Global startup funding rounds 2019–2026: a cleaned, deduplicated dataset

We describe a dataset of 80,911 venture-style funding rounds raised by 47,428 companies worldwide between January 2019 and June 2026, extracted from business news and cleaned with a reproducible pipeline. Repeated reports of one deal are merged, recaps of earlier rounds are dropped and every amount is checked against the sentence it was read from; 66,936 rounds carry a disclosed amount, together US$2.3T. Each row keeps the source sentence and, where retrievable, the article link, so any record can be audited. A hash-chosen 25% sample (20,227 rounds, whole companies) is public under CC BY-NC 4.0, together with four regional and thematic companions (AI, India, United States, United Kingdom). We report the extraction and cleaning method, the amount-verification rules, validation counts and known limitations. Datasets described Global Startup Funding Rounds 2019–2026: Kaggle, Hugging Face, DOI 10.5281/zenodo.23431596 AI Startup Funding Rounds 2021–2026: Kaggle, Hugging Face, DOI 10.5281/zenodo.23431429 Indian Startup Funding Rounds 2015–2026: Kaggle, Hugging Face, DOI 10.5281/zenodo.23431359 US Startup Funding Data 2019-2026: 31k Rounds: Kaggle, Hugging Face UK Startup Funding Data 2019-2026: 6.7k Rounds: Kaggle, Hugging Face Data licence CC BY-NC 4.0 (attribution "Fliar BI"); this paper CC BY 4.0. Contact: [email protected].