The list
Launch500 2026
Once a year we publish the startups building software worth paying for. Not the best funded, not the loudest, not the ones who asked. A place cannot be bought, applied for or voted into: it is earned on the Crash Test, reviews, customer feedback, and our engineers’ view of how useful the product really is.
Published February 2027
as an editorial covering every startup selected for 2026
- 0
- In contention
- 120
- Listed
- –
- Top score
passed a Crash Test in 2026
companies we cover
nothing tested yet
How a company gets considered
You cannot buy a place. You can earn one.
- 1
Book a Crash Test
A named tester runs your category's published scenarios on a dated version, and the report is published whatever it says. You pay for the test, never the result.
- 2
Pass it
A pass puts you in the 2026 field. A conditional or failed verdict does not: fix what broke, tell us, and take the free re-test.
- 3
The editorial decides, in February 2027
Our engineers weigh four things: the Crash Test, reviews, customer feedback, and how useful the product is in practice. There is no quota, so some years will be a handful.
What the judgement weighs
Four inputs, and no formula
Ours
The Crash Test
Named tester, dated version, hours logged, and everything that broke written down. The entry barrier, and one input of four.
Ours
How useful it is
Our engineers' view of whether the product does a real job better than the alternatives, under real use, not whether the feature list is long.
Theirs
Reviews
Verified reviews from people who paid the bill, weighted by how much proof sits behind each one.
Theirs
Customer feedback
What the product's own customers tell us about support, reliability and value, beyond what is posted here.
Nobody yet
The 2026 field
No product has qualified for 2026 yet
No product has passed a Crash Test this year, so the field is empty and this page says so. It fills as reports are published.
An empty list is worth more than a full one nobody can check. The first name on it will have earned the position in the only way the list allows.
Before you ask
What the list is not
It is not for sale.
A Crash Test buys the test, never the result, and a listing buys a page, never a place. Sponsored spots are labelled and never appear on the list.
It is not a popularity vote.
Recommendations and votes live on product pages. They never touch a Crash Test score or a place on the list.
It is not a ranking of companies.
It judges products, not the companies behind them. A good company can ship a product that fails.
It is not a fixed number.
Five hundred is the name, not a target. A quota forces a list to be padded in a thin year, and the moment a place has to be filled it stops being earned.