The Future of Venture: Industrialization or Back to Basics?
Justifying massive AUM aggregation with the power law is a fallacy; smaller, craft-driven funds actually outperform the mega AUM model.
TL;DR On the heels of Google, Facebook, and the ZIRP era, a whole generation of venture capitalists narrowed their pattern recognition so tightly that they fueled an age of excess we are still cleaning up. As the dust settled, many of us observed a new era of VC was dawning. However, the approach espoused by some large players (rationalized incorrectly by the power law but actually driven by a desire to aggregate AUM) couldn’t be farther from our worldview. We believe the best venture investors and founders — who ultimately drive sustainable outsized venture returns — will succeed practicing the fundamental craft of venture, which has an asymptotic scale well below the mega-billions being raised today.
In 2022, I wrote a manifesto describing the forces driving change in venture and how this led to Avila VC. In 2023, Hemant Taneja and Fareed Zakaria published a much more eloquent piece on the same topic. Our diagnosis was quite similar, but the remedy diametrically opposed: back-to-basics craft of venture vs. massive conglomerate-style assets under management (“AUM”).
The appeal to a general partner (“GP”) of amassing AUM is obvious — large fixed fees for a long time irrespective of returns. The rationality for a limited partner (an “LP”) less so — asset allocators must pick the best managers for each strategy and size vehicles appropriately to maximize returns and align incentives. If they are not doing so they aren’t allocators, they are abdicating allocation to others.
Imagine a world with only 2 restaurants, McDonald’s and a Michelin-star one. If you could only invest in one, you should certainly pick McDonald’s. But if you were searching for the best meal the latter is certainly better. As a “meal allocator”, your optimal strategy would be to eat primarily at Mickey D’s, and go to the Michelin-star for infrequent special occasions. Meshing them both together would ruin each product and expecting a Michelin-quality product at McD scale would be absurd.
As must be painfully obvious, I believe venture requires a Michelin-star approach to investing. There are only so many elite founders with world-changing ideas, and the craft of finding them, winning their trust, and supporting them on their journey requires skill and devotion. Venture investing requires actual expertise and investment judgment (hence the rightsizing of venture happening now), and is not a mass-producible activity (which should point to right-sized firms). However, the 2/20 model for private investing encourages asset aggregation. It gives the GP a lot more guaranteed funds (the 2% is fixed regardless of returns, and recurring vs. upon distributions), and simplifies the workload for LPs by having fewer relationships. One might imagine that the lessons many purported to learn in 2021 (“the laws of business always apply in the end, venture is a long game, over-paying is not a sustainable strategy, we must do better with oversight, etc.”) would have lasted a little longer. But while we are still dealing with the cleanup from the last bubble, investors have allowed the excitement of AI to drive capital concentration in a way unlikely to lead to long term venture scale returns. Jamin Ball’s excellent post explains the misaligned incentives.
The Fallacy of Justifying AUM due to the Power Law
Many analyses have been used by smart GPs to convince investors that large AUMs will not hurt them. But these are mistakenly interpreted.
The argument usually goes like this: as a VC, it doesn’t matter how many mistakes you make or how much you overpay, so long as you are in the winners you will be a top performer (and conversely if you aren’t in those you cannot be a top performer). However, this is a simplistic misunderstanding of the data.
If ownership and entry valuation didn’t matter, all the “spray and pray” funds which included a big winner alongside a giant amount of crap would be top decile, self-evidently not true
Funds that did not invest early but late into the “unicorns” typically did not do well. Given current entry prices and CapEx demands, early rounds are more comparable to those data points
The data points in these analyses are skewed by: rapid exits within the mis-priced ZIRP-era and a handful of low-capex/high-margin businesses that are unlikely to be replicated in the next wave
The justification that top decile funds could afford to lose money on about 50% of their initial investment dollars worked because entry valuations were such that the winner(s) covered multiple mistaken at-bats. If entry check sizes and valuations balloon, the affordable loss ratio will be much lower
The power law is real, but without discipline it will not deliver sustainable returns. And with a giant bucket of 2% fees, the impetus to work insatiably for the 20% might be muted.
Momentum vs. Patience
Part of the myth to justify mega-funds is that value is increasingly concentrated in a small handful of firms, as evidenced by the trillion+ valuations of AAPL, MSFT, NVDA, AMZN, GOOG, META. And there is much truth to the fact that the biggest outcomes have increased by at least an order of magnitude. However, its also important to recognize that many of these huge successes took decades to reach their current potential, with quite non-linear paths.
Apple nearly went bankrupt in the late 90s. NVIDIA was founded in 1993, went public in 1999, and was worth $2B in 2000. Tesla and SpaceX, both founded in the early 2000s, nearly went bankrupt in the latter part of the decade. FB and Google, while never close to existential risk, also had major valuation fumbles with the transition to the app economy.
What this means for venture is that the right comparison is: “how much value can be created within a 5–15 year holding period?” After that, venture-backed companies that survive move on to other asset classes, be it private equity or public markets. It also means investors need much more patience.
Some Words of Caution re: AI
Like many, I’m enthralled by the transformative possibilities of AI. I believe it can be a super-cycle that will transform our lives as much as previous ones and that will take decades to play out just like others before. I also believe that while history never repeats itself, it does rhyme. We have lived through previous bubbles, each time rationalizing why it’s different but always returning to business and economics fundamentals. It can be BOTH true that we are in an AI bubble because incentives are causing investors and leading players to overspend, AND that AI will be a life-changing force for society over time.
Given the risk-reward of AI and the inherent CapEx required to advance its development, I believe investors should not forget the lessons from the recent correction: picking winners is difficult and requires investment judgment, investors must provide actual oversight, companies take time to nurture, valuations and burn actually matter, geopolitical and macroeconomic factors can affect an investment, etc. — and proceed with caution.
As an aside, and probably worthy of a separate blog post re:AI, its black-box nature and existential risks suggest we should focus on “human-assisted” AI as a stepping stone before we turn over our businesses and knowledge to an indecipherable black box. The fact that models get it wrong more often than right, and that the choices we make wrt AI development today can change the nature of our societies (Yuval Noah Harari’s excellent book Nexus is worth a read on this topic) should drive us towards an implementation that turbocharges human productivity and learning rather than jumps immediately to dis-intermediating humans from their jobs and roles in society. This responsible approach also better aligns capital deployment with returns, which will create a positive flywheel rather than a post-bubble crash.
Final thoughts: Future of Venture Capital
The most successful funds have kept their sizes fairly small — Benchmark, USV, FF, to name a few. Even today’s largest ones delivered their best returns with much older vintages with smaller vehicles. While the recent flight to massive scale (A16Z and General Catalyst took half of venture dollars this year) indicates some LPs are unconcerned, I worry this could lead to venture asset class underperformance, which would in turn lead to a lack of capital to fuel the innovation that drives the growth of our country and increasingly the entire world. However, I’m optimistic with new funds such as Chemistry, Defy, 20VC, Refactor, Voyager, Theory Ventures, and others that the actual craft of venture will continue and lead the way for successful venture investing.
Ultimately, what matters for society is that we channel the capital towards the right founders to build the companies that will change our future for the better. I believe a return to basics, with a reasonable number of rationally-sized venture firms run by investors with expertise and investment judgment is the path forward. Let’s hope near-term hubris does not prevent getting it done.
To discuss or learn more about Avila VC please reach out to us.
Follow us for sporadic postings on LinkedIn and X, and Patty Wexler on Medium
Love this. Totally agree that while venture capital faces significant changes from AI and market concentration, successful investing still requires discipline, patience, and thoughtful scaling. Great case for: return to fundamentals rather than chasing trends or excessive scale.