Key takeaways
- Amplitude counts events, not storage: admins and managers read Events This Month and Events Last Month for each project under Settings > Organization settings > Projects.
- Report GB if you send data to Amazon S3 or Snowflake; otherwise report total events as records and note that the number is events, adding your monthly counts across the months you have tracked.
- The full history goes to S3 or Snowflake through a manual backfill, or comes in zipped JSON from the Export API, limited to 365 days and 4 GB per request; Free keeps 1 year and Plus 2 years.
- You don't need to export anything to apply: have your size or event count, years of history and user count ready, and the data vendor anonymizes the records before any AI lab sees them.
01
What is Amplitude data, and why do AI labs and data vendors buy it?
The short answer
Amplitude data is the record of what people did inside a company's app or website, kept as a long list of events, each with a timestamp, a user and a set of details. AI labs (companies that build and train AI models) and data vendors (companies that anonymize business records and package them for those labs) buy a copy because the events show, in order, how real people move through a product, which is hard to rebuild from any other system.
What an Amplitude project records
Amplitude is a product analytics tool, software that shows a company how customers use its app or website. The company adds Amplitude's code to the product, and from then on every action a person takes, such as creating an account, opening a feature or paying, is sent to Amplitude as an event.
Each event carries the time it happened, the user or device that did it and properties, which are extra details such as the plan the person is on or the type of phone they use. Events are collected in a project, and one company can have several projects, for example one for its app and one for its website.
Why ordered events teach an AI model something
An AI model learns by studying a very large number of examples and finding the patterns in them, and those examples are called training data. Years of events show how real people move through onboarding (the first steps after signing up), which features they come back to and the point at which they give up. Public text on the internet does not contain that behavior, and no one can recreate it after the fact.
AI labs now want to train AI agents, software that carries out a job step by step, and an agent learns a job from records of that job being done. A data vendor can turn anonymized event histories into worked examples of how people complete tasks in software, or into a practice copy of a product in which an agent is scored on each attempt. The full mechanism is in AI training data: why AI companies want your company's data.
How Amplitude fits into a dataset
An Amplitude project is usually one of several sources in a company's dataset, next to email, documents and tickets. Companies are paid $100K to $2M or more per dataset, according to micro1, the data vendor we typically refer companies to. The price covers the whole dataset, not each source on its own, and what your data is worth explains what moves it.
02
Where does Amplitude show how much data you have?
The short answer
Amplitude measures your data in events, not in storage. Under Settings > Organization settings > Projects, an admin or manager can open each project and read two counts, Events This Month and Events Last Month. If your data already flows to Amazon S3 or Snowflake, report the size of those files or tables in GB. If it does not, report your total events as records and state that the figure is events.
How to read your event counts
The steps below are done by an Amplitude admin or manager. They follow Amplitude's guides to managing organizations and projects and to limits and quotas.
- Sign in to Amplitude as an admin or manager. Viewers and members can read project information apart from the API and secret keys. Editing project settings and setting up destinations is reserved for admins and managers.
- Open Settings > Organization settings > Projects and click into each project. Its General tab lists Events This Month and Events Last Month, and you write both down.
- Open Settings > Organization settings > Plans & Billing to confirm your plan, your contract dates and your monthly event limit.
- If a copy of your data already goes to Amazon S3 or Snowflake, check how large the Amplitude files in the bucket, or the Amplitude tables in Snowflake, add up to. That is your figure in GB.
Who can see what
| Role | What it can do |
|---|---|
| Viewer or member | See the information about each project, except the API and secret keys. |
| Manager or admin | Edit project settings and set up the destinations that send data to S3 or Snowflake. |
Why the page cannot give an all-time total
The project page shows only two months of counts, so the all-time figure is not on any screen. Without an export, estimate the total by adding the monthly counts for every month you have tracked. Report that number as records, and add a note that it counts events.
A project can define 2,000 event types, and events beyond that limit still count toward your volume. Amplitude still ingests them, although they can't be charted.
03
What should you know about your Amplitude data before you apply?
The short answer
Have three facts ready: the size of your Amplitude data in GB if you have an S3 or Snowflake export, or your total events if you do not, how many years of events you hold and how many users are in your Amplitude organization. They go into the data manifest, the written summary of which sources hold your records, how many years each covers and roughly how much. You don't need to run an export to find them.
The size in GB, or the events
Report GB when you have an export, by adding up the size of the Amplitude files in S3 or of the Amplitude tables in Snowflake. Without one, add your monthly event counts for all the months you have tracked, report the result as records and mention in your notes that it counts events.
The years of history
The years of history start at the first month with events in your oldest project. How far back Amplitude keeps your data depends on your plan, so the count may be shorter than the years you have used the tool; ask your admin which plan you are on.
The number of seats
The seat count here is the number of users in your Amplitude organization, meaning the people on your team who sign in to Amplitude.
What you don't need to do first
You don't need to export anything to apply, and you don't need to clean or scrub the records. The data vendor anonymizes them before anyone else sees them, and nothing moves until a lawyer-reviewed license is signed. You only need the size or event count, the years and the number of users for the first conversation. The application, the intro call and each later step are laid out in how to sell your company's data.
04
How do you export your full Amplitude history?
The short answer
An Amplitude admin or manager sets up a warehouse destination, either Amazon S3 or Snowflake, under Amplitude Data > Catalog > Destinations, and runs a manual backfill over the full date range to send the history. For a smaller pull, the Export API returns events in zipped files, with each request limited to 365 days.
Send the history to Amazon S3 or Snowflake
A destination is a place Amplitude sends a copy of its events to, and a backfill sends the events from before the destination was connected. The steps follow Amplitude's guides to the Snowflake destination and the Amazon S3 destination.
- In Amplitude Data, open Catalog and then the Destinations tab. Under Warehouse Destination, choose Amazon S3 or Snowflake and set it up with your own bucket or account details.
- For S3, open the connection's Backfills tab, choose your full date range and click Start Backfill. Manual S3 backfills have no limit on size or date range, Amplitude aims to export 5 million events per hour, and any overlapping data is de-duplicated.
- For Snowflake, open the integration, press Export Data and pick your date range. A Snowflake backfill can take from a single day to several weeks.
- When the files or tables are in place, add up their size for the GB figure.
Pull a smaller range with the Export API
The Export API is a web address that returns events directly, which suits a smaller pull. It needs your project's API key and secret key, which an admin or manager can see. The Export API reference describes it.
- Call the API with a start and an end hour in UTC, as in the example below. It selects events by the time Amplitude received them, so use the same hours that you want to cover:
curl -u API_KEY:SECRET_KEY -o export.zip 'https://amplitude.com/api/2/export?start=20260101T00&end=20260131T23'
Projects on the EU server useanalytics.eu.amplitude.comas the base address. - Split a long history into requests of 365 days or less, because that is the most one request covers.
- Total the size of the zipped files that come back.
How far back you can export
How far back you can export depends on your plan, because each plan includes a set period of data access. Amplitude's pricing page lists them.
| Plan | Data access |
|---|---|
| Free | 1 year |
| Plus | 2 years |
| Growth | Unlimited |
| Enterprise | Unlimited, with custom data retention available |
Formats and limits
The Export API returns zipped JSON files with one event on each line. A request covers at most 365 days, it is refused if the result would be over 4 GB, and it answers 404 when there is no data in the range. Data becomes available about 2 hours after Amplitude receives it, and the API allows 4 concurrent requests and 360 requests per hour.
Downloads of a chart as CSV stop at about 10,000 rows, so they are not a way to export the full history. Event types beyond the 2,000 per project are still ingested, and the Export API can pull them.
How long an export takes
A manual backfill to S3 runs at about 5 million events an hour, so a company with hundreds of millions of events should plan for it to run for some days, and a Snowflake backfill can take from a day to several weeks.
Sources: Amplitude documentation, Export API, Limits and quotas, Snowflake destination, Amazon S3 destination and the Amplitude pricing page. Steps checked against these pages in November 2026.
05
What stays private when Amplitude data is licensed?
The short answer
The company decides which projects and date ranges are shared, and the data vendor removes or replaces identifying details before an AI lab sees anything, including user IDs, device IDs, IP addresses and personal properties on users or events. The company keeps ownership of its data, and the license covers defined records for a defined use.
You choose what is in scope
You decide which projects are sent to a destination and which date range is backfilled, so a project that holds sensitive events can stay out. The data vendor also deletes records too sensitive to clean safely instead of anonymizing them.
Names are replaced before a lab sees them
Anonymization means replacing every detail that identifies a person or a company, such as a name or an email address, with a stand-in that is the same every time it appears. The records still link to each other and the dates still run in order, so a lab can learn from them, but they no longer say who anyone is. The diagram shows one set of records before and after.
The method, including what is deleted outright and how you review samples, is in how company data is anonymized for AI training.
Ownership and payment
Licensing gives the buyer limited rights to a copy, while the company keeps ownership of the originals. The terms that matter, such as exclusivity, permitted use and deletion, are covered in data licensing. micro1 pays in full within 30 days of accepting the data, usually as a single payment.
Questions
Frequently asked questions
Amplitude shows events, not a size. What should I report?
If you send data to S3 or Snowflake, report the total size of what lands there in GB. If you don't, enter your total events as records and add a note that the number is events.
Does my plan limit how much history I can export?
That depends on your plan. Amplitude's pricing page gives Free 1 year of data access, Plus 2 years, and Growth and Enterprise unlimited access, and custom data retention is an Enterprise feature.
Are events beyond my event type limit lost?
No. Once a project passes 2,000 event types, the extra ones still count toward your volume and can't be charted, yet Amplitude ingests them and the Export API returns them.
Can I export everything with a CSV download from a chart?
No. A chart's CSV download stops at about 10,000 rows. For the full history, send the data to Amazon S3 or Snowflake with a backfill, or use the Export API.
Why does my Export API request fail?
The usual causes are a range longer than 365 days, a result over 4 GB, which Amplitude rejects, or a range with no data, which returns a 404 error. Split the range into smaller requests, and remember that data appears about 2 hours after Amplitude receives it.
Do we need to export Amplitude before we apply?
No. The application takes about two minutes and shares no data. You only need your size in GB or your event count, your years of history and your number of users for the intro call, and the project page gives you the event counts. The full export is needed later, when the records are transferred after you sign the license.
Can we leave out events or projects that hold sensitive information?
Yes. The company chooses which projects and date ranges are sent to a destination, and the data vendor removes or replaces personal properties before an AI lab sees anything.
What is Amplitude data worth to AI labs and data vendors?
It is priced as part of a whole dataset rather than on its own. Companies are paid $100K to $2M or more per dataset, according to micro1, and the price depends on the years of history and how completely the records show each job. What your data is worth explains the factors.
Bottom line
Know what is in Amplitude, then see what it is worth
You don't need to export anything to apply. Apply in two minutes, describe your records on the intro call, and micro1, the data vendor we typically refer companies to, prices the whole dataset and sends an offer 1–2 days later.
Not legal, tax or financial advice. Figures come from micro1's published pages and public reporting as of November 2026.