Peter Pan and Neverland VC Returns

As we start 2023, the infamous phrase “software is eating the world” turns 12 and fast approaches its teenage years. After a glorious decade-plus run, I believe the typical angst, awkwardness and entitlement awaits many parts of the industry in its teenage phase. As the VC cloud ecosystem evolves from being a “do no wrong” innocent child into a mature adult that has to deal with the realities of the outside world, I believe the industry (and its returns) are likely to have its Peter Pan moment where it fights mean reversion while still trying to fly around in Neverland. The emerging consensus view that “now is the best time to invest in venture” is unlikely to materialize — although that doesn’t mean those at top need to worry much (at least for the next decade) or that some savvy firms can’t do more than deliver market beta.

To say the last decade was relatively straightforward in venture and the industry truly did FLY would be an understatement. Money was cheap, inflecting secular trends (cloud, mobile) were massively underappreciated and the VC industry had not been industrialized – most firms were just a handful of investors with minimal multi-stage aspirations, agency models, platform teams, etc. This field of opportunities for venture investors was wide open. As long as you were swinging at the ball (aka investing in cloud / software) earlier in the decade, you probably made contact and did alright given the tailwinds at your back and multiple expansion (until lately). Many smart folks did more than alright, built great companies / firms, got liquid one way or another and generated significant wealth for themselves and those around them. Many lucky folks also confused their fortune for success, but that’s another story (and an unfair characterization of the greatest and most dedicated to the craft). Nonetheless, this turn VC into the equivalent of Neverland — generate great returns and significant money with limited risk? Sign me up!

Obviously a lot has now changed and this narrative (and its underlying conditions) are aging out into adulthood. Much of the discussion in techland of late has focused the boom/bust narrative around how post-GFC easy money policy and its reversal (thanks, Powell!) are the chief culprits of recent and future pain. While partially true, I think that narrative only captures some of the nuance and understanding the path going forward will be more complicated than simply adding a 10Y treasury curve to revenue multiple charts and saying “whoops!” This success and VC industrialization over the last decade+ catalyzed a subtle, but meaningful, change that is likely to dictate the “teenage” years of the software VC industry and deflate returns without some course correction. In the course of this boom, I believe VC investing structurally shifted from incentivizing “non-consensus” thinking to becoming a “consensus-driven” industry. Investors could have their cake and eat it too so they put significant resources behind this ostensibly successful strategy, played the numbers game and tried to scale it to its limits. This showed up in many ways, but most notably is the unstated but observably true notion that investing appears to have become an exercise in optimizing dollars at work in the “best” companies versus optimizing returns.

Since pretty much the inception of VC investing, excess returns have been predicated on the idea that a small number of companies drive a large percentage of returns (aka the “power law”) due to their asymmetric upside. A corollary to the asymmetric upside is that above market returns / alpha is driven by making investments that are “non-consensus” where the investor ultimately ends up being correct. This article from Marc Andreeseen and Tren Griffin in 2014 highlights this notion that was originated by Howard Marks with one of my favorite charts:

The takeaway is that the best returns come from betting on companies / things that seem unlikely to work and then turn out to be right in doing so. This was more or less described cloud / software in the post GFC period (2010s). While it was on the radar, many underestimated how significant and central it would become for IT. AWS had <$1B in revenue in 2011, now it has $80B!  ICONIQ, Accel, a16z, Sequoia and others played this trade well, made early bets and rode the wave.

Fast forward to cloud software’s teenage years in the mid 2020s and most of the people, companies and investors around this world have shifted to the top left of the chart above. Even if we’re only part of the way there in terms of cloud adoption, the institutional VC industry got incredibly good at finding, funding and paying up for companies all the way to their earliest stages of development (just follow the github stars!). Entrepreneurs also got better at building companies given all the success and an increasingly large number of operators who have “been there and done that.” Along the way, they realized they could raise capital at increasingly higher valuations because of the rising VC competition and leverage they had as talented founders.

Now, we’re at a point where everyone effectively understands the trends in the space, has gotten much better at being “correct” in riding these trends and are fighting for their share of the pie instead of taking risks and trying to make non-consensus bets with asymmetric payouts. This has invited capital, scale and competition into the software investing arena in potentially irreversible ways and it's all chasing the same opportunities. It’s become a market share battle amongst increasingly large firms and this will drive out alpha. In essence, software VC has moved from a place where one could find abnormal returns to a place where it makes more sense to optimize fund size and try to index the market. Everything becomes consensus very quickly (see: AI) . For sectors like software, it may be time to remove “venture” from “VC” and just call it “capital”. 

As with many financial behaviors, the incentives are driving the outcome. I’ve illustrated via the sample economics of all this below comparing a $500M fund charging 2/20 that returns 3x to a $2B fund that returns 1.5x This is an oversimplification in many ways with a lot of assumptions in here, but the 3x fund returns 12% annually and $250M to its GP whereas the $2B fund returns 4% annually and $400M to its GP. Even though 4% is below the long-term equity market average of ~7%, it seems irrational as a GP to pursue the first option if you could potentially choose the second.

And that is exactly what many firms are doing. Fueled by the pandemic hysteria, top tier private funds built warchests on the back of a decade-long boom and elevating multiples. While the fundraising data below is incomplete and imperfect, you can see that many major funds raised 3x-6x as much LP capital in the last few years compared to the prior four and are clearly demonstrating a willingness and capability to choose or accept the “large fund” model for rational reasons. Using these funds, the firms staffed up, built competitive portfolio services, expanded teams and paid themselves handsomely. It's a good model :)

Further illustrating this industry growth point is the slide from ICONIQ’s fundraising deck that Eric Newcomer shared over the summer. ICONIQ is one of the best pound-for-pound SaaS investors over the last decade. Their returns below speak for themselves. They got in early, harvested great <$1.5B funds (3-10x is quite impressive), and are now running funds of $2.6B, $4.1B and $5.8B over the last few years. I’m curious how well this approach will scale but admittedly quite skeptical. We already saw what happened to Tiger and Softbank and I’m guessing the rest of the VC universe is about to run into a similar wall. By the way, I don’t mean to pick on ICONIQ (they’re insanely good at what they do), but unfortunately they're the only fund whose version of this slide leaked.

As the tide goes out in the current washout, investors will get more selective and focus only on the “best” and “hottest” companies. Investors will realize they’re sitting on record funds they need to deploy and yet only really want to be involved with a select handful of companies. The means the competition to invest in top decile (or quartile or top 5%) companies will only increase and more dollars will be squarely pointed at this opportunity and away from the excessive number of lower tier businesses funded these past few years. This will keep those prices artificially elevated and mute returns since these ARE the power law companies and everybody knows it. The $290B of dry venture powder is great for companies and terrible for investors. This mismatch - between fund sizes raised reflecting peak multiples and a smaller pool of smaller outcomes - will almost certainly squeeze out returns.

As a thought experiment around this, I looked at the early rounds (Seed, A) for Figma. Figma built an impressive software company over the last decade and probably had the most investors lined up out the door to put money in (e.g., it was a consensus bet). Its $20B+ strategic sale in a down market illustrates its positive attributes. Hypothetically, if Dylan (Figma’s founder) wanted to build another company (Figma 2.0), he could probably raise funding at 2x the price of his Figma original seed round ($11M) and A round ($34M) given his track record and prestige. If he had similar success with Figma and built it into another $400M revenue company, it might be able to fetch a 20x ARR multiple in a steady state environment (and I think that's being generous). This means the back of the envelope intrinsic value is $8B. Using some of the share price / return numbers from Rolfe Winkler, here:

Rolfe Winkler @RolfeWinkler

What investors first paid for @figma, which @Adobe buying for ~$40.20 per share: $0.088: @dannyrimer/@IndexVentures, @semil, Jacobsen/OATV A $0.199: @johnolilly/@GreylockVC B $0.332: @mamoonha/@kleinerperkins C $1.098: @andrew__reed/@sequoia D $4.619: Peter Levine/@a16z

This math would wipe out $1.4B of returned capital from the seed investors (was a 457x, now it's a 91x) and $2.3B from the Series A investors (was a 202x, now it's a 40x). While both cohorts of investors still saw strong returns (hard to argue with a 40x), those are still significant return dollars getting competed away in an era where funds are 5x+ as large as before.

Most private investors will refuse to acknowledge all this (at least for a while). I mean, why would they? It’s hard to learn new tricks and there’s a strong financial incentive to keep the dream alive, raise big funds, get paid pretty well and *maybe* deliver returns. Nobody wants to be an angsty teenager trying to figure out who to become as they grow up and need to change their ways (especially if nobody is pushing them to do so). It’s much easier to suspend one’s disbelief, desperately pray that AI is the next platform shift and live care-free in Neverland for the next decade while collecting significant management fees. It’s not like IT will stop modernizing so may as well play the averages.

Whether the industry knows it or not, it’s evolving into an indexing strategy unless it decides to tackle new frontiers and take on new risks. Some firms are thinking this way, focusing on new sectors (like climate, robots, healthcare) and doing a better job of managing fund size growth. But many aren’t and remain content with the indexing approach, which I believe is why we’re at a Peter Pan moment. Capital finds a home and giving VCs too much access to capital incentivizes poor allocation decisions and eventually lower returns. Realistically, the industry is prone to capital cycles like anyone else, so the question becomes a function of who can survive, evolve and deploy responsibly versus resting on their laurels until they realize the old playbook doesn’t work (or that there was a high amount of luck involved the first time around).

We’re starting to see evidence of firms trying to live in Neverland indefinitely in the form of VCs saying they will focus on public stocks (which appears to have backfired thus far), go earlier than their mandate (seed / A) and add new financial products (debt, structured equity, family office services) to their repertoire. We’re seeing way less funds say “hey, we think our fund size is actually too big, we need to rightsize this to deliver the returns that we promised you.” I don’t actually expect anyone to say that, but it will be interesting to see how this plays out in the next 5+ years. 

The (increasingly false) truism and guiding principle that “you simply had to invest in the best companies in software / internet to generate great, outsized returns” has passed. Returns won’t come as easily as they did these past 12+ years and, as it turns out, entry price DOES matter. Sincere kudos to those who were both smart and lucky to go long cloud in 2009/2010. Perhaps a new platform shift creates inflection point returns, but given what we saw with crypto and how quickly the hype cycle around AI reached a fever pitch, I think only a select few will truly prove their ability to deliver abnormal returns by being early / non-consensus in the right spots and companies. I think / hope Sam Lessin nails it in his note by highlighting the outcomes aren’t as massive as expected and the industry will revert to favoring only the most talented “pickers”. After all, how many $50B software companies can there be?  

Alpha in traditional VC is turning to beta. Barring a few trillion dollar outcomes in AI, Neverland may never be the same.