Bugs In The System
On capital allocation, hard problems & incentives
Hey there,
I have been writing about the state of venture capital in Web3 since 2019. In all my years of writing, the data has driven the story. When you don’t understand the meta-game of what is going on, the numbers define how you see the world. The current cycle gives me time to think and reflect on what is happening. This piece summarises some of my observations.
When you study funding data, there are two patterns you would see
Capital allocators are reducing the amount of money that goes into the industry. (Contraction)
There is an aggressive piling of funds into the sector. (Expansion)
This expansion and contraction of money flowing into any industry is a feature, not a bug. People outside the sector often misconstrue it as the collapse of investor interest. But there’s usually more than what meets the eye.
Last week, it was revealed that partners at Sequoia, focused on crypto, are exiting the firm. The announcement came alongside their crypto fund slashing 65% in size. Similarly, Polychain’s latest fund ($350 million) is a fraction of the billions they used to deploy. The data below from Crunchbase is a good outlook on how the frequency of deals and the amount of money have reduced over the past few years. Long story short - the data shows that in terms of money allocated, we are in the worst quarter since Q4 2020.
But is that the case? It is unfair to make that claim without considering two broader points.
Firstly, the stress on the US banking system in Q1 of this year forced multiple ventures to see their lines of credit vanish. In turn, they had to raise money urgently from their existing investors to avoid bankruptcy. The trickle-down effect of that is diminishing venture frequency at the early stages.
AI, VR and AR as themes are far more capital intensive in terms of money allocated to research and hardware. For instance, Meta has spent at least $100 billion in cash on their VR initiatives. Similarly, some 40% of all patents filed by Apple have been in relation to their Vision Pro device. Nascent categories take tremendous amounts of time and capital to become “retail-ready”.
But inspite these broad trends that have driven massive sums of money into AI-specific deals, if you overlay the amount and frequency of deals that have occurred in venture as an asset class, you will see an overlap with what is happening in crypto. The chart below from EY uses data from Crunchbase (yet again). So it becomes easier to see the pattern.
Now I don’t mean to force fit a narrative here, but here is one way to look at it through the lens of yields. Asset classes like venture and digital assets (what we know as crypto) generally tend to produce beta. If the S&P500 is expected to generate 10% in return, a top-performing VC fund may have a return of ~19%.
An asset like Bitcoin may produce anywhere between 40-50% annualised. Surely, these assets have more returns, but it also involves more risk. The challenge emerges when yield rates rise. The capital flowing towards risk-on investments tend to reduce because you could create high returns with a fraction of the risk.
Tomasz Tunguz studied the correlation between yield rates and venture capital investment frequency in 2021. Here’s the crux of the story from his article.
“Here’s the bottom line: as interest rates increase, we should expect venture capital investment to regress. When the cost of capital is low and yields on cash are small, investors seek greater risk to attain a return. As the risk-free rate on Treasuries increases, market forces should engender enough friction to pull venture investment from the stratosphere into the troposphere or below.”
It may sound too theoretical, but the piece had a chart proving his point.
The data above is slightly dirty because
Venture as an asset class has grown over the last two decades. The rates could have remained high, and the money going into VC could have stayed just as high.
It fails into account the evolution of emerging sectors (like AI, blockchains, and greentech) that require more capital than dot-com ventures of the early 2000s.
Keeping that aside, it becomes evident that the low-interest rate environment of 2020 and 2021 created an interim phase where capital allocation to venture boomed. If you take that peak of $200 billion going into the asset class and take any period after, you will see a decline. Because pandemic markets were a unique condition that simultaneously created a perfect trifecta for ventures to absorb money.
Here’s how it played out
More people were using digital products due to the lockdown.
Ventures were showing hyper-growth due to point 1 and thereby commanded higher valuations.
At higher valuations, the tendency to raise more increased to compete with peers in the market and become a dominant player.
The headlines we keep seeing on repeat are designed to induce emotion and are generally not insightful. A different way to interpret this as the culmination of two factors.
An excess of liquidity entering the market thanks to stimulus from the government
The limited number of startups could meaningfully absorb large amounts of capital investments.
The reason we see market contractions often has more to do with industry maturity than investor conviction. During downturns, industries have a period of consolidation. Said consolidation can be of both technology and users. Without a down-trend, there would not be a period of developers & early adopters tinkering with a new category. That early phase of next to no users is where products evolve for a bull-cycle.
This is why both DeFi and NFTs as a sector took off after the bear markets in 2018. Few things convey this message and OpenSea’s userbase chart from 2019.
OpenSea was valued at a peak of $13 billion. It is up to debate whether the valuation is justified, but if it weren’t for the bear markets in 2019, there wouldn’t be time to sufficiently iterate on the product for the retail users that came in large troves in 2022. MakerDAO, Binance & Uniswap could make similar claims, given all of them grew during a bear cycle.
An even better parallel to draw here is Google and Amazon. The search engine was initially a PhD dissertation by Sergey Brin and Larry Page. Facing the challenges of a lack of advertisement models for search engines, the two set out to sell Google to Excite in 1999. But no parties were willing to buy it, even for $750k. 24 years later, Alphabet is the behemoth we know it to be. Similarly, Amazon dropped 90% since listing during the dot-com bubble. Both ventures have become crucial infrastructure for the Internet today.
Sidenote: Here’s a brilliant Ted Talk by Jeff Bezos from 16 years ago comparing the internet to electricity.
All of this paints a rosy picture and does not address the crux of the problem blockchain native ventures have today: A lack of users that pay meaningful fees being retained long enough.
Incentives Rule Everything
Most blockchain ventures we see have a pathway to liquidity in the form of tokens. As Sid often likes to say - ventures are hard to build once a token becomes the product. The DNA required to keep a listed token afloat at high prices & the one needed to retain users that generate fees for a platform are quite different.
What we often refer to as “venture funding” in crypto is, sadly, liquid market investing (into tokens) with lock-ups. Since the liquidity event for most blockchain ventures is tokens, there is often little focus on retaining users for the longer run or building a business the traditional way.
Our focus instead is often on
Weaponising airdrops for short-term accrual of users
Driving narratives with the users and
Issuing a token that ties back to that narrative.
When a narrative becomes the product, the business becomes a collection of stories. And stories, on their own, cannot accrue value on a long enough timeline. Eventually, it creates a prisoner’s dilemma, where everyone’s incentivised to have their tokens vested and sold at the highest price possible.
As I write this, Curve’s founder’s $168 million worth of tokens is facing a liquidation cascade as he chose to take a loan against it instead of selling it. He had taken the loan to acquire two mansions. This is not too different from a Web2 founder selling secondaries. It is even more transparent, but the time to liquidity is the difference.
A founder can expect meaningful liquidity within a few years in blockchain-native ventures. This means the incentives to stick around building decline rapidly for investors and founders. Last I checked, you cannot build a generational firm within two years.
(Ironically, Hopin, the example I linked above was also launched two years back).
When everybody is incentivised to prop up the quickest narrative, the focus on building long-term businesses shifts. It is the reason why we woke up to Twitter influencers talking about $bald (some meme token that went to zero) over the weekend instead of discussing what blockchains can be used for. We are sitting on world-changing infrastructure, battling our instincts to trade meme-tokens and gamble with on-chain Ponzi schemes. In effect, the industry is the biggest marshmallow test there could be.
Hard Problems
I believe this dichotomy makes investing in Web3 native firms an exciting opportunity. Most of the founders we see can be filtered out once you assess
Technical depth
Ability to sell
Possibility of sticking to the venture for a long enough time frame
It is rare to see the three coming together. Why does any of this matter? Because there are opportunities at the periphery that remain untapped. We made and abundance of noise around DAOs, yet two years later, there are less than 70k active voters (over the last six months) on the three largest DAOs. Smaller ones, have less than 100.
It is hard not to observe how primitives like DAOs and smart-contract-based equity allocations could make a meaningful dent in the lives of thousands of startup employees worldwide. Why isn’t there a Deel or Carta that runs on-chain? The discourse is less focused on what the world needs and more obsessed with what seems intellectually intriguing.
While intellectual pontification can drive Twitter clicks, it rarely accrues large user bases the way a thoughtfully designed application can. In other words, we trade real users and meaningful revenue for traders' attention. The business model is less predicated on what a venture can do today and more so on what it can enable a decade later. (FWIW, all early-stage investing is on the basis of what can occur in the future. The difference is, with crypto, you can have the incentives (or liquidity) here and now).
A natural extension of this is what we have been noticing with Real World Assets (RWA) as a theme within crypto. There is some ~$3 billion in RWA-linked assets on-chain—less than 0.1-0.05% of the actual market size. If you treat blockchains as financial infrastructure enabling real-life use cases instead of gambling on-chain, you could build meaningful companies that make a dent in the universe. But the incentives are not stacked up to do so.
One way this translates to capital allocators is with how capital raises for protocols far supersedes money that goes into consumer applications. We often see teams struggle to raise money even while building products at par with the Venmos of the world as capital allocators are far incentivised to deploy to protocols that may have tokens down the line. Often, investors have a simple line of thinking.
A protocol with a token could go up 100x on listing ($20 million to $2 billion)
An application without a token has a meager chance of surviving, let alone being a unicorn.
This problem usually translates to a lack of capital flowing into applications. It could mean users don’t have good experiences when coming on-chain. That, in turn, translates to protocols lying unused. In essence, we create a self-destructive flywheel that primarily runs on narratives. Naturally, there is no instant fix for this. One of the ways this has fixed itself in the past is with a new crop of investors emerging.
If you read Sebastian Mallaby’s Power Laws - it becomes painfully evident that a new crop of investors usually leads the pack when an emergent sector formalises. Sequoia’s Don Valentine started the firm when the traditional investors of the 1970s failed to fill the need for capital felt by tech founders of the time.
Tiger Global and Softbank’s Tech funds emerged during the mid-2000s to fill the gap left by PE funds unwilling to invest in growth-stage ventures. One place I witnessed this was with Tiger’s Lee Fixel leading a $1 billion round for the first time in an Indian venture named Flipkart. It changed the landscape in India as we knew it and set the stage for over 108 unicorns to come from the market in the following decade.
Closer to crypto, Paradigm, Pantera, and Polychain have filled that gap. But because crypto has shorter life cycles (as an asset class), the pace at which a fund evolves to be the size of a PE fund is much faster. All three funds I mentioned handle billions of dollars, and the desire to build alongside early-stage organisations that can absorb smaller sums of money while facing years of uncertainty may no longer exist. This leaves an opportunity gap for new investment funds to capitalise on.
One way this “gap” emerges is with large venture funds that traditionally operate outside crypto. Analysts at the firms often pattern-match through the same lens they use in other sectors when analysing deal flow that goes to partners or investment committees. By the time these funds build a network (for deal flow sourcing) and acquire the skill sets to deploy actively, the whole market cycle will have played out. This is partly why their entry often marks the top and departure marks the bottom.
The winners in crypto in the next cycle will not differentiate themselves by the size of their funds alone. Startups need three things to go from 0 to 1—capital, distribution and secrets that stem from insights or research. We have an abundance of capital in the market today. The opportunity is with distribution and insight. A new class of investors have been addressing those problems specifically.
Operators, researchers and media houses are well-optimised to solve for insight and distribution today. Robot Ventures (by Tarun Chitra and Robert Leshner) & Bankless’ $30 million fund are two instances of this. Paradigm’s focus on research is another example of a fund building a differentiator through sectorial expertise. Work by Dan Robinson combined with the firm’s capital helped fuel what we eventually came to know as DeFi
To make a long story short - there isn’t a lack of capital in digital assets. What we have instead is a mix of factors at play.
There is a broader contraction in capital going into venture capital.
The market needs time before it can mature enough to take more capital.
Incentives are skewed for allocators to sit on the sidelines when it comes to ventures that may take forever to create a return.
And there may be opportunity costs in not deploying directly into liquid assets.
All of this ignores the internal friction at large organisations that were historically deploying. How do you meaningfully create conviction at an investment committee once the venture you deployed so heavily into is exposed for fraud? These are things that take time to resolve. And unless those wounds heal, it is unlikely that we see the same players (of the last cycle) lead massive rounds in crypto.
For the moment, here’s what I do know. Technical innovations are occurring in the industry that are worth backing. We are at a very 0 to 1 phase when onboarding users interested in tools that do not involve speculation.
And for what it’s worth, to borrow how Steve Jobs would have put it: we are nowhere near having made a dent in the universe. These factors leave me bullish on what can still be built, scaled and exited. If you are building in the industry, leave your decks here—we (really) like the builders.
I’ll see you guys later in the week with a long form on narrative cycles & the psychology behind capital flows.
Joel