What happens to the tech startups that never go public?
... and what lessons can we apply to the AI market today?
Over the last year, there has been a feeding frenzy amongst AI startups, as dozens of different investment firms have clamored to deploy capital in the space. In the “post ChatGPT era”, every subsector and incumbent technology provider felt disruptable, and every business being founded was the next big thing.
Only time will tell whether 2023 will represent a golden age for founding AI companies, or a graveyard. We won’t know the answer to that for years, but what we can see already is that the frenzy has slowed. Many VCs have now made a few bets in the space (whether thoughtfully, or just in an effort to stay relevant), and are starting to pull back to wait and see how those pay off.
Regardless of how strong the vintage is, the reality is that there will be some winners, and many losers. But what happens to the companies that fall somewhere in the middle?
Past performance may be indicative of future results
In the past, technology companies had been funded on the basis of fundamentals, with a path to enterprise value growth via steady growth and annual improvements in profitability.
Over the last fifteen years or so, however, low interest rates led to significant expansion in the valuations of technology companies.
In general, tech companies grow much faster than companies in other sectors. Due to their higher growth rates, most of their enterprise value is tied up in future earnings (or “terminal value” in a DCF), instead of present earnings. This means that high growth, unprofitable companies are far more sensitive to fluctuations in interest rates than companies in other sectors. (If you’re interested, we discussed this concept in depth a few weeks ago).
With an extended low interest rate environment that began in the aftermath of the global financial crisis, technologies companies became far more valuable than they had been in the past (as the market valued “future earnings” more favourably). While this made the very best businesses more valuable, it also made the tier two and tier three businesses appear to be great as well.
As interest rates have normalized over the past two years, investors are starting to realize that some of these tier two and tier three assets may have been decent businesses, but they probably weren’t actually venture backable. This has resulted in a glut of growth stage technology companies that are marked at unsustainable valuations at best, and are revealing themselves to be unsustainable businesses at worst.
The below slide from our friends at Battery Ventures illustrates this problem visually: there are 1,000+ software unicorns globally today, but less than 100 have been able to go public over the last decade:
What happens to these companies?
A small handful of these companies will indeed go public. Given the current anti-trust environment (which has kneecapped Big Tech’s buying spree), I suspect even fewer still will exit in a large strategic acquisition. Some of these companies are proving to have broken business models, and are on a fast track to zero.
But the reality is, more than half of these companies have a path to becoming good businesses, but will have to endure the pain of a valuation reset to get there. How does that work? There are many possibilities, but lets discuss a few of the most likely:
Down Round Investment: The Company is able to raise another large round of funding, but at a lower valuation that is massively dilutive to the Founders and existing investors.
This is easier said than done, as growth investors will still look to underwrite a 5-10x return (and a potential public markets exit). Even if management is open to the dilution of a down round, Companies likely have to be posting 50%+ topline growth today for this to be an option
Sometimes rounds will get announced at a flat price / headline valuation to save face publicly, but contain significant structure (e.g., PIKing interest). The takeaway is largely the same either way
Subscale Strategic Exit: With most Big Tech buyers off the board, there are a handful of well positioned public tech companies who are likely to be making acquisitions in the next few years, but at much lower valuations than startups are hoping for
In this case, the revenue growth is less relevant than the quality of the technology, product, and team. If strategic acquirors believe they can resell a company’s products through their own existing distribution channels, they might be willing to transact
It’s worth noting that the premium strategics are willing to pay is probably not much higher than where they themselves trade today. As most tech companies are trading at 6-7x ARR right now, this likely means that these acquisitions will occur at prices below the valuations of prior rounds
Private Equity Exit: In what I believe will be the most common outcome, many of these tech companies will get snapped up by private equity firms who believe they can turn business around when they get under the hood
These investments will almost certainly be predicated on meaningful cost-out plans (finance jargon for firing people), in which a new owner cuts all the fat out of a business to make it profitable or near profitable, often at the cost of reducing topline growth
This has been the bread and butter of a few firms such as Vista Equity Partners and Thoma Bravo over the last decade, but I would expect to see a feeding frenzy in the coming months as many buzzy startups are rightsized
Clearly, none of these are favourable outcomes for Founders or existing investors. Almost all private companies will be valued in-line with public markets multiples, which will be a rude awakening for companies that raised at 100x+ ARR in the last round, even if they’ve grown well.
What can AI founders and investors learn from this?
Despite the SaaSacre in Tech generally, we’re clearly seeing shades of 2021 irrational exuberance in the AI market right now. I am far from the first investor to propose this, but massive investments continue to flow into these companies regardless.
How can the above learnings shape the way we think about the AI market today? Below are a few of my ideas:
Raise less money (at lower valuations). It’s sometimes hard to turn down the shiny object / turn down the money when it’s offered, but this Silicon Valley clip tells the story better than I ever could. Lower valuations mean you can set achievable growth targets, and reduce the likelihood that your last round valuation overhang is the albatross that kills a good business (yes, this argument is incredibly self-serving, but it doesn’t make it wrong).
Stretch out runway. Most companies try to budget for two years of runway. When money comes easily, we often see founders looking for ways to spend more money in order to “go faster”, but without any conviction that the extra dollar spent actually helps. If you do happen to raise a massive round at a large valuation that will be difficult to grow into, there’s nothing that says you can’t stretch out your runway to four or five years to give you the time and space to grow into a valuation that makes sense. This bears repeating: you don’t have to spend it!
Be meticulous about tracking spend and ROI. Almost nobody tracks ROI in the world of AI right now. Just like in the old school era of SaaS, tracking what you dollars are getting you and treating all spend as opportunity cost is worth doing sooner rather than later, to identify money pits long before they become a problem
Negotiate on employees salaries. Many hot AI startups today are just matching the cash comp given out by Big Tech, instead of convincing employees to take less cash and more equity (in the last three months alone, I’ve met multiple Seed stage companies who are planning to pay some of their next hires seven figure salaries). Even when there’s a talent war going on, keeping salaries under control is key to managing burn. Employees should be joining startups because of the equity upside, not to match the cash comp at a cushy Big Tech job.