Zombiecalypse in Startupland
For investors and founders who are tuned into the current VC pulse, there’s no question that the industry is in a very weird state right now. On one hand, top funds are still competing for AI deals at a ferocity reminiscent of the 2021 peak, but on the other hand, there are 1400+ minted unicorns waiting in the wings for some sort of liquidity event. These very conflicting forces are causing quite a bit of consternation amongst the investment community.
In part 1 of this this post, I cover the following:
I start by taking a public equities lens and explore the current depressed multiples in the world of public SaaS companies
Next, I analyze why we’re seeing a lack of IPOs for growth stage companies, and why $100M+ in ARR (annual recurring revenue) isn’t enough for an IPO
I then explore why, despite the lack of exit liquidity, VCs are still aggressively deploying capital into AI startups, which could spur a repeat of the 2021 funding madness
What the lack of liquidity means for VCs, especially emerging managers as they raise their next funds
In part 2 of this post, I investigate specific strategies/portfolio construction required to succeed in the current investment environment.
Let’s dive in.
Just where are we in the SaaS cycle?
From a public equities perspective, there’s no question that the broader SaaS market is facing significant headwinds, with growth rates collapsing since the Covid highs. There are multiple reasons for this:
The overall SaaS market has matured as SaaS growth rates converge to overall enterprise software growth rates. Therefore, the growth of one SaaS company must naturally come at the expense of another, which I wrote about in a prior post, Investing in the Age of Generative AI
Despite the excitement (and corresponding spend) on generative AI, the SaaS market is going through a period of digestion and optimization, so sales cycles are naturally elongated and budgets more closely scrutinized
Given the broader macroeconomic environment (e.g. layoffs), there’s natural seat contraction. In the tech sector alone, there were ~109k tech layoffs year-to-date, and ~263k in 2023 according to Layoffs.fyi, which reduces the number of seats that buyers need
As we see from Clouded Judgement’s Q4 summary above, median YoY (year over year) growth rates have fallen ~50% from 2021. I used Meritech’s excellent comps table to slice things a bit further, bucketing SaaS companies by growth rates.
There are several things to highlight here:
Median EV / NTM (enterprise value / next twelve months revenue) multiples for sub 10%, 10-25%, and 25%+ growers are 3.6x, 6.3x, and 10x, respectively. This is significantly lower than the multiples assigned to hot private growth stage companies
There are only 9 companies growing at 25%+, and currently no SaaS company is growing 30%+
Public companies are operating at some scale. The median company across all three buckets generated $700-800M in LTM (last twelve months) revenue
This has a profound bearing on why we’re seeing a dearth of exits.
SaaS multiples compression has crushed growth stage investing
As we saw above, the market clearing price for public SaaS companies is severely compressed relative to that of private growth stage companies. This has resulted in a logjam of startups waiting for some sort of exit.
The issue with many growth stage companies is that their latest rounds were done at extremely high multiples (in some cases in excess of 100x ARR!), which makes it almost mechanically impossible to exit and yield good returns for investors, founders, and employees. Let’s use three hypothetical growth stage examples to illustrate this (a normal grower, fast grower, and hot 2021 deal done far ahead of traction). I make the following assumption:
All three companies raised $500M in aggregate funding, at a 1x liquidation preference (this means that investors get at least 1x of their money back before common shareholders, e.g. founders and employees)
I dispense with post-money valuations for simplicity
All three companies raised their last round of funding in 2021, meaning they’ve all experienced significant macroeconomic headwinds in the subsequent years
Companies exit at year 4 (this doesn’t include the years it took to get to their current revenue scale), so they would be valued via year 5 revenues multiplied by the median EV/NTM multiple according to the company’s growth rates (see the Meritech comps table above). As an example, our fast grower, growing at 26% in year 5 would be priced at 10x NTM revenues
Fast grower: This is every founder / VC’s dream. Our fast grower coming out of $100m in LTM revenue is still growing quickly with strong growth endurance (growth endurance is defined as the current growth rate, as a percentage prior year’s growth rate). The idea here is that it’s difficult to maintain high, compounding growth rates indefinitely. If our fast grower IPOs 4 years after hitting $100M in ARR, it would roughly be valued at ~$5.85B ($585M NTM revenue * 10x NTM multiple). In this case, everyone (VCs, founders, and employees) gets to eat. OneStream is one such company in this weight class.
Unfortunately, of the 1400+ unicorns, there just aren’t that many of these fast growing, premium assets that can function as a “release valve” of sorts for VCs. Databricks and Stripe are among the handful of assets that are growing at scale and capable of returning significant dollars back to LPs (limited partners). Even these companies are currently not risking the scrutiny of the public markets. Sequoia, in Stripe’s case, had to roll its Stripe shares to a newer fund to give its 2009-2011 LPs some liquidity.
Normal grower: The normal grower is a more common occurrence, meaning the company’s revenue growth “collapses” to 35% in year 1 given the lower 70% growth endurance. Many growth stage startups, like their public SaaS peers, have experienced this kind of growth slowdown in the past 2 years. The issue here is that these companies might be reluctant to ramp up their S&M (sales and marketing) spend given the broader market slowdown. As a result, these startups are stuck in a bit of a death zone where their anemic growth rates make them unattractive acquisition or IPO targets. In our specific example, our “normal grower” would be valued at ~$860M ($239M NTM * 3.6x EV/NTM multiple).
While this isn’t a terrible exit, the founders/employees will end up with less than envisioned. In an M&A scenario, late stage investors would likely take their 1x liquidation preference, while early-stage investors might elect to convert to common, so founders and employees might end up with significantly less than $360M ($860M minus the $500M capital raised).
The issue here is when a company of this weight class tries to go public. Public markets investors would have practically no interest in this class of company — why would they invest in a sub-scale company growing at less than 10% YoY (and in many cases not yet profitable)! There are dozens of public SaaS companies that investors can choose for better risk-adjusted returns! For late stage investors who put money at $1B+ valuation, they would actually lose money as their preferred shares get converted to common stock.
“Hot” 2021 deal: this is truly dangerous territory. During Covid, a surprisingly large number of deals were done at more than 100x+ ARR. As the macro shifted, our “hot” company experienced a significant slowdown in growth (60% growth endurance). An example here might be an online events startup that experienced high growth during Covid, but slowed down once the world re-opened.
If we price our hot startup relative to public market multiples, the company would be only worth ~$491M ($78M NTM revenue * 6.3x EV/NTM multiple). In fact, such a company, unless acquired by strategics, would be worth less than $491M because its revenue scale is an order of magnitude lower than public market comparables! For these previously “hot” companies, it ends up being practically impossible to get any sort of exit unless it’s via an acquihire, which would result in common shareholders getting wiped out. All this said, it is important to note that for truly “special” companies (e.g. Salesforce, Databricks), high entry prices do make sense, so an investor’s “picking” skills are as important as ever.
So, to summarize:
There aren’t a ton of fast growers, and the few fast growing companies may not want to deal with the fluctuations in the public markets
Normal growers, if their growth has slowed down, wouldn’t be able to attract buyers on their IPO roadshows, because those investors have better options (e.g. existing public companies with higher growth rates, greater revenue scale, and actual free cashflow)
Hot “2021” deals have it even worse, because there’s essentially no chance that they can go public, and any sort of acquisition would likely wipe out common shareholders
As a result, the public markets have been essentially devoid of software IPOs in the past few years.
The M&A market is equally challenged. Looking at the 2019-2021 timeframe, exit values in the US via M&A stood at $241B, compared to $88B in the 2022-2024 window (data drawn from Pitchbook’s 2024 Q2 Venture Monitor). Figma ($20B) and Wiz’s ($23B) acquisitions would have juiced M&A exit values, but unfortunately both deals fell through.
Now, I do want to stress that compressed multiples aside, that getting to $100M ARR is insanely hard (much harder than the job of investing!), and it’s even harder to grow at scale! The crazier thing is, even the cream of the crop startups that do go public end up being small fish in the deep waters of the public markets!
The cycle repeats itself in the age of generative AI
Less than 3 years after the last “hype cycle”, VCs have reverted to aggressively deploying capital, this time to AI startups, and we’re once again seeing sub-scale startups getting assigned unicorn valuations. This essentially becomes a repeat of the “hot” 2021 type rounds. The risk here is that we’re piling onto existing backlog of zombie unicorns with even more high priced AI rounds, many of which will likely struggle to sustain revenue growth at scale (a material portion of generative AI use cases are project based, customers experimenting with different solutions, or startups buying from other startups, so the quality of revenues for many of these startups is unclear). This is all happening when the market hasn’t really begun to digest our high priced, 2021 companies yet! Now imagine if we end up with 2000 unicorns that can’t properly exit!
A similar dynamic is happening with large tech companies and their approach to AI CapEx (see image below). The key difference here is that these companies have real profits to support their AI adventures! Like large tech companies, VCs might feel that they can’t afford not to play.
So what does that mean for VCs?
If the current logjam persists, there’s a real risk that LPs actually get tapped out without significant liquidity events. What will end up happening is that our startup Zombiecalypse would end up creating zombie VC funds as smaller / emerging funds struggle to raise subsequent funds (though does mean that funds that do survive will reap the rewards). We’re already seeing a bit of this dynamic playing out, where large investment platforms with a prior history of DPI (distributions to paid in capital) end up raising most of the capital from LPs. So far in 2024, 5 VC funds pulled in 45% of all LP capital!
Now, the open question is how much of the expected rate cuts in September are priced into public SaaS multiples. If for some reason rate cuts are more aggressive than expected, then we’ll likely see some sort of multiples expansion, and at least a portion of our 1400+ startups that are “stuck” would be able to exit under a more “relaxed” interest rate regime. Otherwise, there’s real risk that the VC/startup ecosystem experiences protracted pain. That being said, I don’t think VC firms are powerless by any means. There are real structural and strategic innovations that funds can adopt to generate alpha in this environment. I plan on exploring this topic in a subsequent post. Signing off till then!
Huge thanks to Will Lee, Maged Ahmed, Aaron Wong, Yash Tulsani and Andrew Tan for the feedback on this article. If you want to chat about all things ML/AI and investing, I’m around on LinkedIn and Twitter!
Really enjoyed the analysis--and definitely praying for another ZIRP. Quick question--where was the growth endurance benchmarked from? Or was it just an estimate? Taking a quick look at a company like OneStream, it seems like they had a ~87% revenue endurance.