IPO watch · November 19, 2026 · Product data through October · Model evaluations checked November 19
Anthropic priced its IPO at $83 per share on November 18. On November 19, ANT’s first Nasdaq trade came out of the opening cross at $111.50 around 11:48 a.m. New York time, 34.3% above the offer price, and the stock closed at $97.17, up 17.07%. A year earlier, Claude ranked sixth by website visits among eight general-purpose AI assistants tracked by AlphaMoat. It now ranks third, and higher still in coding evaluations. How often do these different entry points end up on the same enterprise customer’s bill? Pricing release · First-day quote · First-trade log
Four changes to watch at the IPO
Jump to a sectionClaude rose from sixth to third, reaching third in March and holding it through October.
Downloads compare the same calendar month a year apart. Extension data begins in April 2026, so it is a six-month comparison.
Claude 4.5 took 56 days and 5.5 took 15. In November, Fable 5.2 Max led the Artificial Analysis Intelligence Index with 65 points.
Large customers are those generating at least $100K in trailing twelve-month revenue.
01 Web reach grew faster than the peer group
Claude.ai rose from 197.0M visits in October 2025 to 1.3B a year later (+536.3%). Its share of the same eight sites went from 2.33% to 11.63%, a gain of 9.30 percentage points. Growth across the entire peer group explains only part of that increase.
The climb came in stages. Claude.ai passed 600 million visits in March and approached one billion in May. June through August held around 950–970 million. September was the first month above a billion in this window, and October added roughly 241 million more visits. The annual comparison contains both a spring step-up and an autumn acceleration.
02 Downloads, in-app purchases and browser adoption diverged
| Metric | Baseline · value | 2026-10 | Growth | Oct MoM |
|---|---|---|---|---|
| Claude.ai visits | 2025-10 · 197.0M | 1.3B | +536.3% (YoY) | +23.8% |
| App downloads | 2025-10 · 1.3M | 21.2M | +1,509.7% (YoY) | +14.8% |
| App IAP revenue (USD) | 2025-10 · $3.9M | $71.5M | +1,734.6% (YoY) | -10.5% |
| Chrome extension users | 2026-04 · 7M | 21M | +200% (Apr–Oct 2026) | +31.3% |
Mobile started from a smaller base. October downloads reached 21.2M, or 16.1 times the year-earlier level. April was the high point in this window at 29.2M. October rebounded +14.8% from September yet remained 27.5% below April.
App-store IAP revenue peaked in September, five months after downloads, and eased to about $71.46 million in October. Downloads count acquisitions in the period, while purchases also include renewals. Dividing one month’s revenue by that month’s downloads would not yield a conversion rate.
The official extension grew from 7 million users in April 2026 to 21 million in October, a 200% increase. October alone added 5 million, more than a third of the six-month gain.
Explore monthly web, app and browser extension data
| Month | Claude.ai visits | App downloads | App IAP revenue (USD) | Panel share |
|---|---|---|---|---|
| 2025-10 | 197.0M | 1.3M | $3.9M | 2.33% |
| 2025-11 | 180.3M | 1.4M | $5.0M | 2.18% |
| 2025-12 | 172.7M | 1.4M | $5.8M | 2.07% |
| 2026-01 | 202.9M | 2.6M | $9.8M | 2.27% |
| 2026-02 | 290.3M | 7.7M | $18.4M | 3.37% |
| 2026-03 | 613.7M | 25.1M | $73.6M | 6.14% |
| 2026-04 | 823.5M | 29.2M | $31.6M* | 8.12% |
| 2026-05 | 952.6M | 27.4M | $75.2M | 9.12% |
| 2026-06 | 946.8M | 18.2M | $75.4M | 9.39% |
| 2026-07 | 969.3M | 18.1M | $75.7M | 9.71% |
| 2026-08 | 950.2M | 18.1M | $79.1M | 9.42% |
| 2026-09 | 1.0B | 18.5M | $79.9M | 9.72% |
| 2026-10 | 1.3B | 21.2M | $71.5M | 11.63% |
* April 2026 IAP is a coarse source estimate, less precise than other months.
| Month | Chrome extension users |
|---|---|
| 2026-04 | 7M |
| 2026-05 | 8M |
| 2026-06 | 10M |
| 2026-07 | 12M |
| 2026-08 | 15M |
| 2026-09 | 16M |
| 2026-10 | 21M |
03 Faster rollouts, and the time behind a flagship score
Sonnet opened the 4.5 rollout in autumn 2025. Buyers had to wait until Opus arrived to compare all three tiers. In 2026 Opus went first, Sonnet followed six days later and Haiku nine days after that. Teams could compare three options for capability, latency and cost about two weeks into the new generation.
Artificial Analysis Intelligence Index v5.0.1 · as of 2026.11.19
Claude Fable 5.2 (Max, Default Fallback)
The index combines nine evaluations across areas including knowledge work, coding and science. At maximum effort, the model averaged 41.6 minutes and $26.54 per index task. For a product that needs an immediate response, both figures matter alongside the top score. Model results · Index methodology
| Task | Model and effort | Rank |
|---|---|---|
| Web development | Claude Fable 5.2 (xHigh) | #1 |
| Web development | Claude Opus 5.5 (Max) | #2 |
| Overall text | Claude Opus 5.5 (High) | #5 |
Fable 5.2 and Opus 5.5 took the first two spots for web development. On the overall text board the highest Claude entry was Opus 5.5 High in fifth. A team prototyping websites and a team building an everyday chat assistant have good reasons to consult different boards.
04 From chat to code, design and bots: how many ways in?
A team can hand a report to Cowork and turn it into slides in the same conversation. An engineer can start in the terminal, while a colleague delegates work from Slack. Claude reaches across existing tools even as some tasks converge in a single interface.
One name deserves care. ClaudeBot in a website log is Anthropic’s training crawler. @Claude in Slack is a collaboration surface. Counting crawler requests as human adoption would mix two different activities. Crawler documentation
05 The commercial question: will a customer buy both layers?
The prospectus provides a comparison beyond traffic. More than 950 customers generated at least $100,000 each in trailing twelve-month revenue as of September 2025. A year later that count exceeded 11,900. Third-quarter 2026 company revenue was $17.0 billion, up 1,127% year over year, across the platform, applications and other revenue sources. Final prospectus
Claude Code gives a clue to spending intensity. In Q3 2026, Enterprise users who did not use Code spent less than $75 per month on average. Among Enterprise users who used Code, the top 5% spent about 20 times that amount. Comparing a heavy-use tail with another group’s average does not tell us how an ordinary user’s bill changes after adopting Code. Enterprise spending disclosure
Share using Claude Platform and at least one first-party application
Cross-adoption is more revealing. In that customer tier, about 30% used Claude Platform plus at least one first-party app in Q4 2025. By Q3 2026 the share exceeded 50%. Model selection by developers and app adoption by knowledge workers increasingly coexist in the same customer account. Customer cohort disclosure
Three readings can test this thesis next quarter: whether Claude keeps a double-digit share of the web panel, whether mobile IAP returns to its September level, and whether more large customers use both platform and apps. They measure distribution, store spending and enterprise adoption respectively. If they diverge, the first one to turn matters more than another headline growth rate.