February 6, 2023

The Anatomy of a VC Bubble

There are many lessons to be learned from this VC bubble as the process of inflating the bubble charts a predictable course that suggest how the market could play out for for the Generative AI market - the newest input to the VC  fund machine.

The Anatomy of a VC Bubble

The numbers reflecting VC investments in crypto startups reads like the epilogue to a post-apocalyptic novel. In 2022, VCs poured over $21B into crypto startups. At mid-year 2023, that number is hovering at a paltry $500M.

A deeper look at the crypto-craze reveals the inner workings of an economic bubble involving some of the biggest and most well-known VC firms in Silicon Valley. It's hard to decompose tech bubbles because they typically play out over 10+ years. In the research brief I detail the chronology of events that unfolded in the Crypto market over a much shorter time span, which illustrates a visible microcosm of the macro software startup market bubble.

There are many lessons to be learned from this VC bubble as the process of inflating the bubble charts a predictable course that suggest how the market could play out for for the Generative AI market - the newest input to the VC  fund machine.

FTX Bursts the Bubble

As the dust settled around the shocking collapse of FTX, the financial world got a first-hand look at raw venture capital carnage.

There they were—Insight, Sequoia, Lightspeed, Temasek, NEA, Tiger, BlackRock, and more. Billions of LP dollars evaporated into thin air as FTX imploded."We are in the business of taking risk," proclaimed Sequoia in a letter to its LPs. "Some investments will surprise to the upside, and some will surprise to the downside."In what reads like a ho-hum mea culpa, the letter writer shrugs, "We do not take this responsibility lightly. [We] do extensive research and thorough diligence on every investment we make."The Internet was set ablaze in the wake of the Sequoia proclamation, casting doubt about such diligence on FTX. It’s not useful to speculate about what Sequoia knew and didn’t know but the inescapable fact does remain that hundreds of millions of LP dollars were invested into what now appears to have been a grift of epic proportions.

It's not like the current state of affairs just materialized. The FTX scandal burst an economic bubble 20 years in the making, which can be traced through the evolution of the modern venture capital investment thesis. We need to go back in time to understand this more clearly.

The Unicorn Fallacy is Eating the World, Not Software

VC software and Internet investing emerged reborn and armed with the lessons of the dot-com collapse in the present Internet epoch - what my friend Sangeet Choudhary dubbed the "Platform Economy."I remember it like it was yesterday. "Software is eating the world," proclaimed Marc Andreessen in 2011. "That's the big opportunity. I know where I'm putting my money," he wrote. Fast forward over a decade, and Marc's venture capital firm Andreessen Horowitz (a16z) has become the most successful fund of all time with no close second, accumulating a staggering amount of assets under management in record time.

Perhaps the vision of a savant, or maybe just being at the right place at the right time, the genius of a16z's complete remake of the VC investment thesis is undeniable.But the evidence piling up on Sand Hill Road suggests the macroeconomic "bubbles" may be greatly exacerbated by the nature and interests of contemporary venture capital defined by a16z.

Modern venture capital firms make money by charging fees that scale based on the size of their fund, not only the size of the fund's return. Consequently, the venture capital investment thesis has metastasized into a system predicated on seeking outsized returns from portfolio investments, which, organically speaking, happens very infrequently. So infrequently, in fact, that there are not enough "unicorns" in the world to soak up the amount of available VC dry powder in the market.This is the fallacy of the unicorn VC investment strategy. The strategy imposes quantitative constraints on portfolio selection. As a result, our view is that VCs tend to invest too much, in too many opportunities, and too fast, with low probabilities of actual unicorns in the return distribution. As you might expect, the consequence is systemic fund underperformance - the Cambridge Associates US Venture Capital Index averaged just 5.06% per year between 2000 and 2022, compared to the S&P 500 at 5.91%.This cycle of fundraising and fee stacking has been going on for decades, and the significance of the market crashes we have experienced over that time may be correlated to the proliferation of the unicorn fallacy fund thesis.

The Crypto Bubble Decomposed

Just a few years ago, King Midas himself, Marc Andreessen, wrote the playbook on crypto investing much the same way he wrote the book on software investing over 12 years ago. "I compare it to the Internet," he told CNBC in 2014. "Money is a very big deal, and so if you can build a new way to deal with money, it's important and valuable."Crypto investing led by firms like a16z and others exposes a fascinating look into the anatomy of an economic bubble and the potential role venture capital firms play.For the uninitiated, it's essential to understand what these VCs were investing into.A crypto-asset (or "token") is a digital asset designed to work as a store of value, like a traditional "security" might. These assets utilize different cryptographic principles to secure transactions, govern the creation of additional tokens, and verify the underlying tokens' transfer.

Bitcoin was the world's first major example of a crypto-asset.

Crypto transactions occur on a distributed ledger technology ("DLT") platform such as a blockchain. While each DLT platform may have unique features, rules, or permissions, the typical design goal is to achieve some form of decentralized trust and governance.The inability of government regulators to keep up with the volume and velocity of crypto securities and the inherent complications has resulted in a window of opportunity for capitalists, opportunists, and criminals alike. This has been playing out since the origins of the "Initial Coin Offering" (" ICO") craze of 2016 when over the course of 2017, the value of crypto assets like Ethereum rose over 17,000% based mainly on speculative hype.It wasn't until a little-known or understood project called Dfinity showed up that mainstream institutional venture capital revealed the true appeal of crypto investing.

The chronology of the events in this story illustrates the leading role that VC firms play in the orchestration of bubble that in aggregate begin to make a macroeconomic impact.Conspicuously dubbed the "Internet Computer," Dfinity claimed to have invented a new Internet that solved all of the old Internet's problems while making it easier and more cost-effective for developers to build distributed applications (or "dApps"). Dfinity's founder, Domenic Williams, is no stranger to the bombastic hype that seems to trigger Unicorn thoughts among venture capital investors. He once claimed the platform would be "humanity's primary compute platform for building software."

Dfinity attracted the attention of leading institutional VCs a16z and Polychain Capital in 2017. Between 2017 and 2018, the two firms combined to invest $102M in Dfinity despite there being no customers of record or no publicly available platform for customers or enthusiasts to trial. This is bizarre behavior since anyone who has ever attempted to raise capital from VCs will tell you that it is almost impossible without customer validation.

It is worth noting that before taking the millions from VCs, Dfinity founder Domenic Williams made his thoughts about an ICO abundantly clear. "The lack of discrimination between good and bad projects means there is very poor price discovery," he opined. He feared that if Dfinity were to run an ICO, they would be steering directly into "the legal and regulatory hornet's nest that has been created." "We don't want to have our project … distracted by legal problems."

Williams stoically proclaimed, "We are wary of running an ICO and having our reputations tarnished by some of the bad ICOs that were run, or becoming embroiled with problems with regulators," even after completing the transaction with Polychain and a16z.

Whether Dfinity planned to operate a public ICO in a classic bait-and-switch scheme or if they were pushed to do it by venture capital investors remains unclear. Either way, this is exactly what Dfinity did, on May 10, 2021.

With overt public support from the VC firms, Dfinity embarked on what can only be described as an epic public roadshow one would expect to see from a legitimate IPO candidate for a full year before the launch. In the final push, things had become fervently evangelical in nature. Talking heads proclaimed the Dfinity network represented "the dawn of the new open and free Internet."Dfinity's ICO became one of the most successful public offerings ever when the token price crested at $730.

According to fillings, initial institutional VC investments amounted to approximately $0.03 per token, with subsequent transactions priced at $0.62 and $4.16, respectively. The data suggests Polychain and a16z may have experienced mind-blowing 2,436,566%, 117,803%, and 17,472% returns on each pre-ICO investment.The price of the Dfinity token (" ICP") began to drop almost immediately after the ICO launch. Within six weeks, the token had lost 95% - over $300B - of its value. In the hotly debated June 28, 2021 report published by the analyst firm Arkham Intelligence, industry experts concluded, "the Treasury and suspected insiders have continually sent millions of ICP worth billions of dollars to exchanges, totaling 75% of the total deposited, possibly driving ICP's price collapse."

The Dfinity ICO became the subject of multiple class action lawsuits as retail investors who lost billions of dollars after the price collapse sought relief. Top among the demands included full transparency regarding the selling activities of insiders.

The story did not end there.a16z made a $152M investment in October 2021 in another project called "Axiel Infinity," prompting John Reed Stark, an 18-year veteran of the SEC Enforcement Division, to tweet, "in my opinion, the Ponzi enterprise known as @axielinfinity is one of the most shocking grifts I've ever seen. An egregious and transparent exploitation of the downtrodden, brazenly and shamelessly orchestrated in plain view." I’m in no position to comment on such allegations, but it is hard to ignore Stark's assessment as the investment world came to grips with the extent of the FTX collapse.

A16Z has also taken direct aim at influencing policy in Washington, DC, in an effort to preserve the crypto bubble. According to a Forbes analysis of financing disclosures, "the firm's leaders donated almost $2.2 million to financing the campaigns of more than a dozen lawmakers and a political action committee (PAC) that supports them." One such lucky politician was Ritchie Torres, who is vying for a seat in Congress. According to sources, a16z's top GPs threw Torres a swanky fundraiser in New York, after which Torres appeared to publicly voice support for a16z crypto portfolio boondoggle created by notorious WeWork founder, Adam Neumann, called "Flowcarbon." Vox journalist, Neel Dhanesha, described Flowcarbon as "a scam within a scam."

The Winners and Losers in a VC Bubble

The obvious winners in any VC bubble are venture capital firms. How much profit are we talking about in these cycles of VC bubbles? It was recently revealed that a16z is raking in over $500M in management fees per year. In the now infamous 2012 venture capital expose published by the Kaufman Foundation, research revealed that “VC funds receive nearly two thirds of their revenues from fixed fees rather than from performance-based carry.”

"Under the existing 2 and 20 structure, many institutional investors pay GPs well to build funds, not build companies," the authors concluded. If you follow the money, this model seems to incentivize the creation of bigger funds. Applying Kauffman's analysis to what we've observed with the crypto bubble makes the rest of the story entirely predictable.

Billions of dollars poured into VCs looking to replicate the winning formula in the wake of a16z's massive investment success in the Dfinity crypto scheme. If this sounds familiar, it should. The same pattern took place as VCs chased a16z's Internet software company investment returns a decade ago. A16z, not to be outdone, announced a fresh $2.2B fund dedicated to crypto investing while retail investors were losing billions in the ICP price collapse. This also included a brazen announcement of a new all-star cast of obvious insiders recruited to help lead the fund, including a former advisor to Hillary Clinton and Joe Biden, the former Under Secretary of the Treasury, and the former Director of the SEC's Corporate Finance division.The losers in this game are the limited partners (" LPs")."65% of financings fail to return 1x capital," says venture wonk Seth Levine. "And perhaps more interestingly, only 4% produce a return of 10x or more, and only 10% produce a return of 5x or more." Levine further illustrates that real data do not support the classic heuristic of the VC J-curve return projection.

Buyer Beware: AI as the Next Bubble?

Decomposing the crypto bubble within the broader arc of VC software investing reveals a fascinating pattern for predicting future economic bubbles, which can serve as a potential model for LP risk management."

We should drive AI into our economy and society as fast and hard as we possibly can, in order to maximize its gains for economic productivity and human potential," Marc Andreessen recently published in his latest investment manifesto. "Today, growing legions of engineers … are working to make AI a reality… My firm and I are thrilled to back as many of them as we can, and we will stand alongside them and their work 100%."

Groundhog Day, anyone?

AI startups have raised $15.5 billion this year, according to PitchBook data, which has already surpassed last year's total. In addition, a close examination of median post-money valuations reveals a 109.8% increase from last year.

This "book value" feeds the VC machine and further exacerbates the Unicorn Fallacy. There is no reason to see AI as anything but the next great VC bubble on a long arc of software investing.

In a recent interview with PitchBook's Leah Hodson, Kindred Capital's founding partner Leila Zegna poignantly confessed that "when the returns you thought you had vanish and you're questioning whether you're any good at [VC], putting your fist on the table to advocate for something that others don't see is a tough ask."

Indeed, the sound you hear is the thunder of a herd.