When does public training data run out?

Short answer

Epoch AI, a research group, estimates that the stock of public human-generated text data could be fully used for AI training between roughly 2026 and 2032, pushing buyers toward private and licensed data.

When does public training data run out?: overview of The estimate, What happens after, What this means for partners
Covered on this page: The estimate · What happens after · What this means for partners

The estimate

Epoch AI, a research institute studying AI trends, estimates that the stock of public human-generated text could be effectively exhausted for training purposes between 2026 and 2032, depending on how fast models and training budgets grow (Epoch AI).

What happens after

When public data stops being enough, buyers turn to:

  • Licensed content — publishers, libraries, platforms.
  • Synthetic data — generated by models themselves, with limits.
  • Private operational data — the records of real work inside companies.

What this means for partners

The third category is where SourceX operates. Companies with 50+ full-time employees at peak usually hold years of private operational data. Introduce one to SourceX; if a deal closes and SourceX collects its fee, you can earn 25% of that fee, up to $100,000 per company.

Refer a company · How rewards work

  1. Step 1Share your linkSend your personal link to a company you know.
  2. Step 2Company appliesThe company applies itself at /apply.
  3. Step 3Buyer selects and paysThe buyer selects and pays for the data and SourceX receives its fee.
  4. Step 4You get your rewardYour share of SourceX fees becomes payable.

Common questions

When could public human-generated text data be exhausted for AI training?

Public human-generated text data could be exhausted for AI training between 2026 and 2032, according to Epoch AI. This timeline depends on how quickly models and training budgets expand over the coming years.

What data sources will AI labs use once public data runs out?

Once public data is insufficient, AI labs and data buyers will turn to licensed content from publishers and platforms. They will also rely on synthetic data generated by models, and private operational data from companies.

What type of companies have valuable private operational data for AI training?

Companies with 50 or more full-time employees at their peak often possess years of valuable private operational data. This data represents records of their real work and operations.

How can partners earn rewards by referring companies to SourceX?

Partners can earn 25% of SourceX's collected fee, up to $100,000 per company, by referring eligible companies. Rewards are paid if a deal closes and SourceX receives its fee, after the partner introduces the company.

Free resources

By SourceX Partnerships Team · Published 2026-10-09 · Updated 2026-10-09

Know a US company with valuable proprietary data?

Become a referral partner from anywhere we support, get your link and introduce an owner or authorized decision-maker.

Refer a company →

I own a business

Explore licensing your company's data to AI developers worldwide. Start a short assessment; no uploads needed.

Start an assessment