In the 28 days ending May 18, 2026, GPTBot overtook ClaudeBot to become the third-largest AI crawler on the web — reversing the pattern that held through all of Q1 2026. That single data point reframes the ChatGPT-versus-Claude story most other stats roundups still get wrong. Combined with OpenAI's 900 million weekly active users, $25 billion annualized revenue, and 50 million paying subscribers, the May 2026 picture looks like this: ChatGPT is bigger than ever — and the infrastructure footprint behind it is finally pulling ahead of Anthropic's, after months of being behind.

Key findings (May 2026 update):

• GPTBot 11.97%, ClaudeBot 10.67% of all AI bot traffic on Cloudflare's network over the 28 days ending May 18 — GPTBot is ahead for the first time since we started tracking. Source: Cloudflare Radar bots/summary/user_agent.
• ChatGPT has been ranked #1 in the Generative AI service category every single day for at least the past 85 days (since Feb 23, 2026 — the start of our 12-week tracking window). Source: Cloudflare Radar internet-services-timeseries.
• 900 million weekly active users confirmed by OpenAI on Feb 27, 2026 — on pace for 1 billion before year-end. -- TechCrunch
• $25 billion annualized revenue as of March 2026, on a $2-billion-per-month run-rate. Targeting $29.4B for full-year 2026.
• 50 million paying subscribers worldwide — the figure most "ChatGPT Plus only" stat sheets still under-report by 5x.
• Bytespider (ByteDance) more than doubled its AI crawl share from 3.83% to 8.71% in 28 days — fastest-growing crawler on the entire AI list.
• chatgpt.com sits in Cloudflare's top-200 domain bucket at rank 25 — up from a niche-ranked domain 18 months ago, but lower than the top-10 figures circulating in older roundups.

What changed since the April 10 version of this report

I last pulled this dataset on April 10, 2026. I re-pulled every Cloudflare Radar metric on May 18 to see which numbers held, which moved, and which reversed. Several of the headline claims from April are now wrong — that's the value of dating your data and refreshing it. Updated May 18, 2026.

Metric	April 10	May 18	Δ
GPTBot crawl share	11.05%	11.97%	+0.92 pt ⬆️
ClaudeBot crawl share	11.78%	10.67%	−1.11 pt ⬇️
Bytespider crawl share	3.83%	8.71%	+4.88 pt ⬆️⬆️
Googlebot AI crawl share	31.73%	27.85%	−3.88 pt ⬇️
Meta-ExternalAgent share	16.29%	13.96%	−2.33 pt ⬇️
Crawl purpose: training	50.69%	52.54%	+1.85 pt
chatgpt.com Cloudflare rank	reported as #10	rank 25, top-200 bucket	reset
OpenAI paying subscribers	10M ChatGPT Plus only	50M paying (all tiers)	re-stated
OpenAI total raise / valuation	(not in April version)	$122B raised, $730B+ valuation	new

Three of those changes matter the most. The ClaudeBot-versus-GPTBot reversal is a structural shift in OpenAI's web-data footprint. Bytespider doubling means ByteDance is now a top-five AI crawler at a scale none of the April analyses caught. And the 50 million paying subscriber figure — confirmed in OpenAI's funding round announcement — is roughly 5x what most "ChatGPT Plus subscribers" stats articles report, because they only count one of OpenAI's four paid tiers.

ChatGPT statistics 2026: key highlights

These are the ten numbers I keep coming back to. I cross-referenced each one against Cloudflare Radar (for crawl and ranking data), OpenAI's official statements (for users and revenue), SimilarWeb (for traffic share), and Sensor Tower (for app metrics).

#	Statistic	Value	As of	Source
1	Weekly Active Users	900 million	Feb 27, 2026	OpenAI
2	Paying Subscribers (all tiers)	50 million	Q1 2026	OpenAI
3	Annualized Revenue	$25 billion	Mar 2026	OpenAI / PYMNTS
4	Monthly Revenue Run-Rate	~$2 billion / month	Q1 2026	Sacra
5	OpenAI Valuation (post-money)	$730+ billion	Apr 2026	OpenAI
6	Total Raised	$122 billion	Apr 2026	OpenAI
7	Generative AI Ranking	#1 every day since at least Feb 23, 2026	May 18, 2026	Cloudflare Radar
8	GPTBot Crawl Share	11.97% of AI bot traffic	28d to May 18	Cloudflare Radar
9	2025 Mobile Revenue	$1.35 billion (+673% YoY)	FY 2025	Sensor Tower
10	2026 Revenue Target	$29.4 billion	FY 2026	DemandSage / OpenAI

If you're evaluating large language models for enterprise use, this is the baseline. ChatGPT is one of the most-used web properties on the planet, and that scale affects everything from API reliability to model training data breadth.

ChatGPT user growth and adoption (2022–2026)

ChatGPT's adoption curve is the fastest in consumer technology history. One million users in five days. Nine hundred million weekly active users in three years. Nothing else has scaled this fast.

The growth timeline

Date	Milestone	Time to reach
Nov 30, 2022	Launch day	Day 0
Dec 5, 2022	1 million users	5 days
Jan 2023	100 million users	~2 months
End of 2023	180 million users	13 months
Feb 2025	400 million WAU	~27 months
Jul 2025	700 million WAU	~31 months
Dec 2025	800 million WAU	~36 months
Feb 27, 2026	900 million WAU (confirmed by OpenAI)	~38 months

According to TechCrunch's coverage of OpenAI's February announcement, weekly active users grew 350% in the 18 months prior. At that pace, 1 billion weekly active users is achievable before the end of 2026 — making ChatGPT the fastest consumer product in history to reach that milestone.

Who's actually using ChatGPT?

The user demographics have shifted. The gender split has flipped: at launch roughly 80% of users were male, but by 2026 about 52% are women. That signals mainstream adoption well beyond the early tech-adopter demographic.

Seventy percent of ChatGPT usage isn't work-related, according to Zapier's analysis. Close to half (49%) of all queries fall under "asking" — general knowledge questions, advice, explanations. Despite enterprise ambitions, ChatGPT's growth engine is still powered by personal use.

One pattern we see at TechnologyChecker when scanning technology stacks across millions of domains: ChatGPT integration signals (API calls, embedded widgets, OpenAI SDK references) have increased 3x in the sites we monitor since January 2025. Businesses aren't just using ChatGPT casually — they're building it into their products.

ChatGPT revenue and financial performance

OpenAI has gone from a research lab to one of the fastest-growing revenue engines in tech. Annualized revenue crossed $25 billion in March 2026, on a roughly $2 billion-per-month run-rate. The April 2026 capital raise of $122 billion valued the company at over $730 billion post-money — a number that would have been ridiculous to anyone reading this article a year ago.

Revenue growth timeline

Period	Annualized run-rate	Year-over-year growth
2023	~$2 billion	First meaningful revenue
2024	~$6 billion	~200%
End 2025	~$20 billion	~233%
March 2026	$25 billion	tripled YoY
2026 target	$29.4 billion	per OpenAI internal projections

According to PYMNTS reporting, OpenAI's ARR tripled in 2025 alone. The pace has not slowed in Q1 2026. That trajectory puts OpenAI on track to match companies like Salesforce and Adobe in total revenue within 2-3 years, though profitability remains a separate question.

Where the revenue comes from: the 50 million paid subscribers nobody talks about

This is the stat most ChatGPT articles get wrong. OpenAI confirmed 50 million paying subscribers across all tiers in its April 2026 announcement. Most "ChatGPT Plus subscribers" stats articles still report 10 million — because they're only counting one of OpenAI's four paid SKUs:

• ChatGPT Plus ($20/month) — the original consumer tier
• ChatGPT Team ($25/user/month) — multi-seat for small teams
• ChatGPT Enterprise (custom pricing) — large-org deployment
• ChatGPT Pro ($200/month) — heavy users and o1-series access

At 50 million paying users, the average monthly revenue per paying user comes out to roughly $35–$40 — which lines up with the rough $2B/month run-rate when you exclude API revenue. The mix is now substantially weighted toward higher-priced enterprise and Pro tiers, not the $20 Plus tier most stat sheets implicitly assume.

Mobile revenue: $1.35B in 2025, $221M in a single US month in March 2026

ChatGPT's mobile app revenue grew 673% year-over-year. According to WytLabs, mobile app revenue surged from $174 million in 2024 to $1.35 billion in 2025. Sensor Tower's State of Mobile 2026 report puts ChatGPT at number 2 globally by downloads, with 148% growth in downloads, 254% growth in in-app purchase revenue, and a 426% surge in total time spent.

The Sensor Tower platform-level data for March 2026 alone (US only): 23 million iOS downloads with $221 million in revenue, plus 55 million Google Play downloads with $54 million in revenue. That's a single market and a single month — and ChatGPT pulled in $275 million in mobile revenue from US users alone.

The profitability question

Despite $25 billion in annualized revenue and a $730 billion valuation, OpenAI reportedly isn't profitable yet. Training and inference costs for the GPT-4o and o1 model families consume enormous compute resources. For B2B buyers evaluating AI adoption trends, this matters: pricing stability depends on OpenAI finding a sustainable cost structure. Inference cost optimization has been improving about 10x per year, but whether subscription prices hold steady through 2026 is unclear.

If you're forecasting AI spend, budget for tier restructuring. OpenAI's introduction of the $200/month Pro tier suggests the $20 Plus tier may eventually become more limited.

Traffic, engagement, and usage metrics

ChatGPT doesn't just have users. It has some of the highest engagement metrics of any web platform — deep, habitual usage rather than casual drive-by traffic.

Monthly visit volume

Month	Monthly visits	Source
January 2026	5.72 billion	The Digital Elevator
February 2026	5.35 billion	DemandSage / SimilarWeb
Current estimate	~5.6 billion	Exploding Topics

ChatGPT receives approximately 2.5 billion prompts per day globally. That breaks down to roughly 2.78 prompts per weekly active user per day, which suggests regular multi-turn sessions rather than single-question visits.

ChatGPT usage by country

The India numbers stand out. At 198 million users, they're nearly matching the US, and the 100% growth spike after ChatGPT Go launched at $4.50/month shows what happens when you price for the market.

According to Siana Marketing, the top five countries by ChatGPT user count are:

Country	Users	Share of global traffic
United States	205 million	15–19%
India	198 million	~9.5%
Brazil	69.6 million	—
Canada	64.8 million	—
France	51.6 million	—

The US leads at 205 million users (~17% of the global total). India's user base doubled in a single month after OpenAI launched ChatGPT Go, a localized tier priced at $4.50/month — well below the $20 Plus subscription. If OpenAI replicates the ChatGPT Go model in other price-sensitive markets, the path to 1.5 billion users becomes realistic before 2028.

Conversion rates: ChatGPT traffic vs Google

This is the stat B2B marketers should care most about. According to a Seer Interactive case study, traffic referred from ChatGPT converts at 16% compared to Google Organic's 1.8%. Nearly 9x higher.

The caveat: AI traffic still accounts for less than 1% of overall organic traffic. The volume is tiny. But the intent signal is strong — people who arrive at your site through a ChatGPT recommendation are far more likely to take action.

What to do: set up UTM tracking for AI-referred traffic today. Even if the volume is small, understanding how ChatGPT users behave on your site gives you an early advantage.

What Cloudflare's May 2026 crawl data reveals about ChatGPT

At TechnologyChecker we monitor crawl patterns and bot activity across millions of domains. When I re-pulled the Cloudflare Radar data for this May 18, 2026 update, three numbers stopped me cold — and one of them outright reverses the headline finding from this article's April version.

AI bot crawl volume (28 days ending May 18, 2026)

These percentages represent each bot's share of total AI bot traffic observed by Cloudflare Radar in the 28 days ending May 18, 2026:

Bot	Share of AI bot traffic	Operator	Δ from April
Googlebot	27.85%	Google	−3.88 pt
Meta-ExternalAgent	13.96%	Meta	−2.33 pt
GPTBot	11.97%	OpenAI	+0.92 pt ⬆️
ClaudeBot	10.67%	Anthropic	−1.11 pt ⬇️
Bytespider	8.71%	ByteDance / DeepSeek	+4.88 pt ⬆️⬆️
Bingbot	8.22%	Microsoft	+0.23 pt
Applebot	7.13%	Apple	−0.02 pt
Amazonbot	4.96%	Amazon	+0.75 pt
OAI-SearchBot	1.67%	OpenAI	−0.46 pt
Other	4.86%	Mixed	—

Combined OpenAI bots reach 13.64% (GPTBot 11.97% + OAI-SearchBot 1.67%), now ahead of Anthropic's 10.67%. In April it was the other way around. This is the structural reversal I keep coming back to. OpenAI's crawl footprint pulled ahead of Anthropic's in just under a month.

The numbers above are a frozen 28-day window. The chart below is live — Cloudflare Radar updates it daily, so if you're reading this weeks after publication you'll see whether the GPTBot/ClaudeBot lead held.

GPTBot vs ClaudeBot vs Bytespider — the crossover story

I've been watching these three bots trade positions daily for 12 weeks. The chart above plots their daily share through the entire February 23 – May 18 window.

Three patterns to notice:

1. GPTBot and ClaudeBot were roughly tied through Q1 2026, swapping the lead from week to week.
2. In the week of May 9, the gap blew open. GPTBot surged to 18.71% of AI bot traffic on May 9 alone, while ClaudeBot dropped to 7.82% the same week. That's a 10.89-point gap that emerged in seven days. The most likely explanation: OpenAI ran a large training crawl burst that week, possibly tied to a new model checkpoint.
3. Bytespider tripled. ByteDance's crawler went from 3.62% on Feb 23 to a peak of 11.78% on May 8. The most likely explanation: DeepSeek (which shares the Bytespider infrastructure) is scaling up its training pipeline aggressively to catch the leaders.

I'm not fully confident the May 9 GPTBot spike represents a sustained shift — it could be a one-time crawl burst. But the underlying 28-day rolling average has GPTBot ahead of ClaudeBot whether or not you exclude the surge week. The trend was already there before May 9 amplified it.

What AI crawlers actually do

Training is the single largest purpose of AI crawling in May 2026 at 52.54% of requests — up from 50.69% in April. Mixed-purpose crawls (models that index for both training and retrieval) account for another 36.06%. Only 8.41% of AI crawler traffic is classified as search — real-time web retrieval on behalf of a user prompt — and that's actually up from 7.28% in April, which suggests AI assistant search products (ChatGPT Search, Perplexity, Claude) are growing their share of the crawl mix.

User-action crawls (browser-using agents like ChatGPT's Operator or Anthropic's computer use) remain small at 2.36% but continue to be the fastest-growing crawl purpose category.

The crawl-to-referral problem

This is where the data gets uncomfortable for publishers. Cloudflare Radar tracks how much each AI platform crawls versus how much referral traffic it sends back. As of the April 13–20, 2026 measurement window:

Platform	Crawl-to-referral ratio	What it means
DuckDuckGo	1.5:1	Sends traffic back frequently
Google	5:1	Established search model
Microsoft	29:1	Moderate return
Mistral	53:1	Low return
Perplexity	95:1	Modest return
OpenAI	1,252:1	Very low return
Anthropic	13,528:1	Lowest return by far

Anthropic crawls 13,528 pages for every 1 referral it sends back. OpenAI's 1,252:1 ratio is bad; Anthropic's is roughly 10x worse. Perplexity's 95:1 looks healthy by comparison and is the highest crawl-to-referral conversion among the AI-first platforms.

OpenAI's ratio has actually improved over time: it was 785:1 in October 2025, peaked at 1,851:1 in December 2025 as crawl volume spiked, then dropped to roughly 1,252:1 in mid-April 2026. That improvement correlates with OAI-SearchBot generating more outbound clicks as ChatGPT Search gains users.

Domain ranking context: where chatgpt.com actually sits

I pulled the Cloudflare Radar domain rank for chatgpt.com on May 18, 2026. It returned rank 25, in the top-200 bucket. That's a high-traffic ranking — chatgpt.com is one of the busiest properties on the Cloudflare-visible internet — but it's not the "#10 globally, ahead of Amazon" figure that circulated in late 2025 statistics roundups. The methodology behind Cloudflare's domain ranking changed in early 2026, and chatgpt.com's bucket placement reflects current top-200 status, not its earlier reported top-20 position.

What it means for publishers: check your site's robots.txt and server logs. Our robots.txt AI crawlers blocking report breaks down exactly which bots to allow and which to restrict. Blocking a bot with a 13,528:1 ratio costs you almost nothing in referral traffic.

ChatGPT vs Claude vs Gemini vs Perplexity vs DeepSeek — May 2026 market share

The Cloudflare Radar generative AI service category provides a clean ranked view of which AI services attract the most traffic on the Cloudflare network. The ranking is updated daily. Here's the picture as of May 18, 2026.

Generative AI service rankings (May 18, 2026)

Rank	Service	Operator	Notable
1	ChatGPT	OpenAI	#1 every day since at least Feb 23, 2026
2	Claude	Anthropic	Locked at #2 every day in the 12-week window
3	Perplexity	Perplexity	Locked at #3 every day in the 12-week window
4	DeepSeek	DeepSeek	Above Gemini since late April
5	Google Gemini	Google	Slipped from #4 in some weeks
6	Grok / xAI	xAI	Stable in the 6–7 band
7	Character.AI	Character.AI	Recovered from #9 in April
8	GitHub Copilot	Microsoft	Stable
9	Doubao	ByteDance	NEW entrant in May 2026
10	Suno AI	Suno	AI music platform

ChatGPT has held the #1 ranking for at least 85 consecutive days through May 18, 2026 — and the trend likely extends well beyond the start of our 12-week window. Claude and Perplexity have held #2 and #3 for the same window without a single position change. The interesting movement is below that: DeepSeek has stabilized at #4 above Google Gemini, and Doubao (ByteDance) appeared in the top 10 for the first time in May.

The SimilarWeb web-traffic decline you should actually pay attention to

According to SimilarWeb's March 2026 data, ChatGPT's share of generative AI web traffic fell from 77.43% in early 2025 to 56.72% by March 2026 — a 20-point drop. Gemini surged from 6% to 25.46% over the same window. Claude grew from 1.4% to 6.02%.

That's the second-largest market share rebalancing in any consumer technology category in the last decade, behind only the iOS/Android share war of 2009-2012. But the rebalancing is uneven by surface:

• On web traffic (SimilarWeb): ChatGPT lost 20+ points to Gemini, mostly. The Gemini surge is real.
• On the Cloudflare network ranking (live data): ChatGPT held #1 every day, Gemini slipped to #5. The two datasets are measuring different things.
• On enterprise win rates: Claude reportedly wins ~70% of head-to-head enterprise deals against OpenAI, per SaaStr's coverage.

I'm cautious about the SimilarWeb decline framing because Gemini is pre-installed on Android and Workspace, which inflates "web traffic" share for users who didn't actively choose it. The Cloudflare ranking, which reflects what users actively go to, tells a different story.

The dual-platform reality

Here's the stat that reframes the whole competition narrative. According to AI Business Weekly, 79% of OpenAI users also pay for Anthropic. That confirmed what I've been seeing in our stack data at TechnologyChecker. Enterprises aren't picking one AI vendor — they're running ChatGPT and Claude side by side.

eMarketer reports that Claude users spend 34.7 minutes per daily session — the highest engagement of any AI platform. That session depth, combined with Claude's enterprise win rate, explains why Anthropic can charge premium prices with a smaller user base.

For practical purposes: if you're integrating with one AI platform, ChatGPT's user base offers the largest addressable audience. But Claude's growth trajectory and enterprise win rate make dual-platform support increasingly important.

Head-to-head: ChatGPT vs Claude for developers

Based on benchmark data and developer feedback across forums, here's where each platform stands for technical use cases as of May 2026:

Dimension	ChatGPT (GPT-4o / o1)	Claude (3.5 Sonnet / Opus)
Code generation	Strong across languages	Stronger for long-context code
Reasoning (math/logic)	o1 excels	Competitive, slightly behind o1
Writing quality	Good, formulaic at times	More natural, preferred by writers
Context window	128K tokens	200K tokens (advantage)
API pricing	$5–15 per 1M tokens	$3–15 per 1M tokens
User base	900M WAU	Growing rapidly (#2 in Gen AI)
Web crawl footprint (May 2026)	13.64% combined	10.67%

For B2B SaaS teams: build for ChatGPT's user base, test against Claude's quality, and monitor DeepSeek and Doubao for cost-optimized alternatives.

Enterprise adoption and industry use cases

Consumer usage drives headlines. Workplace integration drives revenue. Here's what the enterprise data looks like as of May 2026.

According to Christian & Timbers, ChatGPT has been adopted by 93% of Fortune 500 companies in some capacity. That figure includes everything from individual employee use to department-wide deployments to formal ChatGPT Enterprise contracts.

The 7 million workplace seats number tells the real story — that's 9x year-over-year growth in enterprise seats, way faster than the consumer growth rate. OpenAI now counts 1.5 million business clients globally, with over 1 million of those outside the US. Enterprise adoption grew 340% year-over-year by OpenAI's own accounting.

PwC deployed 100,000+ ChatGPT Enterprise seats. Morgan Stanley and Coca-Cola are also named enterprise adopters. These aren't pilot programs anymore — they're organization-wide rollouts.

The top enterprise use cases from OpenAI's published workplace adoption study: data analysis and reporting (60% of enterprise users), content generation including text and images (51%), personalization and customer experience (42%), code generation and debugging (~35% among developer teams), and internal knowledge management (~28%).

Industry-specific adoption patterns

Based on TechnologyChecker's technology detection across 50M+ domains, the top adopting sectors break down as follows: travel and hospitality leads at 18%, followed by retail and CPG at 16%, IT at 14%, health and lifestyle at 13%, and food at 13%. That travel/hospitality number didn't surprise me — those teams deal with massive volumes of customer queries and content translation, exactly the tasks where ChatGPT saves the most time.

Software and technology leads the IT portion by a wide margin. Nearly every SaaS product we scan shows some form of AI integration — OpenAI API calls, embedded ChatGPT-powered features, or AI-assisted documentation. See our top technologies used in software development for the broader stack.

The SaaS integration wave

According to SQ Magazine, 42% of new SaaS platforms with AI capabilities launched in 2025 rely on OpenAI models. That includes GPT-4o, the Assistants API, and the embeddings API. When nearly half of all new AI-enabled products choose your API, you've become infrastructure.

Developer and API adoption

OpenAI now has 3 million active developer accounts, a 5x increase since 2022. According to Index.dev, 1.3 million US developers alone have created projects on OpenAI's platform.

At TechnologyChecker, OpenAI API signatures show up in about 3x more sites than they did in January 2025. The 3 million developer figure matches what we're seeing in the wild — that's roughly the same number of developers who use Stripe's API. OpenAI has reached "developer infrastructure" scale in three years, which took most payment and cloud platforms a decade.

Key takeaways for AI leaders

Three patterns from the May 2026 data matter most for 2026 planning.

1. The OpenAI crawl footprint just pulled ahead of Anthropic's. GPTBot 11.97% vs ClaudeBot 10.67% in the 28 days ending May 18 is a structural shift, not noise. Combined OpenAI bots reach 13.64% of all AI bot traffic. This matters because crawl share is the leading indicator of which models will have the freshest, most comprehensive training data — and therefore the most reliable answers — six months from now.

2. ChatGPT's dominance is real but more contested than the headline numbers suggest. ChatGPT held #1 in the Cloudflare GenAI category every single day for 85+ days, and has 900M weekly users. But SimilarWeb shows ChatGPT's web traffic share fell from 77% to 57% in 12 months. The two datasets measure different things — Cloudflare reflects what users actively visit, SimilarWeb captures broader web behavior including Android-default Gemini sessions. Pay attention to which dataset matches your decision context. For consumer engagement, the Cloudflare ranking matters more.

3. Mobile and enterprise are the growth vectors to watch. The 673% growth in mobile app revenue, 93% Fortune 500 adoption, and 7 million workplace seats (9x YoY growth) signal that ChatGPT's next phase of growth won't come from more consumer users — it'll come from deeper enterprise integration, geographic expansion through localized pricing like ChatGPT Go, and developer ecosystem lock-in through the 3 million active API accounts.

For sales and GTM teams specifically: understanding which technologies your prospects use — and how they're integrating AI tools — gives you a genuine competitive edge. At TechnologyChecker, we track these adoption signals across 40,000+ technologies and 50M+ domains, turning raw detection data into prospect intelligence that connects directly to buying intent.

If you want to see which companies in your target market are actively adopting AI tools, start a technology lookup to identify prospects whose technology stack signals readiness for your product.

Methodology and data sources

I pulled these statistics on May 18, 2026 from 25+ sources including:

• Cloudflare Radar for AI bot crawl share, crawl purpose breakdown, generative AI service rankings, and chatgpt.com domain ranking (live data, refreshed for this May 18 update)
• OpenAI's official announcements for weekly active users (900M, Feb 27, 2026), paying subscribers (50M), annualized revenue ($25B), and the $122B capital raise / $730B valuation
• TechCrunch, 9to5Mac, eWEEK for coverage of OpenAI's user milestone announcements
• PYMNTS, Sacra for revenue and ARR breakdowns
• SimilarWeb for web traffic share (ChatGPT 56.72%, Gemini 25.46%, Claude 6.02% as of March 2026)
• Sensor Tower for mobile app downloads and revenue (including the 2025 $1.35B mobile revenue figure and March 2026 US platform-level data)
• Christian & Timbers, OpenAI's workplace study, SQ Magazine, Index.dev for enterprise and developer adoption

How I verified the data: every Cloudflare Radar number was pulled live via the Radar API on May 18, 2026 — no third-party summaries. External claims (user counts, revenue, market share) were cross-referenced against at least two independent sources before being included. Where sources disagreed, I noted both figures and explained the methodological difference (see, for example, the Cloudflare vs SimilarWeb framing of market share).

Limitations: Cloudflare Radar data reflects traffic through Cloudflare's infrastructure, which while extensive, doesn't capture 100% of global web traffic. User count estimates from non-OpenAI sources use different methodologies and may not be directly comparable. Revenue figures for OpenAI are based on reported annualized run-rates, not audited financials. Market share percentages differ significantly across sources depending on whether they measure unique users, session volume, or app downloads — I've explicitly noted which surface each figure measures.

Last updated: May 18, 2026. I'll refresh this article as new Cloudflare Radar data comes in. The next planned update is mid-June 2026.

---

About the author: Emma Davies is a Data Analyst at TechnologyChecker with 5 years of experience in technology data analysis and prospect intelligence. She holds a BSc in Data Science from Cardiff University along with Google Data Analytics and Tableau certifications. Her work involves analyzing technology adoption patterns across 10M+ domains to identify trends and buying signals for sales teams.