in enterprise, mobile, fintech, and cloud. He has invested in 11 unicorns, with 3 of those companies going public. Zhao and his previous fund were early backers of Marqeta, the leading fintech company that powers the "on-demand economy" like DoorDash and Uber, and ultimately served on its board before the company went public. In addition, Zhao also backed unicorns like Lime and Workboard since the early days.

In 2020, Zhao observed that most venture funds were increasingly chasing rounds and clout while over-indexing on consensus signaling instead of first-principles thinking. "I thought most venture firms were missing the mark on identifying the next wave of transformative AI companies because they were becoming disconnected from the technical fundamentals of a business, leading them to potentially overlook the most innovative builders," says Zhao. He believed that to build an intergenerational firm investing in AI, the firm needed to be structurally technical and research-driven, capable of making early contrarian bets based on deep technical understanding rather than market momentum.

To realize this vision, Zhao teamed up with Jenny Xiao, who was previously an early employee and researcher at OpenAI and received her PhD from Columbia University. Zhao brings the institutional investment discipline and company-building experience of a seasoned venture investor, while Xiao brings the technical depth and research mindset essential for identifying genuine breakthroughs in artificial intelligence. In addition, Leonis also has 8 research fellows that includes PhDs and researchers from Stanford, MIT, Berkeley, Harvard, etc, helping the firm in advancing fundamental research.

"We don't just talk about technologies, we immerse ourselves in them," explains Xiao. Through their research-driven approach, Leonis Capital is able to generalize their learnings into a thesis-driven approach that is continually evolving as the AI landscape changes. The firm regularly publishes comprehensive analysis and research pieces that are widely circulated among Silicon Valley AI founders, VCs, and researchers. Some of their more notable works include their "Zero or Hero" essay, a 10,000-plus-word analysis that outlines how the firm values foundation models and AI applications, distinguishing between incremental improvements and genuine breakthroughs, and their 40+ page report profiling OpenAI's evolution from a non-profit research lab to the "Everything Platform" in AI. Their approach feels more like academic researchers developing and testing hypotheses than traditional VCs chasing deals.

"The Leonis team has clear and well-thought-out ideas about where AI is going and why. That stands out in an industry that is sometimes prone to hype chasing instead of deep thinking," said Todor Markov, an early OpenAI researcher now at Anthropic, and also at Leonis Capital LP.

The firm's research-driven approach has already yielded impressive results in its first fund.

Fund I, a 2021 vintage fund, is already performing in the top 1% of venture funds in its vintage, according to Pitchbook data. The portfolio has already generated notable returns. Some of Fund I's early bets have also achieved unicorn status, with its biggest hits sitting at 50x - 200x markups and some portfolio companies have also exited to big tech giants like Nvidia and Google.

Standout investments include MaintainX, an AI-powered maintenance operations ERP platform that was later backed by Bessemer and Bain Capital Ventures, reaching a $2.5B valuation; Motion AI, an AI-first project management platform, has gained significant revenue traction since Leonis' seed investment and later attracted investment from Sam Altman and Scale Venture Partners. Similar trajectories apply to other Fund I portfolio companies.

"The Leonis team engaged with our AI roadmap at a technical depth that distinguished them from most other VCs we have interacted with. Their ability to connect technical possibilities with market realities helped us build something that's not just innovative but commercially viable, the type of guidance we truly appreciate," says Harry Qi, Motion AI's founder.

Beyond the impressive investment results, the Leonis team is laying the foundation to build an institutional-grade platform. "For us, Fund II represents an opportunity to not only hone our research-driven methodology but to start enhancing our decision quality with the latest AI tech stack," says Zhao. The firm has taken steps to build internal software tools and frameworks that systematize its approach at scale. "We are building an AI system that we internally call "Alfa," trained on all of our thinking, success, learning, and experience. We expect Alfa to improve over time as we build out more technology, like our proprietary signal tracking systems that monitor GitHub contributions, academic papers, and research conference presentations," explains Xiao. "Systems like Alfa help us identify promising technical talent before company formation, it also helps us engage in more in-depth conversations with technical founders when an important AI-native building block becomes available for builders (such as multi-agent workflow, MCP, etc)."

"What impressed me most about Leonis is how rigorously they think," says Stefano Corazza, Head of AI Research at Canva and former Head of AI at Roblox. "As both a researcher and operator, I see their technical depth and thesis-driven discipline as rare in venture, and exactly what the next generation of AI founders need."

So far, LPs have been impressed not only with Leonis' returns but with the knowledge edge they gain from the firm's research-driven approach. "Leonis has built one of the most technically sophisticated investment platforms I've seen: deeply technical, methodical in research, and unafraid to go early when it counts. From an LP perspective, they're not just keeping up with the frontier, they're helping define it," says serial founder and former AI lead at Nvidia Jeff Tseng, who is also an LP in the fund.

Zhao and Xiao both understand how quickly the AI space is moving and how profound and lasting its impact will be on society. "With each technology paradigm shift, there is a need for a new class of capital providers, ones who truly understand the underlying physics of the change," reflects Zhao. "The best founders in AI are looking for more than capital, and as a firm, we see ourselves as the first partner for technical founders. We lead with conviction, and we are research-driven at our core. Ultimately, that is what it takes to build an enduring institutional venture franchise in the AI era."

About Leonis Capital: Leonis Capital is a research-driven venture firm based in San Francisco, focused on AI-native companies. Since 2021, it has partnered with technical founders at the earliest stages, combining deep technical expertise with institutional investing experience.

SOURCE: Leonis Capital

Information contained on this page is provided by an independent third-party content provider. XPRMedia and this Site make no warranties or representations in connection therewith. If you are affiliated with this page and would like it removed please contact [email protected]