Speed-to-revenue
A new breed of startups rises, and other AI-related things
Some AI-related things
insights from Stripe: speed-to-revenue
a new breed of startups arises
techcos, but also a place of business
to the consultants go the ai spoils
ai productivity gains
👉👉👉Reminder to sign up for the Weekly Recap only, if daily emails is too much. Find me on twitter, for more fun. A plug for our sponsor, AlphaSense
AlphaSense is actually an amazing one-stop-shop for:
expert calls (especially since it acquired Tegus),
company-filings, and
analyst research;
plus, it’s got a pretty neat generative AI-search tool that I’ve been playing with a lot (which makes searching across the corpus of information much easier than before).
It’s a genuinely fun and very useful platform, and as part of the sponsorship, Random Walk readers have ACCESS TO A FREE TRIAL, which you’d be a fool not to try (just by clicking through the link below):
FREE TRIAL! FREE TRIAL! FREE TRIAL!Give it a shot. It’s a great platform.
Speed-to-revenue
It’s Friday Fun-Day, so instead of victory laps and pre-recessionary indicators, we’re going to do some AI tailwinds.
First, from the always fun, Stripe annual letter is more striking evidence of “speed-to-revenue” for AI companies:
Median time to $5M arr for the Top 100 AI companies is 2 years, or 33% less than the top 100 SaaS Cos.
AI companies are achieving product-market-fit far more quickly than ever before (at least amongst the winners).
Why that is, one can only speculate. But presumably it has something to do with the novelty of the technology (and the new value that it unlocks), but also the relatively quick development speed of AI-assisted software development.
An new and different VC model to emerge
That the barriers to entry for tech businesses (and software development) are now lower than ever (and getting lower) is a really good thing.
It’s also supportive of a new model of techco investing that provides an alternative path forward for VCs, still very much mired in the duldrums (even if VCs are not the most innovative, in the main, when it comes to their business).
Smaller, less-capital consumptive companies that become “real businesses” without years of net-burn are possible.
Less capital, means less-dilution, which means end-markets (i.e. TAM) do not have to be quite so large, to generate outsized returns.
Plus, “exit” (over an 8-12 year time horizon) is not the only path to generating realized returns to investors: real revenue means old-fashioned stuff, like dividends and buybacks.
Or, alternatively, continue to compete with a16z by offering an undifferentiated product to a scarce-few “venture scale” opportunities, followed by 10-year illiquid holds periods, during which you will attempt to raise funds II and III and maybe IV based on paper-markups. LPs totally love that.1
To wit, Stripe highlights the growth of “vertical saas,” especially for SMBs, a strategy that has historically sounded very good, but has struggled to work.
Look at all the small businesses using some software that uses Stripe! (Don’t look at the decline in “Wedding Services,” though, because that’s depressing.)
But, back to the story of why speed-to-revenue (and the steady growth of small-businesses using software) is a tailwind for a different kind of VC product.
Vertical SaaS fails because small businesses are (a) small; and (b) fragmented. It turns out it’s really hard for the techco to acquire enough customers in a cost-effective way to achieve the sort of scale that mints actual multi-billion dollar companies—yes, there are a lot of small gyms in the aggregate (for example), but good luck getting all of them to buy what you’re selling.
So, you end up with niche software businesses that may well generate ~$50M in revenue, but that’s not nearly enough to achieve the “venture scale” outcomes VC need to drive returns.
That’s no good.
Real businesses, instead of cash-burning unicorns
The math changes, however, if you can achieve $50M in revenue with just a single, smaller round of financing (and a lean team).
It changes even more, if you were never trying to “hyper-growth” your company in the first place, and/or target some unachievably high TAM, which makes burning-cash a self-fulfilling prophesy.
I mean, one of the key learnings (and truly surprising outcomes) of the past few years is how nimbly cash-burning companies became cash-generating ones, once their incentives changed. The lesson, in most cases, is that they shouldn’t have been burning quite so much in the first place.2 The trade-off is that sustainable growth (as opposed to hyper-growth) still leaves the wildly overpriced equity underwater, but just because the previous equity paid way too much, doesn’t mean these are bad or uninvestable businesses.
To the contrary, compound earnings 10% yoy, and that dog will definitely hunt.
“Non-venture scale” is sacrilege in VC-land (for good reasons, mostly, but also some bad ones), but times are changing, and I think this will look obvious with the benefit of hindsight.3
There’s a whole new class of investable techcos out there, and AI is going to help make it happen.
To the consultants go the spoils
Random Walk (and now others) is worried about job growth.
You know the drill: outside of healthcare (and the retailers who feed and clothe them), there hasn’t been much.
So where does demand for new worker bees come from? Does it come from transformational technology, like AI?
Well, hopefully, but not yet. There are some green shoots, though.
GenAI job postings are growing like a rocketship:
GenAI job postings have almost 30X-ed their share of total postings since 2023 . . . from one rounding error, to a slightly larger rounding error.
OK, so 0.3% of total isn’t much, but it is more than before. So, that’s good.
Who are the big winners? Is it engineers? Data scientists?
Nope. It’s Management Consultants:
Management consultants have captured greater than 10% of all GenAI-related job postings, which is more than double the next closest role (ML engineer).
What this means, I couldn’t really tell you, but we already knew that consultants were making a ton of money working with clients on implementing AI, or at least, talking about it a lot.
I suppose that continues to be the case, and the premium to workable AI skills, will presumably continue to rise. On the other hand, consultants are the classic “we don’t know what to do, but we must do something, until this blows over” solution, so let’s hope it’s not that.
Productivity gains
And finally, from the St. Louis Fed, workers are reporting pretty big productivity gains from AI.
20% of respondents reported 4 or more hours of worked saved per week:
Put another way, 66% reported at least 2 hours of work-saved per week.
It’s just a survey, but hey, that’s not bad.
Unsurprisingly, the more you work with AI, the more time you save
“Computer and Math” reported the highest usage, and the greatest share of time savings.
That AI is a massive force-multiplier for coding seems indisputable, at this point. That alone is a pretty great leap forward (when you consider the leap forward that software generally, has been).
Other links
Imports from China are probably higher than we think. The discrepancy between how we count imports and how China measures exports has grown substantially since the ‘trade war’ began. The “missing” imports can’t be explained by reroutes to Vietnam and/or Mexico. The likely source of the gap are the duty-free de minimis imports via Temu, et al.
FDIC quarterly banking profile. Mostly banks look fine. NIM rebounded with the yield dis-inversion. Overall earnings continue to be mostly flat. Credit quality is pretty solid.
S&P base-case expects speculative grade default rate to fall to 3.5%. Lower yields are a big driver of the optimism, but also underlying strength. Definitely some red flags though.
Brex turnaround driven by ‘less is more’ approach. Lean, mean, corporate card generating machine.
Japan’s birthrates fall further . . . but South Korea’s increased for the first time in a decade! Idk. Bad news and good news.
Patrick McKenzie talks Data Center power economics with Azeem Azhar. Is it a bubble like before? It’s got some things in common, but also some things that are not. Great stuff.
Previously, on Random Walk
Random Walk is an idea company dedicated to the discovery of idea alpha. Find differentiated data, perspectives and people, and keep your information mix lively. A foolish consistency is the hobgoblin of small minds. Fight the Great Idea Stagnation. Join Random Walk. Follow me on twitter. Follow me on substack:
I considered doing a round-up of all the lamentations “gosh, it’s still bad,” and “what if the IPO market doesn’t open?” but it’s not nice, and I didn’t have the heart.
I’ve been writing about this since 2023, and every subsequent check-in has been some version of “it’s tough out there, as I predicted” “gosh, it’s still tough out there for all the same reasons” “can confirm, it’s tough, just like before” and “oh goodness that j-curve,” and all that. I do attempt “help is on the way” or “silver linings” or “looking up,” when I get the chance, but there haven’t been too many.
The reality is that the industry has contracted to a size more appropriate to the risk profile, and that’s going to leave plenty of managers and companies out in the cold (and already has). The ‘21-’22 vintages are busted, and unless you’re very early (where the possibilities are endless) or very late (where the path is clear), you’re not getting funding. The mushy-middle of alphabet soup rounds (which were really pass-the-baton cash-burning rounds) is basically gone.
With some rare exceptions, the best outcome for the recent vintages is getting some portion of the original investment back, but even good businesses leave most of the wildly overpriced equity under water (and the common isn’t getting squat). It’s demoralizing for everyone, and no returns (now or later) mean LPs are tapped out, which means fewer new funds.
But, it’s also a dream-selling industry, which means plenty of folks refuse to face the music. Lina Kahn was very bad, but she wasn’t the reason your portcos couldn’t exit. And rates aren’t going anywhere close to zero again, under any conditions that are good.
To be clear, a correction is good, even if it’s painful. Likewise, there are most definitely paths forward, but you have to adapt to survive. Personally, I think that’s exciting—the VC momentum trade was probably the least interesting VC has ever been, and new n’ different is (in all events) my jam.
Plus, the only way to compete with the name-brand multistrats (who are the only funds big enough to absorb the massive checks institutionals have to write) is to not compete: be very differentiated, or go where they’re not.
Sounds fun.
“Hey, maybe we should be building real businesses instead of setting money on fire for eight years” has most definitely entered the zeitgeist.
Whether its GPs focusing on more liquid structures for LPs, and/or less capital intensity for founders, or just companies celebrating their own “seed strapping” people are starting to get it.
The good ones are the basic math of pareto outcomes and fund sizes: if you’re only going to hit it big on ~1 investment, thEn that outcome has to be so massive that it “returns the fund.” If your fund is e.g. $100M, then you need to generate $300M to 3x returns. The only way to generate $300M from an initial investment of $2-5M (because again, 90% of VC returns come from 10% of the companies), is if that ~$12M company becomes a $1B+ company (i.e. 100X, net of some dilution). If that’s the case, then it doesn’t make sense to invest in a company that has no chance of generating that kind of outcome (so the story goes).
The bad reasons are that if you’re in the AUM businesses, then deploying capital to raise more capital is a pretty good model, so long as the music keeps playing. Fred Wilson of USV is famous for keeping funds relatively small, but others have not resisted the temptation to raise ever more (and clip that 2% management fee).
Forgive me if I've asked this before, but how many of the private companies that are waiting (perhaps in vain) to IPO are staying alive by borrowing from private credit investors? I smell a calamity approaching in private credit, although it may be years away. They are getting huge inflows and are desperate to generate the marginal asset, which is never good for credit. The big problem is that if (when) the cycle turns and they start getting redemptions, they can't nurse projects along (create money) as banks can. If they don't have plenty of cash they will have to sell the loan into a bad market. I know that there are an infinity of different structures for these investors, but still I think it's a problem waiting to happen.