Yariv Adan, Common Companion, ellipsis enterprise
There has by no means been a greater time to be an AI engineer. In case you mix technical chops with a way of product design and a eager eye for automation, you might need even constructed a extremely helpful app over a weekend hackathon. So, is it time to pitch VCs? Widespread knowledge says that if you could find a market hole, ship actual worth, and ship rapidly, you will have the recipe for a venture-backed startup. You might be probably watching numerous friends do precisely that. However earlier than you be part of the hunt for a billion-dollar unicorn, you must ask your self: would you be higher off herding donkeys?
and startups are altering. Not incrementally, however basically. Over the previous yr, we’ve met group after group doing every thing proper: transferring quick, constructing helpful merchandise, focusing on actual buyer ache, delivering actual worth. And but, we handed on lots of them. Not as a result of the groups had been weak, however as a result of the moats that will defend their worth have basically eroded.
Essentially the most primary rule of enterprise hasn’t modified: an organization wants differentiation and defensible moats to maintain high-margin success at scale. However what counts as a defensible moat has shifted dramatically, with the bar rising to a a lot larger stage. If your small business lacks a real moat, whether or not proprietary information or distinctive experience that may face up to a military of highly-skilled AI brokers, it is going to inevitably face disruption inside the commoditization kill zone.
Two years in the past, we coined the time period Commoditized Magic to explain the long run we noticed AI portray. Know-how and merchandise have gotten really magical, unlocking beforehand unattainable capabilities but they’re virtually fully commoditized by frontier fashions. We stay optimistic in regards to the “magic” half: it introduces an enormous financial alternative by unlocking worth that was beforehand inaccessible. However the commoditization danger is actual and disruptive, making whole areas uninvestable.
On this piece, we need to unpack that commoditization dynamic: why the unicorn is even more durable to hunt within the present panorama. However we additionally need to counsel {that a} new creature, or reasonably, a really acquainted one, is about to emerge: herds of donkeys.
Commoditization from Each Path
AI is consuming software program and providers, however on the similar time, the unit economics of making worth are drastically altering. The associated fee, experience, time, and general assets required to convey a product to market are spiraling down. That adjustments every thing, and commoditization is speeding in from all sides.
The person as builder. There’s a new class of apps changing beforehand bought software program: the ephemeral app. Whether or not it’s a easy immediate that creates an artifact, a Claude Code session, or some mixture of expertise, instruments, and plugins customers can now construct any app they will think about. Any skilled engineer is aware of that constructing even essentially the most complicated module for a single, one-time person is trivial; the standard complexity and experience kick in solely when making it modular, generic, scalable, and maintainable. A single user-builder is a formidable competitor to a complete SaaS firm in terms of constructing precisely the app she wants at a given second. This scales to groups as nicely, and thru organizational reminiscence, past that.
The explosion of rivals. As coding brokers enhance and attain the extent {of professional} human engineers at a lot decrease value and complexity of administration the entry barrier to turning into a SaaS firm drops dramatically, resulting in orders of magnitude extra rivals. The result’s crowding at each stage, and we already see it in our dealflow. Each use case now has quite a few startups attacking it, every ranging from a small beachhead the place they’ve some unfair benefit, hoping to broaden and win the market. However once they increase their heads, they see beachheads throughout them, with no clear differentiation. These firms could ship actual worth, some could even be worthwhile however they don’t make sense as venture-backed companies.
Enterprise and startups have at all times been a numbers recreation of hits and misses. However when the ratios shift by orders of magnitude, with much more firms, solo founders, and tiny groups all enabled by the identical instruments, the outdated guidelines break down. You find yourself with many extra misses than hits, to the purpose the place the VC mannequin itself stops working.
“It’s All About Distribution” Or Is It?
An argument we frequently hear is that in a world the place software program is a commodity, it’s all about distribution: transfer quick, seize these first clients, and also you win. Sadly, commoditization and AI are rewriting the principles of go-to-market and distribution as nicely.
First, there may be the crowding downside. In case you can transfer rapidly, quickly prototype an MVP, and signal a pilot, all in 4 weeks with two folks, so can your many rivals.
Second, not solely does AI unlock ephemeral, hyperpersonalized apps, however integrating conventional software program has additionally turn into a lot simpler, faster, and cheaper. Conventional SaaS merchandise arrive generic and require complicated, costly integration tasks, a significant supply of stickiness and first-mover benefit. Within the new world, the place these integrations might be automated or regenerated on the fly, these moats are quickly disappearing. As lock-in results weaken and the shopper not wants to fret as a lot about future assist and compatibility, they will concentrate on what they want now, and who does it greatest, particularly in extremely commoditized and aggressive markets.
In consequence, we count on software program procurement AI brokers to emerge that exchange outdated, human-led strategies. These brokers may bid and check in actual time for required capabilities, threatening to render model, distribution, and first-mover benefit largely irrelevant. The economics are clear: when switching prices strategy zero, loyalty follows.
Lastly, Massive Tech is transferring up the stack and throughout verticals. Think about how frontier mannequin suppliers and platform house owners, assume e mail, chat, and docs within the enterprise, or cellular, search, and social for shoppers, can now construct vertical use circumstances themselves, sooner and higher than ever. Google including AI capabilities straight into Workspace, Microsoft embedding Copilot throughout Workplace, Apple integrating intelligence into iOS. These giants are transferring into territory that after belonged to startups, leveraging distribution benefits that startups merely can not match. The power to develop at a lot larger velocity applies to Massive Tech as a lot because it does to a two-person startup, and Massive Tech begins with a billion customers.
That is the brand new actuality within the software program and providers market, as helpful intelligence turns into a commodity.
Donkeys, Not Unicorns
Is that this the top of entrepreneurship, is there no path ahead for robust small groups who can ship fast worth to underserved markets? Removed from it.
There’s clearly an enormous alternative for brand new unicorns, simply with the next bar. That’s the chance we’re targeted on as a VC. However we additionally consider that the superpowers and pace of AI have unlocked one other avenue for entrepreneurs, one which doesn’t require enterprise capital in any respect.
What if, as a substitute of chasing a single elusive unicorn, you used brokers and the low value of growth to automate and scale the creation of value-generating companies? Can a solo founder construct a herd of passive-income-generating donkeys at scale?
Take into consideration what that appears like in observe. You automate ideation and market analysis to generate, prioritize, and prune a pipeline of concepts. You automate person analysis and interviews, buyer outreach, speculation technology, prototyping, experimentation, and evaluation. You bootstrap these companies, run them in parallel, kill the losers, double down on the winners, and adapt as wanted.
Think about a founder operating fifteen micro-businesses concurrently, every serving a slim area of interest focusing on an underserved market phase they’ve entry to: one automating compliance studies for small European fintech companies, one other producing customized coaching supplies for logistics firms, a 3rd managing invoicing workflows for freelance consultants. Most likely even with geographical focus. None of those is a billion-dollar market. None of them will land on a TechCrunch headline. However every generates regular, sustainable income, and collectively they compound into one thing significant. The founder isn’t managing fifteen groups; AI brokers deal with the construct, the iteration, the shopper assist. The founder’s job is portfolio administration: which donkeys to feed, which to retire, which niches to enter subsequent.
That is the inverse of the enterprise mannequin. As an alternative of concentrating danger into one huge guess, you distribute it throughout many smaller ones. As an alternative of needing a 100x return on a single firm, you construct a portfolio the place the combination final result is what issues. The mathematics is completely different, the chance profile is completely different, and critically, it doesn’t require exterior capital, which implies the founder retains full possession and management.
We suggest this path to groups we meet who’re doing glorious work however working in areas the place the moat merely isn’t deep sufficient for a venture-scale final result. Usually very small and environment friendly, these groups are completely positioned to bootstrap reasonably than increase. The donkey path isn’t a comfort prize. For a lot of founders, it could be the smarter play.
This isn’t a venture-scale play, and that’s exactly the purpose. It’s a brand new avenue for entrepreneurs keen to commerce the dream of 1 huge final result for a portfolio of smaller, sustainable ones, and to make use of AI to make that portfolio manageable at a scale that was beforehand unattainable.
We consider there’s a actual alternative right here, and we’ve began exploring the instruments to make it work. Keep tuned.
