Corporate Events 2019–2026: M&A, Layoffs, IPOs

Key findings

1. 1,339,579 event reports for 210,414 companies, January 2019 to June 2026. The largest types are M&A (360,561), Company announcement (322,768), Contract (202,205), Asset acquisition (111,419).
2. Layoff reports peaked in 2023 with 5,166 reports, against 2,122 in 2019 and 4,065 in 2025.
3. IPO reports peaked in 2021 (8,136); 2025 had 7,978.
4. M&A is two-sided: 184,859 reports name the company as acquirer and 175,702 as target. The most active acquirer is KKR (651 reports), followed by Blackstone and ril.com.
5. Coverage is concentrated: the 50 most-reported companies hold 7.1% of all reports; Amazon alone has 10,280.
6. Lawsuits: 68,322 reports; the company is the defendant in 67% of those with a side.

Events by type and year

Event reports after deduplication, full data. 2026 covers January to June.

Most-reported companies (top 50)

Large groups can appear under more than one domain (for example a national site), as resolved in the company database.

Most active acquirers in M&A (top 50)

Counts are deduplicated news reports, not closed deals: one large deal reported on several days counts more than once.

Layoff reports by sector and year (top 25 sectors)

Layoff reports by country and year (top 25 countries)

Companies with the most layoff reports (top 50)

Lawsuit reports by sector (top 30)

IPO reports by company country and year (top 25)

Reports about listings, offerings and IPO filings, counted by the company's home country.

Contract reports by sector (top 30)

Events by company country (top 30)

Column dictionary

12 columns in the sample CSV.

Method

1. Extraction. Radar's pipeline reads each business-news article, classifies event sentences into types and links the company named to its website domain. One article can yield several events.
2. Scope. Thirteen event types: M&A, company announcements, contracts, asset acquisitions, lawsuits, new products, regulatory decisions, IPOs, job creation, layoffs, share buybacks, incorporations and FDA approvals. Leadership and board events (people data), earnings, and funding events are excluded.
3. Bulk back-fills. Days carrying more than five times the usual number of reports (archive back-fills stamped with an ingest date rather than the publication date) are dropped.
4. Dates. The event date is the article's publication date, read in Indian Standard Time. Rows dated before 2019 or after 30 June 2026 are dropped (the source holds some impossible dates, up to 2077).
5. Deduplication. The same company, event type and normalised sentence is kept once (earliest report). Reports of one company and event type on the same day whose sentences share 60% or more of their words are merged.
6. Text. Sentences shorter than 25 characters, sentences naming a third-party data provider and sentences carrying an email address or phone number are dropped; long sentences are cut to 300 characters.
7. Company fields. Name, country and sector come from Fliar's company database, joined on the domain. Countries are normalised (USA / US → United States).

8. News columns. Publisher and URL come from the article the sentence was read from, when it can be retrieved at release time.

9. Sample. Each companie key is ranked by sha1(key). Keys are taken in rank order until the file holds max(25%, 5,000) rows (capped at 100,000).

10. Profile links. Links are built from Fliar's company index and checked against it. A row whose company has no public profile has an empty link.

Limitations

• Derived from news: unreported events are missing and coverage favours large, English-reported companies.
• A row is a deduplicated news report, not a verified transaction. A big deal reported on several days appears more than once, and a sentence can recall an older event in a newer article.
• Events are linked to the domain the company database resolves; a multinational can appear under a national domain, which also sets its country.
• Event types come from a classifier and some sentences are mis-typed.
• About 9% of domains have no match in the company database, so their name, country and sector are blank.
• 2026 covers January to June only.
• news_source and news_url are blank when the article could not be retrieved at release time.

Cite this dataset

@misc{fliarbi_corporate_events_2019_2026_2026,
  author    = {{Fliar BI}},
  title     = {{Corporate Events 2019–2026: M&A, Layoffs, IPOs}},
  year      = {2026},
  publisher = {Zenodo},
  doi       = {10.5281/zenodo.23432171},
  url       = {https://doi.org/10.5281/zenodo.23432171},
  note      = {Sample release. Licence CC BY-NC 4.0}
}