Anthropic IPO: What Claude’s Web and App Data Show

Using Similarweb’s web and app data, we looked at how Claude attracted users, whether they kept coming back, and how often they reached usage limits or upgrade pages as the platform grew.

This analysis reflects Claude’s consumer-facing position; its relative standing in API consumption, which sits beyond our measurement, could deviate from market share levels implied by consumer data.

Figures cited from press reports include company statements and third-party estimates. See full disclaimers and disclosures at the end of this post.

Core insights

• Claude’s share of website visits among major AI chatbots rose from 2% to 9.6% in the 12 months through August 2026. ChatGPT’s share fell from 78% to 57%.
• User acquisition and upgrade signals accelerated. In August 2026, sign-up flow visits reached 21.4M, +504% YoY, and upgrade pages drew 14.7M unique visitors, +476% YoY. Corporate login sessions grew 166 times YoY.
• Retention improved as the user base grew. 45% of users who joined in February 2026 were still active six months later, up from 25.8% for the January 2025 cohort. The gap to ChatGPT narrowed from 38 points to 14 points.
• The share of Claude users who also used ChatGPT in the same month fell from 78% to 61%, while the share of ChatGPT users who also used Claude rose from 2% to 14%.

What web and app data shows about Anthropic ahead of a potential IPO

Anthropic may be heading toward one of the most valuable public listings on record. Press reports suggest that the company could seek a valuation of more than $2 trillion, potentially placing it ahead of the valuations attached to the public debuts of SpaceX and Saudi Aramco.

A year ago, Claude accounted for about 2% of visits across the major LLMs in our analysis, while ChatGPT accounted for nearly 80%. By August 2026, Claude’s share had risen to 9.6%, while ChatGPT’s had declined to 57%.

Those figures show how quickly Claude’s position in the market has changed. Claude’s growth also extends beyond headline traffic. Sign-ups, visits to plan-upgrade pages, corporate-login activity, and user retention all increased as its audience grew.

Over the past several months, Similarweb’s Investor Solutions team analyzed Claude’s web and mobile app activity. We looked at how Claude’s audience, engagement, and retention changed across its consumer, developer, and enterprise-facing properties.

Anthropic filed a confidential S-1 in June, and until it’s public, investors will not have an audited view of its revenue, margins, customer concentration, or financial position. Similarweb data cannot answer those financial questions, but our data shows Claude’s current growth across the web and mobile apps.

Claude gained share for twelve consecutive months

ChatGPT remains the clear category leader, but its total visit count declined year over year. DeepSeek, Perplexity, and Grok also remained below their peak months, while Claude and Gemini were the only two products in the category gaining share.

Between February and April 2026, monthly visits to claude.ai increased from 203 million to 824 million. Similarweb counts each session as a visit, so one user may account for multiple visits. One user visiting multiple times counts multiple times. The increase coincided with major product launches, a Super Bowl campaign, and heightened news coverage that pushed the Claude app to number one in the U.S. App Store based on reporting from CNBC.

Traffic did not decline after the initial surge. More recently, claude.ai recorded approximately 950 million to 969 million visits per month. Claude has narrowed the gap with ChatGPT, but ChatGPT still leads.

Claude’s audience is becoming more distinct

The audience overlap data provides another view of how Claude’s position has changed. The share of Claude users who also used ChatGPT in the same month fell from 78% to 61%, while the share of ChatGPT users who also used Claude rose from 2% to 14%. Claude is becoming more common among ChatGPT users while building a larger audience beyond that overlap.

Traffic to sign-up, upgrade, and corporate pages increased

Claude’s sign-up activity in August 2026 was six times higher than the prior baseline. User visits to Claude’s upgrade pages, where free users are presented with paid plans, reached 21 million and came from 14.7 million unique visitors.

Claude’s usage-limit page, which appears when users reach their quota, recorded 52 million visits in August 2026. The page was introduced in mid-2025 and registered fewer than 10,000 visits in its first month. Its growth since then partially reflects the expansion of Claude’s audience and the category.

Corporate-login sessions substantially increased every month from January through August 2026, reaching 163 times their level from the prior year. This shows considerably more activity through Claude’s corporate access, but it should not be read as a measure of contract value or volume.

Retention improved as Claude’s audience grew

Breakout growth does not always produce a durable audience. Clubhouse and BeReal both attracted significant media and user attention, and then struggled to sustain that momentum. Claude has avoided that pattern so far.

Among groups of users who signed up in the same month, those with a complete six-month history show retention rising from 25.8% for people who joined in January 2025 to 44.8% for those who joined in February 2026.

ChatGPT still leads at the six-month mark, with retention at 59% compared with Claude’s 45%. ChatGPT’s user retention remained broadly flat over the two years while Claude’s rose by 18 percentage points. Claude’s more recent cohorts show some softening at months three and four, which is worth watching as the 2026 intake matures.

What the data cannot tell

Similarweb’s global web and mobile app data provides a view of consumer behavior, developer-product usage and activity through enterprise-facing login pages. It cannot directly observe API consumption, contract values, revenue, margins, or customer concentration.

This matters because press estimates suggest that API usage and enterprise contracts account for roughly 80% of Anthropic’s reported revenue. Those parts of the business appear only indirectly in observable web and app activity.

CNBC and TechCrunch reported that Anthropic’s annualized revenue increased from approximately $9 billion at the end of 2025 to $65 billion by July 2026. Over the same period, Claude’s web visits increased 5.6 times, unique visitors increased 4.9 times, and app monthly active users increased 8.9 times.

Digital growth alone cannot validate Anthropic’s reported revenue.

Details to review in the S-1

The S-1 will show a view of Anthropic’s financials, and when Anthropic files, disclosures in four areas could help put the usage signals into context:

• Consumer subscription revenue. Visits to Claude’s upgrade pages grew 5.8 times, while usage-limit pages accounted for 5.5% of visits. Any disclosure of paid subscribers, conversion rates or consumer subscription revenue could help show whether that activity is translating into sales.
• Revenue mix and enterprise customers. Corporate-login sessions reached 166 times their year-earlier level, albeit from a small base. Details on Anthropic’s consumer, API and enterprise revenue could provide context for that increase.
• Customer retention and expansion. Six-month user retention reached 45% for Claude’s February 2026 cohort. Customer churn or net revenue retention, if reported, would help investors assess the durability of Anthropic’s revenue. These financial measures are different from web-user retention and should not be read as direct equivalents.
• User acquisition spend from sales & marketing. Paid digital channels played a limited role in Claude’s traffic growth. Sales and marketing disclosures could potentially provide a broader view of Anthropic’s cost base, but they may not isolate customer-acquisition costs.

Similarweb data supports a narrower conclusion: Claude’s growth extended beyond overall visits into repeat use, upgrade-page activity and corporate access. An S-1 could help investors assess how that usage relates to Anthropic’s revenue and cost structure.

Read the full 26-section primer that includes the underlying charts, methodology, and a more detailed assessment of Anthropic’s growth.

DISCLAIMER AND DISCLOSURES
Similarweb Ltd. and its affiliates (collectively “Similarweb”) provides data and analytics products. Similarweb and its affiliates does not hold securities of Anthropic PBC, which is a private company, and has not received compensation from Anthropic for investment-banking services. Anthropic and the peer companies named in this report may be customers of Similarweb’s data products.
This report does not assign a directional rating of any kind, does not produce a revenue estimate or valuation, and does not constitute an offer, solicitation or recommendation regarding any security, including any securities that may be offered in a future initial public offering. Reported revenue, run-rate, mix, customer-count and valuation figures cited in the report are company statements relayed by the press; they are not audited and have not been verified by Similarweb.
Similarweb is not a registered investment advisor or broker in any jurisdiction. All data, reports and related materials provided or made available by Similarweb (the “Materials”), are intended for informational purposes only. The Materials are based on data obtained from third parties, including estimations and extrapolations based on such data, and are not intended to be predictive of the movement of the market prices of the securities of the company(ies) referred to therein. Similarweb has no authority to provide any information or metrics, or to make any representation on behalf of the company(ies) referred to in the Materials, and the Materials have not been authorized or approved by such company(ies).
Any statements, opinions or references in the Materials to the past performance of a company, security or industry do not in any way guarantee, and is not an indication as to, future performance or results or any specific outcome. Under no circumstances are the Materials to be considered or relied upon in any manner as legal, tax or investment advice. Similarweb shall not be responsible for the accuracy of the Materials and shall have no liability for any decision by any third party based in whole or in part on the Materials.
All Materials are the property of Similarweb and may be protected by applicable copyright or other intellectual property laws and treaties. Title and full exclusive ownership rights in the Materials, the underlying data, and any and all other Similarweb intellectual property (including the Similarweb name, logo and other marks) remain solely with Similarweb (except for third party trademarks and logos, which remain the property of their respective owners). All rights not expressly granted herein are reserved.