AI洞察/Anthropic IPO: Claude Beyond the Chat Window
行業洞察2026-11·閱讀約 8 min

Anthropic IPO: Claude Beyond the Chat Window

Thirteen months of Claude web and app data, six months of Chrome extension growth, Artificial Analysis benchmarks and prospectus disclosures show how Anthropic reaches consumers and enterprises.

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 section

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.

Figure 1 · Claude climbed three places from December 2025 to March 2026
Oct 2025
Third by March 
Oct 2026
ChatGPT
Gemini
DeepSeek
Perplexity
Grok
Claude
Doubao
Kimi
Claude’s monthly ranks: 6, 6, 6, 5, 4, 3, 3, 3, 3, 3, 3, 3, 3
1ChatGPT
2Gemini
3Claude
4DeepSeek
5Grok
6Doubao
7Perplexity
8Kimi
Oct2025Jan2026AprJulOct
Claude share of panel visits2.33% → 11.63%
Each column is one month, ranked by website visits within the fixed eight-assistant panel. On wider screens, percentages on the right are October 2026 shares of the panel’s combined visits.

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

Figure 2 · Downloads peaked in April, IAP in September, while Chrome kept climbing
App downloadsPeak 2026.04 · 29.2M
Oct
21.2M
App IAP revenuePeak 2026.09 · $79.9M
Oct
$71.5M
Chrome extension usersPeak 2026.10 · 21M
Oct
21M
Oct2025Jan2026AprJulOct
Rows share one timeline. Each is scaled from zero to its own peak in the window. Chrome extension data begins in April 2026. The dark bar marks each row’s peak; the right column shows October 2026 values. The hatched bar is April 2026 IAP, a coarse source estimate less precise than other months; its dip should not be read as a revenue decline.
Table 1 · Growth from the baseline shown for each metric
MetricBaseline · value2026-10GrowthOct MoM
Claude.ai visits2025-10 · 197.0M1.3B+536.3% (YoY)+23.8%
App downloads2025-10 · 1.3M21.2M+1,509.7% (YoY)+14.8%
App IAP revenue (USD)2025-10 · $3.9M$71.5M+1,734.6% (YoY)-10.5%
Chrome extension users2026-04 · 7M21M+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.

Chrome added 14 million extension users in six months

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
Web and app · October 2025 to October 2026
MonthClaude.ai visitsApp downloadsApp IAP revenue (USD)Panel share
2025-10197.0M1.3M$3.9M2.33%
2025-11180.3M1.4M$5.0M2.18%
2025-12172.7M1.4M$5.8M2.07%
2026-01202.9M2.6M$9.8M2.27%
2026-02290.3M7.7M$18.4M3.37%
2026-03613.7M25.1M$73.6M6.14%
2026-04823.5M29.2M$31.6M*8.12%
2026-05952.6M27.4M$75.2M9.12%
2026-06946.8M18.2M$75.4M9.39%
2026-07969.3M18.1M$75.7M9.71%
2026-08950.2M18.1M$79.1M9.42%
2026-091.0B18.5M$79.9M9.72%
2026-101.3B21.2M$71.5M11.63%

* April 2026 IAP is a coarse source estimate, less precise than other months.

Chrome extension · April to October 2026, up 200%
MonthChrome extension users
2026-047M
2026-058M
2026-0610M
2026-0712M
2026-0815M
2026-0916M
2026-1021M

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.

Figure 3 · The three-tier rollout window shrank by 41 days
The axis counts days since the first launch in each generation, on a shared 0–56 day scale. Letters mark each tier; links below open the Anthropic announcement.

Artificial Analysis Intelligence Index v5.0.1 · as of 2026.11.19

Claude Fable 5.2 (Max, Default Fallback)

#1Overall rank
65Index score
41.6Minutes per task
$26.54USD per task

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

Table 2 · Arena user preference, November 18
TaskModel and effortRank
Web developmentClaude Fable 5.2 (xHigh)#1
Web developmentClaude Opus 5.5 (Max)#2
Overall textClaude 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.

Grouped by where the work happens; each product links to its official description. The dashed card marks a research preview. ClaudeBot in website logs is a training crawler, not one of these surfaces. Crawler documentation

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

Figure 5 · Joint adoption crossed the halfway mark
Q4 2025
≈30%
Q3 2026
>50%

Share using Claude Platform and at least one first-party application

Each square is 1% of large customers, defined as those generating at least $100K in trailing twelve-month revenue in each period. The company disclosed an approximation and a lower bound; the grids show 30 and 50 squares, and “+” marks that the Q3 2026 share was above 50% by an undisclosed amount. Final prospectus

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

AlphaMoat view

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.

相關閱讀

Updated monthly
AI Traffic & App Statistics

AI chatbot market share, fastest-growing AI tools, categories and countries, plus AI app downloads and in-app purchase revenue.

2026-10
Work vs Personal Agents: AI’s Next Battleground

Work vs Personal Agents: September 2026 traffic, growth and opportunities across 100 products, with Meta Muse, OpenAI Dots, Grok Bot and Manus Cue.

2026-09
AI Gateways: Crowded, Leader Gains (Sep 2026)

74 new entrants in twelve months, traffic up 69% against a falling market — and leader OpenRouter gained share rather than losing it, from 37.4% to 49.3%. The breakdown: 21.6% of the growth was carried in by new supply, incumbent demand grew only 32.7%, and China supplies 19.3% of visits — 4.8x its share of the wider market. Full Top 50 included.

2026-11
GEO/AEO Market: 401 Tools, Six Segments

AlphaMoat maps 401 tagged tools, comparing leaders, traffic gains and July–October 2026 trends across six GEO/AEO segments.

← 返回洞察列表