YC May Have Peaked: Data Suggests A Stumble in Acceleration Impact
A new study examining around 750,000 American startups across 329 accelerators finds that Y Combinator (YC) historically delivered extraordinary value, but its estimated value-add appears to have declined by 2022. The researchers make a deliberate distinction between startup quality and accelerator impact, a separation that strengthens the credibility of their conclusions. Broadly, the findings are sobering: roughly 60–80% of accelerators seem to underperform relative to simply building without an accelerator.
From my perspective, YC may have faced three compounding challenges.
-
Scaling constraints in early batches: Batch sizes grew significantly, but scarce resources—such as partner attention, investor exposure, and personalized introductions—do not scale at the same rate. The study itself notes that accelerator effectiveness can drift due to partner turnover, mentor networks, and program redesign.
-
A shift in founder experience: YC increasingly selected younger founders with less accumulated industry experience. In a landscape where rapid product iteration is common, domain judgment becomes a premium asset. Sectors like insurance, defense, healthcare, institutional finance, and manufacturing still demand years of context and hands-on experience.
-
The AI paradigm shift: Artificial intelligence represents a genuine technology inflection point. For a long period, YC built patterns by backing companies like Stripe, Airbnb, Dropbox, Coinbase, and a slew of SaaS ventures. Those patterns may transfer poorly to a world where a capable founder can produce a compelling AI product in weeks.
Paradoxically, YC’s strongest moat could be amplifying this decline.
The enduring YC advantage likely lies not in generic startup know-how (much of which is publicly available) but in its Bookface-like network: a dense web of exceptional founders, customers, employees, investors, and domain expertise. Networks compound in both directions: they can accelerate value creation, but they can also magnify decline.
If YC today selects fewer defining AI-driven startups, those companies may not reinforce tomorrow’s network. A weaker network, in turn, reduces value for future AI founders, making YC less attractive to exceptional founders who bring deep industry knowledge. This creates a potential negative feedback loop: weaker selection leads to fewer significant winners, a thinner network, and diminished accelerator value-add for future cycles.
It’s important to note that the paper does not prove that YC’s founder selection has deteriorated. What it does show is provocative: YC once appeared extraordinarily transformative, yet by 2022 its incremental contribution seemed comparatively modest.
YC remains a prestigious brand, but the glory may belong more to past achievements than to current, transformative impact.
The key question moving forward is whether YC continues to cultivate a network capable of fueling the next technology cycle, or if it is primarily monetizing the network built during the last one.
P.S. The paper is titled Beyond Demo Day: Sorting and Value Added in Startup Accelerators, and is freely accessible on the NBER website.
