Anthropic versus the 39 IPO winners: prospectus comparison
Snapshot: October 27, 2026. Financial facts come from issuer registration statements filed before the first trading day; Anthropic is its public October 26, 2026 S-1. The 39 are the existing cohort that beat both SPY and QQQ at 90 days and one year, not every one-year winner. These 39 come from 203 issuers with usable one-year paired histories in a 588-issuer census; missing and delisted histories can bias the sample. They were selected using later outcomes, so their common features cannot establish a strategy or predict Anthropic.
Anthropic: October 26, 2026 S-1. The filing does not yet supply a public IPO price range or final capitalization, so no defensible offer-price valuation or return forecast is implied here.
Headline comparison
| Measure | Anthropic | 39 historical winners |
|---|---|---|
| Latest complete fiscal-year revenue | FY2025 $4,585m | Median $390.4m; 31/39 below $1,000m |
| Annual revenue growth | FY2025 +1,087.8% | Median +33.6% across 38 with meaningful comparable prior years |
| Largest disclosed peer growth | — | CoreWeave FY2024 +736.6% |
| Latest complete fiscal-year net income / loss | FY2025 loss $41,965m, including $34,308m fair-value change | 10 positive, 28 negative across 38 issuers with a comparable figure; TW omitted |
| Latest complete fiscal-year operating result | FY2025 operating loss $8,060m | Accounting and cost classification differ by issuer |
| Latest interim revenue | H1 2026 $16,286m, up 1,300.3% year over year | Interim periods differ across issuers; not pooled as full-year revenue |
| Latest interim operating result | H1 2026 operating loss $693m (4.3% of revenue) | Business models and accounting presentations vary |
| Latest interim net loss | H1 2026 $80,748m, including $72,388m change in fair value of convertible notes, derivative liabilities and SAFEs | Do not equate fair-value charges with cash operating losses |
FY2025 Anthropic revenue exceeded the latest pre-listing full-year revenue of 38 of 39 winners; GlobalFoundries reported $4,851m for FY2020. These are nominal dollars from different years, without inflation adjustments. Their accounting periods, reporting currencies and gross/net revenue presentation differ. Revenue scale is not valuation.
Business and risk comparisons
- CoreWeave (CRWV), a cloud/GPU infrastructure provider: pre-trading S-1/A. FY2024 revenue was $1,915.4m, up 736.6%, and operating income was $324.4m; its net loss was $863.4m. Microsoft was 62% of 2024 revenue and the top two customers 77%. It is one of the closer infrastructure-growth analogies in this selected group, although it provides compute capacity while Anthropic sells AI models, APIs and applications.
- Anthropic: H1 2026 revenue was $16,286m; compute and infrastructure expense was $8,945m, included in operating expenses. Revenue was primarily consumption-based. Its top five customers represented approximately 25% of revenue in the nine months ended September 30, 2026, and no single customer exceeded 10%. These concentration figures use different periods and different numbers of customers from CoreWeave. The filing describes multi-year, fixed compute obligations and variable demand. Some total-commitment figures remain blank in this preliminary filing.
- DigitalOcean (DOCN), Kingsoft Cloud (KC) and WhiteFiber (WYFI): cloud infrastructure or hosting make them relevant for compute supply and capacity economics, though they are not frontier model developers. Their latest reported pre-listing annual revenues were $318.4m, $568.3m and $47.6m respectively. Kingsoft’s USD amount is translated from RMB in its filing. WhiteFiber has no meaningful prior-year growth rate because its prior period had zero revenue.
- Zoom (ZM), DocuSign (DOCU), Elastic (ESTC), Appian (APPN) and Zscaler (ZS): useful for software distribution and enterprise adoption, but not close infrastructure-cost analogues. Zoom had FY2019 revenue of $330.5m and net income of $7.6m, whereas Anthropic reported a GAAP operating loss in its latest interim period.
- Anthropic reported approximately 900% net revenue retention as of September 30, 2026. This is its own qualified-customer measure, excluding Claude Pro and Max fixed-price subscribers; it should not be compared directly with differently defined retention figures in another prospectus. The filing reported more than 11,900 customers with at least $100,000 in trailing-twelve-month revenue, more than 1,900 above $1m and approximately 250 above $10m.
All 39 pre-listing prospectus comparisons
The financial year shown is the latest reported full-year figure in the cited financial table; Schrödinger is an audited FY2018 figure while its filing also disclosed preliminary FY2019 revenue. Growth is from the immediately prior reported fiscal year, using local currency where the filing also provides a USD translation. IPO return starts at the first closing price and ends at the paired close at or before the one-year anniversary; it excludes dividends. “IPO return” is not excess return.
| IPO | Business label from census | Latest fiscal-year revenue (USD m) | Annual growth | Fiscal-year net income / loss (USD m) | 1-year IPO return | Filing |
|---|---|---|---|---|---|---|
| ALGM | Magnetic-sensor and analog power integrated circuits | 2020: 650.1 | -10.2% | +37.1 | +88.5% | S-1/A 2020-10-21 |
| APPN | Low-code application development software | 2016: 132.9 | +19.5% | -12.5 | +107.8% | S-1/A 2017-05-12 |
| BILI | Online video, gaming and entertainment community | 2017: 379.4 | +371.7% | -28.2 | +61.0% | F-1/A 2018-03-23 |
| BILL | Cloud financial workflow and payment automation software | 2019: 108.4 | +67.0% | -7.3 | +304.0% | S-1/A 2019-12-10 |
| BL | Accounting and finance automation cloud software | 2015: 83.6 | +61.8% | -24.7 | +48.8% | S-1/A 2016-10-27 |
| CARG | Consumer | 2016: 198.1 | +101.0% | +6.5 | +55.4% | S-1/A 2017-09-29 |
| CRSR | COMPUTER PERIPHERAL EQUIPMENT | 2019: 1,097.2 | +17.0% | -8.4 | +95.6% | S-1/A 2020-09-18 |
| CRWV | Cloud computing | 2024: 1,915.4 | +736.6% | -863.4 | +87.0% | S-1/A 2025-03-20 |
| DAO | Online learning tools and education platform | 2018: 106.6 | +60.5% | -30.5 | +112.0% | F-1/A 2019-10-18 |
| DOCN | Cloud computing infrastructure and developer platform | 2020: 318.4 | +24.9% | -43.6 | +42.0% | S-1/A 2021-03-23 |
| DOCU | Electronic signature and agreement automation software | 2018: 518.5 | +35.9% | -52.3 | +44.1% | S-1/A 2018-04-25 |
| ESTC | Search and data-analytics software | 2018: 159.9 | +81.4% | -52.7 | +22.4% | S-1/A 2018-10-02 |
| FOUR | Integrated electronic payment processing technology | 2019: 731.4 | +30.5% | -58.1 | +183.2% | S-1/A 2020-06-03 |
| GFS | Semiconductor foundry | 2020: 4,851.0 | -16.5% | -1,351.0 | +23.4% | F-1/A 2021-10-19 |
| HNGE | Digital care software | 2024: 390.4 | +33.4% | -11.9 | +47.1% | S-1/A 2025-05-16 |
| HUYA | Video-game livestreaming platform | 2017: 335.8 | +174.2% | -12.4 | +36.9% | F-1/A 2018-05-10 |
| ICHR | Semiconductor manufacturing gas and chemical delivery subsystems | 2015: 290.6 | +16.7% | +5.6 | +151.8% | S-1/A 2016-11-29 |
| IIIV | Business Services | 2017: 262.6 | +31.5% | +0.9 | +46.9% | S-1/A 2018-06-14 |
| IQ | Online video streaming and entertainment platform | 2017: 2,671.0 | +54.6% | -574.4 | +53.8% | F-1/A 2018-03-26 |
| KC | Cloud computing infrastructure and platform services | 2019: 568.3 | +78.4% | -159.6 | +70.8% | F-1/A 2020-05-06 |
| MBLY | Driving software and chips | 2021: 1,386.0 | +43.3% | -75.0 | +25.7% | S-1/A 2022-10-18 |
| MIAX | Electronic financial exchange technology | 2024: 1,140.1 | +9.5% | +102.0 | +41.6% | S-1/A 2025-08-04 |
| PSN | Cybersecurity and intelligence software alongside defense and infrastructure engineering | 2018: 3,560.5 | +18.0% | +239.4 | +29.1% | S-1/A 2019-04-29 |
| PUBM | Programmatic digital-advertising cloud infrastructure | 2019: 113.9 | +14.7% | +6.6 | +31.3% | S-1/A 2020-11-30 |
| RDDT | Social platform | 2023: 804.0 | +20.6% | -90.8 | +129.4% | S-1/A 2024-03-19 |
| ROKU | TV streaming software platform and connected devices | 2016: 398.6 | +24.6% | -42.8 | +210.8% | S-1/A 2017-09-18 |
| SDGR | Scientific software | 2018: 66.6 | +19.7% | -28.4 | +247.9% | S-1/A 2020-01-27 |
| SFIX | Personalized multibrand Internet apparel retail combining algorithms and styling services | 2017: 977.1 | +33.8% | -0.6 | +76.8% | S-1/A 2017-11-06 |
| SITM | Semiconductors | 2018: 85.2 | -15.7% | -9.3 | +341.3% | S-1/A 2019-11-07 |
| SPT | PREPACKAGED SOFTWARE | 2018: 78.8 | +75.9% | -20.9 | +178.6% | S-1/A 2019-12-06 |
| TEM | Clinical data software | 2023: 531.8 | +65.8% | -214.1 | +77.1% | S-1/A 2024-06-13 |
| TW | SECURITY & COMMODITY BROKERS, DEALERS, EXCHANGES & SERVICES | 2018: 657.6 | +30.4% | N/M | +26.9% | S-1/A 2019-04-02 |
| U | Real-time 3D content development software | 2019: 541.8 | +42.3% | -163.2 | +100.0% | S-1/A 2020-09-16 |
| ULS | SERVICES - TESTING LABORATORIES | 2023: 2,678.0 | +6.3% | +276.0 | +61.6% | S-1/A 2024-04-02 |
| UPST | Finance Services - Lending | 2019: 164.2 | +65.3% | -5.0 | +377.2% | S-1/A 2020-12-09 |
| WAY | Healthcare payments software | 2023: 791.0 | +12.2% | -51.3 | +98.8% | S-1/A 2024-05-28 |
| WYFI | Cloud computing | 2024: 47.6 | N/M | +1.4 | +51.9% | S-1/A 2025-08-01 |
| ZM | Cloud video communications software | 2019: 330.5 | +118.2% | +7.6 | +142.0% | S-1/A 2019-04-16 |
| ZS | Cloud cybersecurity platform | 2017: 125.7 | +56.5% | -35.5 | +103.6% | S-1/A 2018-03-13 |
Qualifications: BILI, DAO, HUYA, IQ and KC provide USD translations of RMB reporting in their prospectuses; growth is measured in RMB. FOUR and IIIV report gross payment-related revenue; MIAX’s exchange revenue includes transaction-related costs. TW’s $657.6m is pro forma net revenue: it equals the sum of predecessor and successor periods; its growth uses the filing’s FY2017 net revenue. TW reports net revenue, whereas FOUR reports gross revenue. Schrödinger’s FY2018 audited figures are shown; the filing also disclosed preliminary unaudited FY2019 revenue of $84.9m–$86.2m. WYFI’s earlier period had zero revenue. N/M = not meaningful or not calculated. Latest annual years and currencies are not synchronized, and industry labels are descriptive. Net-income figures generally include noncontrolling interests; DOCN and ROKU are labeled losses attributable to common stockholders and equal reported net losses; UPST is before allocation to noncontrolling interests (its loss attributable to common stockholders was $0.466m); PSN includes noncontrolling interests (its attributable income was $222.337m); ICHR includes a $7.181m discontinued-operations loss; SPT presents net and comprehensive loss together. TW is omitted from the comparable net-income count: its pro forma income was $99.526m and its historical predecessor/successor periods sum to $159.467m. Net income is subject to non-cash accounting charges and is not cash flow.
BILI, CRSR, GFS and ICHR filed abbreviated last amendments. The table cites the most recent earlier substantive pre-trading filing for financial data; both the amendment and financial filing are retained in the evidence index. A dated filing establishes availability before the trading date, not the exact information known by every investor. The final priced prospectus is listed separately in the index and is not used as an input to the pre-trading comparison.
Implications for the original IPO and options question
The prior 39 winners are an outcome-selected sample. Their growth and size describe successful historical cases but do not identify an investable entry rule. A useful next test would compare the same prospectus features with the non-winners in the full eligible, observable cohort, then evaluate a rule using only pre-trading facts in time-ordered holdouts. For cash-secured puts, actual strikes, expirations, implied volatilities, bid/ask prices and options availability after listing are needed; none can be inferred from these prospectuses.
Sources: Anthropic S-1; each issuer filing is linked in the table. The detailed machine-readable ledger records each selected source table, raw revenue row, annual-period caveat and three benchmark returns.
Download the comparison ledger · Download the filing index · Download extracted table evidence