This page explains where every number on Federal Exposure comes from, how we calculate the figures we derive, and what our data cannot tell you. We currently cover 40 companies.

What we measure

We track two independent measures of each company's dependence on federal money.

  1. What the company reports. The share of revenue that a company says came from the U.S. government, taken from its annual report on Form 10-K.
  2. What the contract record shows. Federal contract obligations to the company and its subsidiaries, taken from USAspending.gov and matched to the company by our pipeline.

The first is the number investors usually cite. The second is built from federal records, not from the company's own reporting. Comparing the two checks each one and adds detail, such as which agencies the money comes from.

Definitions

Reported share. The percentage of total revenue that the company attributes to the U.S. government in its 10-K, in the company's own words and on its own definition.

U.S. government revenue. The dollar amount behind the reported share. Where the company discloses the dollars, we use them. Where it gives only a percentage, we calculate dollars as reported share × total revenue and label the result "calculated". Total revenue comes from the company's XBRL financial data.

Adjusted share (ex-FMS). The reported share with foreign military sales (FMS) removed. Under FMS, the Department of Defense buys defense goods and services from U.S. companies on behalf of foreign governments. Some companies count these sales as U.S. government revenue and some do not. We compute an adjusted share only where a company discloses its FMS dollars. Otherwise the adjusted share equals the reported share and is marked "not adjusted".

Contract obligations. The total amount federal agencies legally committed to pay a company on prime contracts in one federal fiscal year, net of reductions. The federal fiscal year runs from October 1 to September 30.

Obligation ratio. Contract obligations ÷ reported U.S. government revenue. A ratio near 1 means the contract record and the company's disclosure roughly agree. It is a check on both measures, not a performance score.

Share change. This year's reported share minus last year's, in percentage points ("pts").

When a company does not disclose a U.S. government share, we show "not disclosed" and explain why. We never estimate a share and present it as reported.

Sources

Every figure on the site carries a source link and an "as of" date. SEC data on this build is as of 2026-10-16. USAspending data is as of 2026-10-16.

How we read the 10-K

Companies state their U.S. government share in prose or in a customer table, and the wording varies widely. A program, the candidate finder, searches each 10-K for sentences and table rows that mention the U.S. government, the federal government, the Department of Defense or the Department of War near a percentage or a dollar amount.

A person then reviews the candidates, picks the right one, and records the exact sentence, the filing link and the review date. The finder never publishes a share on its own. Only reviewed rows reach the site.

For each company we also record:

How we match contracts to companies

USAspending lists recipients by legal entity, and a large company may receive awards under dozens of subsidiary names. We assign each recipient to a company in this order:

  1. Manual overrides. A reviewed list of Unique Entity Identifiers (UEIs) mapped to companies. Overrides always win. We use them for renamed entities, false matches and joint ventures.
  2. Parent rollup. If USAspending lists a recipient's parent and that parent is already matched, the recipient is matched to the same company.
  3. Name aliases. Patterns for each company's known names, checked against the recipient name.
  4. Unmatched. Large recipients that match nothing go on a list we review at every refresh.

Rules that apply throughout:

Each company page lists every recipient name and UEI we matched, with its obligations, so readers can audit the match.

Why obligations and revenue differ

Contract obligations and reported revenue measure different things, so they rarely match exactly.

Our explainer on obligations, outlays and revenue covers these differences in more depth.

Known limitations

Recurring data posts

Some posts, such as weekly roundups of contract actions and watches for new filings, are generated from the data pipeline using fixed templates. Every number in them comes from the processed data, with its source. A person reviews each post before it is published.

Validation

Before any build, an automated check runs on the data. The build stops if any of these is true:

The check writes a report that a person reads at each refresh.

Change log

Corrections, data refreshes and method changes, newest first.