Coolcaesar · CC BY-SA 4.0
Now
~$5.4Tthe compute everything AI runs on
93 companies tracked and counting— browse the list →
Chai Discovery6/12 assessed0 sourced→
Basis6/12 assessed1 sourced→
Applied Compute6/12 assessed1 sourced→
Cognition8/12 assessed2 sourced→
Cyera11/12 assessed6 sourced→
Databricks12/12 assessed0 sourced→Browse all companies →
Which startup is actually worth joining?
A community-built guide to Tech/AI startups: structured company research, sourced key numbers, and first-hand signals from members working inside — so your next move is backed by evidence.
Latest researchlive from database
Every figure carries a provenance label and a source linkWhy startups at all?
Every giant you know was once a startup nobody was sure about
Someone joined Nvidia when it was pitching from a diner booth, and Stripe when it was seven lines of code. Joining the right startup early is one of the highest-leverage career moves in tech — early people compound with the company in ways no big-tech ladder can match.
Coolcaesar · CC BY-SA 4.0
Asoundd · CC BY-SA 4.0
Now
~$4.4Tthe front door of the internet
Andy Li · CC0
AmazonNow
~$3Tthe everything store, plus the cloud
Dllu · CC BY-SA 4.0Then · 2008
“Sold novelty cereal boxes to make rent”
Now
~$100BIPO'd bigger than the hotel chains combined
Coolcaesar · CC BY-SA 4.0Then · 2010
“Two brothers and seven lines of code”
Now
~$160Bthe internet's payments backbone
HaeB · CC BY-SA 4.0Then · 2015
“A non-profit research lab”
Now
~$850Bthe fastest-growing product in history
The honest part: for every company on this wall, hundreds you've never heard of quietly died. Survivorship bias is the whole difficulty — picking well is the game, and it deserves better evidence than vibes.
The next ones are being built right now — see who we're tracking →Photos via Wikimedia Commons — click a credit for the license.
93
companies researched in depth
12
checkable metrics per company
13
actively sponsoring H-1B
0
paid placements · ever
The rulebook is public, down to the last band
No black box, no editor's thumb on the scale. Michelin uses anonymous inspectors and confidential criteria; we do the opposite — a score can only be produced by these rules applied to sourced data, and there is no hand-editable score field anywhere in the system. Disagree with a weight? You can see exactly what it is.
The rubricv2.1
6 × 12Momentum16
Revenue growth · Capital validation
Efficiency20
ARR scale · Revenue per employee · Capital efficiency
Backing8
Investor quality · War chest
Market11
Market (editorial)
Product11
Product (editorial)
Upside16
Equity upside · Valuation / ARR · Liquidity track record
Weights sum to 100 · every metric has published scoring bandsRead the full rubric →See it applied →
Wherever you are on the way in
From idle curiosity to a signed offer, the research is the same — only the stakes change.
🔭
Just exploring
Curious what the startup world actually looks like beyond the headlines. Start with the list — every company researched in depth, no login, no paywall.
Browse the list →⚖️Weighing an offer
One offer or several, this is the diligence you'd want a friend on the inside to do: sourced numbers, scorecards, risks, and the questions to ask in the interview.
Open the rankings →🧩Already inside one
You hold pieces nobody else has. Correct a fact, add an insider signal, and make the picture whole for the next person deciding.
Join and contribute →