The case against accelerators
I have a weird relationship with accelerators, incubators, and really any form of pragmatic company building. On the one hand, I got my start at Entrepreneur First and owe an enormous debt to the EF team and the wonderful founders that I was lucky enough to work with. Had I not been where I was, when I was, I can safely say I wouldn’t be the investor I am today.
I also find the structural realities of running a program fascinating. I could spend hours debating the tradeoffs of target volumes and decision funnels. The perpetual economic tension between quality and quantity of a cohort. The viability Demo Days. S-tier shop talk.
But on the other hand, over the last 5+ years I have seen a steady decline in the importance of accelerators as sourcing channel for (most) VCs. This is just my experience, but I am quite confident that this view is widely shared. It’s not exactly a secret that investor sentiment toward startup-building programs has atrophied over time and here’s my take as to why. TL;dr: Accelerators were a product of the times—specifically, localized capital scarcity—and no longer make sense.
A sign of the times
Nearly all structured company-building programs are a temporal phenomenon. They are a catalyst for activation energy until they’re not; typically within an underdeveloped geography or technology/industry sub-domain (past: space or climate; present: defense). As startup markets mature, capital flows in, funding sources compete and upgrade their offerings and entrepreneurship becomes an increasingly viable career path for talent. With that maturation comes diffused startup guidance (lore?), more experienced or nuanced investment products, and higher (funding) clearing prices. These collectively make the relative offering of a fixed-terms accelerator uncompetitive. Value dissipates, although there is a time lag depending on brand strength.
At some point, the programmatic benefit collapses onto itself and inverts, creating adverse selection. This either takes form in reality through sub-par support/guidance, or in the perception that taking part in a program signals a lower-quality founder who undervalues their potential or can’t perform as needed. This is the reputation death spiral.
The obvious exception
Firstly, I will readily acknowledge that YC is the exception that proves the rule. YC invented the asset class, and accordingly, it enjoys the benefits of the first mover and mindshare leader. When YC speaks, all other programs listen. There are lots of mixed feelings about YC in the investor community, but the reality remains that you can’t be in this game and not check out the batches. Even if the signal-to-noise ratio is totally broken with 200+ companies per cycle and Demo Day being a futile event, the YC badge still carries weight.
Only the innovative survive
I argue that there are only four durable positions that an accelerator can pursue:
Capital moat – There are some niche domains where the inherent barriers to entry naturally demand scale and accordingly provide durable value. Innovating in biology, chemistry, and material science requires lab space, and lab space is expensive. There is a legitimate gap to fill at the day-one stage, and program-based investors can fill it. The same could be true for any equipment-dependent area like quantum or some industrial automation.
Example(s): IndieBio, The Engine
Talent moat – Co-founder discovery is a uniquely powerful offering in the long-term potential of a company. When done right, the 8–10% equity cut is well worth it. To be sure, my past work experience might bias me here, but speak to any entrepreneur who had to look for a co-founder and I bet they would agree.
Example(s): EF
Network moat – Geographic arbitrage is something that has a relatively enduring advantage. Most investors would argue that talent is relatively distributed, but opportunity is certainly not. Because of the importance of network-based sourcing, it is very difficult for any fund to materially scale a geography. Today, the only real play—at least for top talent—is to bring regional talent to the US.
Example(s): , HAX
Service moat - Premium positioning is always viable and accelerators are no different. Great talent (especially second time founders) don’t care about mentorship or pitch practice but they do care about great catering or turnkey service. This isn’t the most scalable strategy, but it’s sound.
Example(s): Hf0
To be clear, even these strategies are not immune to competition. As organic “founder meetups” begin to permeate a city, the relative advantage of a co-founder discovery program falls. YC also offers one for free. As Seed/Series A funds build inroads with local funds, the relative value of geo-curation falls. This is the nature of market competition: adapt or become irrelevant.
Don’t get taxed for what the internet gives for free
The idea that a program can more effectively / efficiently teach entrepreneurship is trash. There is an ungodly amount of educational content available for free forever—and it’s provided by better people too.
With even mild connections, an entrepreneur can find a better cash offer than ~$150K for ~8%. To be honest, if you can’t, you’ll never make it anyway. Even solo GP / micro funds will offer better economics, and you’ll usually get a much more engaged investor because it’s their career, not a job. If you don’t know many investors, change your LinkedIn employment status to “Stealth” or “Working on something new.” I’ll wait…
If you’re still considering a program, then at least do yourself this favor:
Get the latest cohort/batch’s funding stats. It’s really easy to have impressive numbers at a macro level (e.g., “our portfolio has raised a cumulative $2 billion”), but that doesn’t matter if the program is in decline.
Find out who specifically would be advising you, if that’s part of the deal. Prioritize founders who have built something or program people who’ve been doing the role for 5+ years. There is a lot of churn in accelerators, and you shouldn’t let them train on your time.
Take some independently sourced references. It’s easy to get some nice words from the star pupil of the last batch, but you really want to understand what it’s like for folks who struggled to get their narrative/customer/product right.
Accelerators are a necessary but temporal market player. Just as government subsidy should not exist in perpetuity, neither should industrialized company building. Capital should become abundant, and high-quality (pre)seed funds should proliferate to a local maxima. This has been the story of London, and it’s increasingly true of Paris, Stockholm, Berlin, etc. I would like to see programs like this in underdeveloped markets like Poland and Turkey. Ireland could do more to fulfill its potential too. Despite my best efforts, I have not found an investment in Lisbon. I think it’s too hard to build a big company with that much sunshine.
When I got started in 2016, EF was the only game in town to see high tech-risk opportunities. AI was synonymous with deep tech. Seed rounds were less than £1 million. It was a simpler time. If you’re reading this in a city where that doesn’t sound outrageous, then join the cohort and Godspeed. But if you live in any of the cities I’ve mentioned—or joining a given program is a somewhat normal thing to do—think twice.