Samir Kaji: VC Has Diverged into Two Asset Classes
Venture capitalist Samir Kaji pointed out that the venture capital industry has shown a clear "barbell" structure: large top-tier funds and small funds have formed two distinctly different asset categories, requiring limited partners (LPs) to adopt different logics. He also noted that no matter how portfolio strategies are designed, they cannot compensate for poor quality of underlying assets, and returns fundamentally depend on whether one invests in top-tier companies.
He emphasized that AI is rapidly widening the valuation gap, especially in later funding rounds, with capital highly concentrated in a few consensus projects; meanwhile, large VCs are moving towards private equity (PE) models, expanding into secondary markets, credit, and multi-product platforms, leading to increasingly blurred industry boundaries.
Multiple English industry observations and data support this judgment: in recent years, capital has concentrated on a few leading AI companies, with some SPVs generating high fee income through concentrated investments in popular projects, while early-stage funds face return pressures in a high-valuation environment, exacerbating industry divergence.
Source: Public Information
ABAB AI Insight
The "barbell structure" is essentially a result of mismatched capital supply and return distribution. Top funds rely on brand and project acquisition capabilities to secure optimal assets, creating a positive feedback loop; small funds depend on flexibility and early entry to achieve asymmetric returns. Mid-tier institutions lack top resources and lose early cost advantages, leading to a compressed survival space.
AI has intensified this concentration trend. Large models and infrastructure companies require extremely high capital investment, naturally attracting large funds and sovereign wealth participation. Once in the "consensus project" track, capital floods in exponentially, driving up valuations and squeezing future return spaces. This makes "getting into the top tier" the most critical variable, rather than traditional notions of diversified investment strategies.
The shift of VCs towards PE reflects the industry’s search for new return stabilizers. By utilizing secondary markets, credit, and wealth management channels, funds can reduce reliance on a single IPO path while also obtaining management fees and structural returns. This transition signifies that VCs are evolving from "high-volatility equity investments" to "multi-asset management platforms."
Deeper uncertainties arise from the unclear commercialization path of AI itself. The profit structure of foundational models, the moat at the application layer, and the demand for infrastructure all contain uncertain assumptions. When the entire industry allocates capital around an unverified technological paradigm, the misalignment between valuations, narratives, and actual cash flows will gradually be exposed in future cycles.