We Pinged 3,098 Startups That Raised Seed Money in 2024. One in Six Didn't Answer.

What we did

Our database tracks every venture round we can verify, which means we also hold something morbidly interesting: the website of every startup at the moment it raised. So we pulled the full 2024 seed cohort—3,098 companies that closed a Seed or Pre-Seed round between January and December 2024 and listed a website—and requested every single domain on July 3, 2026.

That puts each company 18 to 30 months past its raise. Long enough for the money to run out if nothing worked. Short enough that almost none of them have raised again without us noticing.

Each domain landed in one of four buckets:

• Alive (2,549 companies, 82.3%) — DNS resolves, the site loads, a real page comes back.
• Dead DNS (206, 6.6%) — the domain no longer resolves at all. Nobody is paying for it or pointing it anywhere.
• Parked (6, 0.2%) — the domain resolves to a “this domain is for sale” page.
• Unreachable or erroring (337, 10.9%) — DNS exists but the server times out, refuses connections, or returns an error for its own homepage.

We treat dead DNS and parked as the strongest failure signal—a company that stops renewing its domain has usually stopped existing. The unreachable bucket is murkier (more on that in the caveats), but a startup whose homepage has been down long enough for us to catch it is not a startup having a great quarter.

Crypto is the graveyard. Biotech barely dies.

The headline rate hides enormous variation between sectors. Here is every industry with at least 100 companies in the cohort, sorted by the share that is gone or distressed:

Three things jump out of this table.

Crypto seed startups from 2024 are disappearing at triple the biotech rate. 37.6% of cryptocurrency companies and 24.6% of blockchain companies failed the pulse check. These raised during a recovering crypto market, not the 2021 froth—and they are still dying faster than everything else we track.

The AI boom has a quiet mortality problem. AI is the largest category in the cohort by far—718 of the 3,098 companies, 23% of everything that raised a seed round in 2024. One in five of them (21.4%) is already gone or distressed, and their 9.3% dead-domain rate is well above the 6.8% cohort average. Record AI funding means record AI funerals; the two headlines just run two years apart.

Regulated and physical industries hold on. Biotechnology (8.6%), medical (11.7%), health care (11.7%), and manufacturing (12.8%) all die at half the crypto rate or less. Our guess: these companies raise against milestones that take years, not months, and their seed rounds are sized accordingly. A biotech that raised in 2024 is probably still running its original experiments.

Money buys time, and pre-seed has less of both

Survival scales with the size of the check. Companies that raised under $1M show an 8.1% dead-domain rate; companies that raised $10M or more at seed—there were 112 of them in 2024—die at 2.7%. The medians tell the same story quietly: the companies still alive raised a median of $1.30M, the dead ones $1.12M.

Stage matters the same way. Seed-stage companies show a 6.3% dead rate (16.8% including distressed), while pre-seed companies sit at 8.1% (20.2% including distressed). None of this is surprising—runway is literally purchased in dollars—but we have rarely seen it measured on a single clean cohort rather than assembled from survey answers.

Caveats, honestly stated

A website check is a pulse check, not a death certificate. Specific ways this measurement can be wrong, in both directions:

• Some “dead” companies are alive. Startups rebrand, get acquired, or move domains, and the old URL in a funding announcement goes stale. We spot-checked our dead list and found live, well-known companies whose recorded website was simply outdated. The true mortality rate is somewhat lower than the raw 17.7%.
• Some “alive” companies are dead. A domain with two years of prepaid registration and a static landing page passes our check while the team quietly went back to their old jobs. Zombie startups are invisible to this method.
• The unreachable bucket is noisy. Aggressive bot protection can block an automated check even with a browser user agent. Some fraction of the 337 unreachable/erroring sites are healthy companies with paranoid firewalls.
• Our cohort skews toward announced rounds. Startups that raise quietly and die quietly never enter the dataset.

We are comfortable with the direction of every finding above—the industry gaps and check-size gaps are far larger than any plausible error—but treat the point estimates as having a percentage point or two of slop.

Why this matters if you sell to startups

We built this study on the same data our customers use for prospecting, and the lesson cuts both ways. A company that raised seed money 6 months ago is close to the best lead in B2B: funded, hiring, and building its stack. The same company 24 months later has a meaningful chance of being a ghost—which is why lead lists built on stale funding data quietly rot at roughly the rates in the table above.

If you want the live version of this cohort—the startups that raised in the last 90 days, with verified founder contacts—that is what VCBacked does all day. And if you want to re-run this study yourself next year, the methodology section above is the entire recipe: pull a cohort, wait, ping.