In 1849, thousands of men flooded California chasing gold. Most went broke. The ones who got rich weren’t the miners.
They were the people selling picks, shovels, and denim pants. Levi Strauss didn’t pan for gold. He looked at who was already there, what they desperately needed, and built something durable on top of someone else’s rush.
The Gold Rush didn’t create wealth by producing gold. It created wealth by producing infrastructure that outlasted every mine that ever closed.
Every startup wave works the same way.
We never skip Reading the Failures
When a startup wave crashes, the narrative hardens fast: bad idea, wrong timing, deluded founders, reckless capital. Everyone moves on.
That’s the wrong reading. And it’s an expensive one.
The edtech wave of 2020-2022 didn’t fail because online learning was a bad idea.
It failed because
- the growth was borrowed,
- pulled forward by a pandemic, not anchored in structural behavior change. But in failing at scale, it left behind something nobody is pricing correctly:
- a generation of PMs who know how to build for Tier 2 and Tier 3 learners,
- half-built B2B infrastructure,
- assessment engines,
- credentialing rails,
- LMS APIs, that never found their consumer moment but are genuinely useful to enterprises, and
- a cohort of educators who figured out how to teach on camera.
The fad died. The residue didn’t.
The hyperlocal delivery wave, burning capital at a rate that looks insane in hindsight.
What it brought is extraordinary:
- a trained dark store operators, a workforce that understands cold-chain logistics at the neighborhood level, and
- unit economics data on last-mile delivery that took tens of millions of dollars to produce.
The failure is in the business model. The underlying problem, and the infrastructure built chasing it, remained entirely intact.
The graveyard is a supply depot, if read correctly.
Why Nobody Looks There
Because it feels like contrarianism for its own sake. Because the headlines called it dead. Because the last fund that touched the space took a write-down.
But there’s a more structural reason: most people ask the wrong question. If you look at a failed wave and ask did this work?, the answer is obviously no. If you ask what did this leave behind that the next entrant inherits for free?, you’re asking a different question entirely .
The residue that matters comes in four forms.
Trained talent. Every failed wave produces operators who learned something expensive at someone else’s cost. The fintech wave trained an entire generation of compliance officers, payment engineers, and credit underwriters who now exist in the Indian market and weren’t asking for FAANG compensation because the wave they rode didn’t produce FAANG exits.
Half-solved infrastructure. Almost every failed startup solved part of a problem before running out of runway. The crypto wave produced wallet UX research and on-chain identity experiments that the next generation of financial infrastructure builders is inheriting without paying for the R&D.
Calibrated customers. A failed wave educates the market. Indian consumers today understand UPI, digital lending, and online pharmacy in ways they didn’t in 2017. The failed startups paid the customer education bill. The next founders in those spaces don’t have to.
Revealed unit economics. Failed companies produce the actual numbers: what CAC looked like at scale, where LTV assumptions broke, which segments retained. That data doesn’t disappear when the company does. It lives in the people who survived the crash.
The Filter: Timing Was Wrong, Problem Wasn’t
Not every graveyard is worth excavating. The ones that matter share one characteristic: the timing was wrong, but the underlying problem was real.
The question we ask at Auxano when looking at a failed category: did customers actually want the thing, or did they tolerate it during a moment with no alternatives?
That distinction: tolerated vs. wanted, is the first cut.
The second is simpler: is the residue still there? Talent disperses. Infrastructure decays. The window for capturing what a failed wave leaves behind is usually 18-36 months after the peak. Long enough for the noise to clear, short enough that the assets haven’t fully dissipated.
Auxano at the graveyard
We don’t avoid failed sectors. We read them differently.
When a space is live, the noise is high and the tailwind makes it easy to confuse motion with direction. When a space has crashed, the noise is gone. What’s left is signal. The founders still building in a failed category after the capital has fled are telling you something important, they’ve found what the last wave missed.
That’s not desperation. That’s often the clearest early signal we see.
The map said those spaces were dead. The terrain says otherwise. And at Auxano, we’ve learned to trust the terrain.
The graveyard isn’t a cautionary tale.
It’s often the thesis.
Author,
Kushagra Sharma
