You asked about Big Data
133 people built something like it. 12 are gone.
That is 9% of the ones whose fate we could settle, against 23% across the whole register. 6 were acquired.
| Launched | What it was | Today |
|---|---|---|
| 2016-01-01 | AI + Robotics for the Built World · geckorobotics.com | still up |
| 2024-09-01 | YCArchil The high-performance file system that connects AI to data · archil.com | still up |
| 2023-06-01 | AI that audits AI — catching hallucinations before you ship them. · energent.ai | still up |
| 2023-06-01 | Metadata for GenAI workflows · deasylabs.com | acquired |
| 2025-09-01 | YCSciloop Expert STEM reasoning data for frontier AI labs. · sciloop.dev | still up |
| 2024-01-01 | Deploy trustworthy data agent that know your business · upsolve.ai | still up |
| 2025-06-01 | YCLiva AI Audio & Video Data · theliva.ai | still up |
| 2024-09-01 | Build full-stack applications for business operations with AI · vortexify.ai | still up |
| 2026-06-01 | YCAtlia The AI-native property management company for short-term rentals · atlia.com | still up |
| 2016-06-01 | YCScale AI Data-centric infrastructure to accelerate the development of AI · scale.com | still up |
| 2022-01-10 | databar.ai | still up |
| 2023-09-13 | docs.lantern.dev | still up |
| 2022-01-31 | ideas-generator.com | gone |
| 2021-01-01 | YCEncord The data layer for physical AI · encord.com | still up |
| 2022-05-16 | ortto.com | still up |
| 2022-05-07 | layer.ai | still up |
| 2023-08-11 | cambioml.com | still up |
| 2022-02-01 | ango.ai | still up |
| 2022-06-15 | deep-talk.ai | still up |
| 2026-01-01 | Physical AI data lab · humanarchive.ai | still up |
| 2020-01-01 | YCAirbyte Context layer for production-grade AI agents · airbyte.com | still up |
| 2023-12-06 | lume.ai | still up |
| 2021-06-01 | YCAnakin One API to get clean data from any website for your AI agents at scale · anakin.io | still up |
| 2020-01-01 | YCDatasaur Datasaur builds private AI/agents for enterprise/government · datasaur.ai | still up |
| 2015-06-01 | We use human data and AI to develop better drugs faster. · vergegenomics.com | still up |
| 2018-01-01 | PROVEN is using AI, data and personalization to disrupt CPG · provenskincare.com | still up |
| 2022-01-01 | Compliant-Ready AI for the Enterprise · dynamo.ai | still up |
| 2022-01-01 | YCLanceDB Open-source, serverless vectordb for production-scale generative AI · lancedb.com | still up |
| 2024-01-01 | Accelerate Medication Access with AI · runtrellis.com | still up |
| 2021-01-01 | YCauthzed Authorization Infrastructure securing AI at OpenAI, Zoom, & more · authzed.com | still up |
| 2023-01-01 | Automate data-driven slide decks and documents with AI · rollstack.com | still up |
| 2020-01-01 | The API to get recordings, transcripts, and metadata from meetings · recall.ai | still up |
| 2025-01-01 | YCMundo AI High Quality Multilingual Training Data for AI Models · mundoai.world | still up |
| 2020-06-01 | YCDatafold Automating manual and repetitive data engineering tasks with AI · datafold.com | still up |
| 2022-01-01 | YCSpade Spade is the data and AI platform for modern finance. · spade.com | still up |
| 2023-06-27 | capeprivacy.com | moved |
| 2021-06-01 | YCSecoda Secoda is the AI layer for your Analytics · secoda.co | still up |
| 2020-06-01 | Financial data layer for AI-powered finance and accounting · blueonion.ai | still up |
| 2013-01-01 | Data and AI platform for Ecommerce & Cross-Border Logistics · semantics3.com | acquired |
| 2023-01-01 | YCFlower Train AI on distributed data · flower.ai | still up |
| 2023-03-07 | yobulk.dev | gone |
| 2021-06-01 | AI Agents for Data Operations · oneschema.co | still up |
| 2022-01-01 | YCEventual Building the AI data engine for any modality and scale · daft.ai | still up |
| 2018-06-01 | Database for AI · activeloop.ai | still up |
| 2024-06-01 | YCEntangl Find & resolve issues in data center engineering & operations using AI · entangl.com | still up |
| 2022-11-29 | airy.co | still up |
| 2023-10-25 | blog.spice.ai | unverified |
| 2024-06-01 | YCSepal AI Data Development for Advanced AI · sepalai.com | acquired |
| 2023-08-09 | hawkflow.ai | unverified |
| 2024-01-01 | ProSights is a data / AI transformation partner to PE-backed portcos. · prosights.co | still up |
| 2022-01-01 | Rapid medical data annotation · redbrickai.com | still up |
| 2023-08-29 | layernext.ai | still up |
| 2023-04-04 | osschat.io | still up |
| 2024-06-01 | YCZeit AI The autonomous data engineer for the mid-market · zeit-ai.com | still up |
| 2023-10-17 | datamotto.com | moved |
| 2023-06-15 | deid.segmed.ai | still up |
| 2024-06-01 | YCDavid AI Data for audio AI · withdavid.ai | still up |
| 2022-06-01 | YCLamin Open data platform for traceable, multimodal AI · lamin.ai | still up |
| 2023-10-17 | app.refuel.ai | unverified |
| 2021-01-01 | Build better computer vision models by building better datasets · segments.ai | acquired |
How this compares
Y Combinator reports its own outcomes, including acquisitions. The other two are judged by asking the domain, so they are not measured the same way and are never averaged together.
Older launches have had longer to die, so this slope is expected. A year that broke it would mean something is wrong with the checking.
The register re-probes every domain it holds. Some that answered last month stop answering. We send that list — what went quiet, and what it was — once a month. Nothing else.
How this was made. Show HN entries are every post from 2021 to 2023 with at least five points that linked to its own domain; code hosts and free hosting are excluded, because a repository outlives the effort behind it. Product Hunt entries needed 100 votes — a far higher bar, so they are the promoted top slice rather than a like-for-like sample. Both are judged by asking the domain whether it still answers: gone means it no longer resolves, refuses the connection, serves a parking page, or says plainly that it shut down. An HTTP error is not counted as gone — a server answering 404 is one somebody still pays for, and an audit of 25 gone-calls found two such sites alive with only a moved homepage. Y Combinator entries use the status YC publishes itself, which is why only those can say acquired. Squatted domains — a lapsed name now serving somebody else's site — are not detected, so these figures run slightly low.