Jenny Xiao, cofounder of Leonis Capital and former OpenAI researcher, has revealed a striking disconnect between AI research and investment trends, warning that the current AI hype cycle lags 3 to 5 years behind the actual technical frontier. Speaking on the Fortune Magazine podcast, Xiao highlighted what she describes as a “massive disconnect between what researchers are seeing and what investors are seeing.”
Xiao’s unique perspective comes from her journey through the AI ecosystem—she dropped out of a Ph.D. program in economics and AI to join OpenAI as a researcher before founding Leonis Capital in 2021. Her venture capital firm was specifically created to bridge the gap between deep academic AI research and investment capital, addressing what she sees as a critical need in the industry.
According to Xiao, discussions at major AI conferences are significantly behind what researchers are actively working on and thinking about. This temporal gap creates challenges for investors trying to understand where the technology is truly headed. “We are so behind the technical frontier, and that’s the gap I really want to bridge,” she emphasized.
The former OpenAI researcher argues that AI requires a fundamentally different approach to venture capital compared to traditional tech investments. Unlike SaaS companies built on stable technology stacks, AI is evolving rapidly, requiring investors to possess technical depth comparable to the founders they’re backing. “With AI, there needs to be a new generation of founders. There needs to be a new generation of VCs,” Xiao stated, noting that this is the first time investors need to provide financial support to both the market and the underlying technology simultaneously.
Xiao also offered crucial advice for investors navigating the AI landscape: AI progress isn’t linear. Instead, it happens in “lumps,” making questions about whether progress is slowing or accelerating overly simplistic. “It’s neither of those two extremes,” she explained. “It’s somewhere in between.” This non-linear nature of AI development means investors need deeper technical understanding to properly evaluate opportunities and avoid being misled by surface-level trends or temporary plateaus in visible progress.
Key Quotes
There is a massive disconnect between what researchers are seeing and what investors are seeing
Jenny Xiao, cofounder of Leonis Capital and former OpenAI researcher, highlighted the fundamental gap between cutting-edge AI research and investor understanding, suggesting this disconnect could lead to misallocated capital and missed opportunities.
We are so behind the technical frontier, and that’s the gap I really want to bridge
Xiao explained her motivation for founding Leonis Capital, emphasizing that even major AI conferences lag years behind what researchers are actively working on, creating a critical need for venture capitalists who can bridge this knowledge gap.
With AI, there needs to be a new generation of founders. There needs to be a new generation of VCs
Xiao argued that AI’s rapid evolution requires a fundamentally different approach to venture capital, where investors must be as technically sophisticated as the founders they back—a departure from traditional VC models.
AI progress isn’t linear. It’s neither of those two extremes. It’s somewhere in between
Xiao cautioned investors against simplistic narratives about AI progress slowing or accelerating, explaining that development happens in “lumps” rather than steady increments, requiring more nuanced technical understanding to properly evaluate.
Our Take
Xiao’s insights reveal a troubling maturity gap in AI investment that could create both risks and opportunities. Her warning about non-linear progress is particularly prescient—we’ve already seen this pattern with transformer models and large language models, where years of incremental research suddenly crystallized into breakthrough capabilities. The 3-5 year lag she describes suggests that what researchers are working on today could fundamentally reshape the AI landscape by 2028-2030, yet current investment theses may not account for these developments. This also raises questions about whether the current AI boom is actually pricing in outdated assumptions. For founders, this creates an opportunity to work with technically-savvy investors like Leonis Capital who understand the true frontier. The broader implication is clear: technical depth is no longer optional for anyone serious about AI investment or strategy.
Why This Matters
This revelation from a former OpenAI insider turned venture capitalist exposes a critical information asymmetry in the AI investment landscape that could have significant implications for capital allocation and startup valuations. If investors are indeed 3-5 years behind the research frontier, they may be overvaluing yesterday’s breakthroughs while missing tomorrow’s opportunities—or conversely, underestimating risks from rapidly advancing capabilities.
Xiao’s emphasis on non-linear AI progress challenges the prevailing narrative around AI development timelines and suggests that both AI optimists and pessimists may be misreading the trajectory. For businesses planning AI strategies, this means relying solely on publicly discussed developments at conferences could leave them strategically disadvantaged. The call for a “new generation” of technically-savvy VCs also signals a potential shift in how AI companies will be funded and evaluated, with deeper technical due diligence becoming essential. As AI continues to reshape industries, this gap between research reality and investment perception could determine which companies and investors successfully navigate the next wave of AI innovation.
Related Stories
- OpenAI Lost Nearly Half of Its AI Safety Team, Ex-Researcher Says
- OpenAI’s Competition Sparks Investor Anxiety Over Talent Retention at Microsoft, Meta, and Google
- CEOs Express Insecurity About AI Strategy and Implementation
- The Rise of AI Distillation and Its Impact on Big Tech’s AI Dominance
- Groq Investor Warns of Data Center Crisis Threatening AI Industry