Small venture funds (<$750M) are mathematically destined to outperform large funds due to the extreme difficulty of generating the massive exit values required by large capital bases.
The venture capital incentive structure is broken, as it rewards General Partners for fund size via management fees more than for top-tier performance, creating a misalignment with Limited Partners.
The best investment strategy in AI is to focus on enabling technologies (e.g., controllers, physics engines, GPUs) rather than attempting to compete with incumbents on building larger foundational models.
The trend of successful tech companies staying private for longer periods concentrates wealth among an elite group of investors and ultimately harms public retail investors, who are left to buy into potentially overpriced IPOs.
Google possesses the strategic ability to fundamentally undermine the business models of AI competitors like OpenAI and Anthropic by aggressively cutting the price of its model tokens.
Early Google Ventures
During the early days of Google Ventures, Maris recalls that Google's leadership viewed the term 'AI' as science fiction and mandated the use of 'machine learning' to avoid 'freaking people out'.
2009-2018
As founding CEO of Google Ventures, Maris oversaw a period where the fund achieved an estimated 4.1x return multiple and used machine learning to design its portfolio construction and fund size.
Post-Google Ventures
Served as Google's vice president of special projects, where he incubated significant long-term projects including Waymo, Google X, and Calico.
c. 2017-Present
Founded the venture capital fund Section32, raising an initial $150 million. He has since launched six funds, averaging approximately $400 million each, all reported to be performing in the top decile for their vintage.
Current
Actively articulates his thesis that small VC funds outperform large ones, critiques misaligned industry incentives, and predicts a massive, compressed evolution for the AI industry.
▶Venture Capital Mechanics and PerformanceJun 2026
Maris argues that the structure of venture capital, particularly fund size, is a primary determinant of returns. He uses mathematical arguments to assert that small funds (<$750M) are structurally positioned to outperform large funds, which struggle to generate the necessary multiples on massive capital bases. This theme critiques the industry's trend towards mega-funds and misaligned GP incentives.
Investors should scrutinize fund size as a key risk factor and be wary of the 'bigger is better' narrative, as incentive structures may prioritize management fees over absolute returns for LPs.
▶AI Investment Strategy and Market DynamicsJun 2026
Maris advocates for a 'picks and shovels' approach to AI investing, focusing on enabling technologies like GPUs and physics engines rather than competing on large foundational models. He sees the current AI landscape as being in its infancy ('Atari stage') and believes incumbents like Google hold immense power to reshape the market through strategic pricing.
This suggests that the most defensible long-term value in AI may lie in the ecosystem's infrastructure layer, and that competitive moats for model-centric startups are more fragile than they appear.
▶Critique of Tech Capital MarketsJun 2026
Maris is critical of how value is created and distributed in the modern tech ecosystem. He argues that the trend of companies staying private longer concentrates wealth among a small circle of late-stage investors at the expense of public market participants, who are left as 'bag holders' for overvalued IPOs.
This perspective challenges the popular narrative of private market superiority and highlights a systemic risk for retail investors and 401k plans, suggesting a potential future correction or regulatory scrutiny.
▶The Pace of Scientific and Technological InnovationJun 2026
Maris contrasts the exponential pace of innovation he foresees in AI with the more measured, linear progress in therapeutics. He predicts AI will undergo a multi-decade leap in just five years, while biological innovation is constrained by the complexity of human biology and regulatory hurdles like FDA testing, at least until a breakthrough like in-silico cell simulation occurs.
This highlights a key distinction for investors: while both fields leverage computation, the timelines for value realization are fundamentally different, with software-based AI offering faster cycles than biology-based therapeutics.