IPO Data 2019-2026: India, US, UK & Global IPOs

Key findings

1. 13,316 company IPOs, January 2019 to June 2026, one row per company. 3,014 have a report that the shares listed or the issue was priced; 186 were reported withdrawn or postponed; the rest were announced or filed.
2. The United States (3,412 IPOs, 26%) and India (2,913, 22%) lead, far ahead of China (841), Canada (723) and the United Kingdom (486). Of the Indian IPOs with a known board, 941 are SME issues and 955 mainboard.
3. Indian SME IPOs took off after 2022: 85 in 2022, 164 in 2023, 276 in 2024 and 268 in 2025. The median SME issue with a verified size is ₹31 cr; the median mainboard issue is ₹650 cr.
4. 2021 was the peak year (2,857 IPOs, 323 of them blank-check SPACs), against 920 in 2019, 1,615 in 2022 and 2,225 in 2025. Completed listings were 697 in 2021 and 557 in 2025.
5. Largest completed IPO: SpaceX ($75.0B, Nasdaq, 2026), then Saudi Aramco ($29.4B, 2019) and Rivian ($12.0B). India's largest is Hyundai Motor Company (₹27,390 cr, 2024).
6. SPACs: 690 blank-check IPOs and 785 companies that went public by merging with one; 47% of the blank-check IPOs are dated 2021.

IPOs by year: status, India, US, China, UK and SPACs

One row per company IPO. "Raised" sums completed IPOs whose size is verified against the report sentence and reported more than once.

IPOs by country and year (top 30 countries)

Largest IPOs worldwide, 2019 to June 2026 (top 50)

Completed IPOs with a verified size, ranked by US$. Sizes are as reported in the news and include any upsizing reported at pricing.

Largest US IPOs (top 50)

US-headquartered companies, any exchange.

Largest Indian mainboard IPOs (top 50)

Rupee figures as reported; a US$-only figure is shown in rupees at ₹83 = US$1.

India SME vs mainboard IPOs by year

Board is known for Indian IPOs whose reports name the exchange or the SME segment; others are counted in the country tables only.

Largest Indian SME IPOs (top 30)

Verified SME issue sizes, any status (most SME issues are reported at subscription rather than at listing).

IPOs by exchange (top 40)

Exchange as reported; dual listings count under the exchange named most often. Companies list on foreign exchanges too (Chinese companies on NYSE, for example).

US IPOs by exchange and year

US-headquartered companies.

UK IPOs by exchange and year

UK-headquartered companies (London Main Market, AIM and overseas listings).

Largest UK IPOs (top 26)

IPOs by sector

Sector is read from the IPO reports, else from the company database, and grouped into 20 broad sectors.

IPOs by sector and year

Column dictionary

25 columns in the sample CSV.

Method

1. Extraction. Radar's pipeline reads each business-news article, classifies IPO sentences and extracts the company, exchange, currency, issue size, price band, issue price, listing date and IPO type. Follow-on offerings, offers for sale on their own, QIPs and rights issues are excluded.
2. One row per company IPO. Reports are grouped by the company's website; reports with only a name join the website their name maps to. Name variants are then merged in four passes: one name's words start the other's or the names are 90%+ similar (same year, compatible country); a short and long name on the same exchange and year (Uber / Uber Technologies); the same verified US$200M+ amount in the same year under related names (ADNOC / ADNOC Gas); a name-only report whose words contain a linked company's full name. Series numbers keep SPACs apart (Acquisition II vs III). A handful of merges were made by hand after reading the largest rows.
3. Names. Descriptions are stripped ("Saudi Arabia's state-owned oil giant Aramco" → Aramco); placeholders, exchanges, people and descriptions ("Zomato-backed startup") are dropped.
4. Status. Completed needs a report sentence that names the company and says the shares listed, debuted or began trading, the issue was priced or oversubscribed, or the IPO raised money, with no plan language (will, expected, plans, filed, draft) unless the sentence also says the stock debuted or traded. Withdrawn / postponed is a company with a quarter or more of its reports about a suspension, withdrawal, postponement or cancellation and no trading report.
5. Dates and year. The listing date is taken from completion reports and must fall within 120 days before to 30 days after the article. The IPO year is the listing year, else the year most completion reports were written, else the year of the first report. IPOs listed before 2019 (recalled in later articles) are dropped, as are companies known only from roundup articles that name six or more IPOs. Reports from archive back-fill days (dates carrying five times the usual volume, old articles stamped with an ingest date; 7,023 reports) are dropped.
6. Issue size. A figure stands when (a) its currency is written in the sentence, (b) it matches a money figure in the sentence (0.6–1.7×), (c) it does not sit next to a valuation, deal-value, revenue or subscription-demand word, and (d) the sentence reports a completed IPO or concrete offer terms (price band, fresh issue, offer for sale, lot size) without target or valuation language. For a completed IPO the completion reports are used. The figure most distinct sentences agree on wins. Amounts above US$80B are dropped (the largest in the data is US$75B). Indian SME amounts above US$60M are dropped. SPAC-merger "sizes" (deal values) are blanked.
7. Hand checks. The largest rows were read against their sentences: a rights issue, a follow-on offering and preferred-stock deals were removed, and amounts that were another figure were blanked (listed in the spec).
8. Exchange and board. About 2,600 exchange spellings are mapped to 80 exchanges. Indian issues are Mainboard (NSE, BSE) or SME (NSE Emerge, BSE SME, or an SME issue whose exchange is not stated).
9. Company fields. Country is the reports' company country, else the company database, else the exchange's country. Sector is read from the reports, else from the company database, and grouped into 20 broad sectors.

10. News columns. The excerpt is the earliest clean report sentence, preferring one whose article can be retrieved at release time.

11. Sample. Each company 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).

12. 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: IPOs that were never reported are missing, and coverage favours English-language reporting. Year-to-year counts also reflect how much news was ingested; the 2025 jump in announced IPOs is partly wider news coverage.
• Most rows are announcements or filings without a completion report (Announced / filed); many of those did list, others never happened. Use listing_status == "Completed" for listings.
• Issue sizes are as reported in the news (some include the greenshoe, some do not) and only about one row in five has one; the US$ figure uses the reported dollar amount or a fixed average exchange rate.
• Company matching is automatic: some companies still appear twice under name variants, and a few rows are linked to the wrong website (a parent or a namesake).
• The full ₹ issue size is shown for Indian rows; US$-only figures are converted at ₹83 = US$1 in the India tables.
• 2026 covers January to June only.
• news_source and news_url are blank when the article could not be retrieved at release time (some article stores were offline).

Cite this dataset

@misc{fliarbi_ipo_tracker_2019_2026_2026,
  author    = {{Fliar BI}},
  title     = {{IPO Data 2019-2026: India, US, UK & Global IPOs}},
  year      = {2026},
  publisher = {Fliar BI},
  url       = {https://fliarbi.com/datasets/ipo-tracker-2019-2026/},
  note      = {Sample release. Licence CC BY-NC 4.0}
}