Google vs Microsoft Market Share 2026: Google Owns the Web, Microsoft Owns the Workplace

Google vs Microsoft market share in 2026 splits by layer, not by a single winner. Google takes 88.55% of search referrals against Bing's 3.53% and leads the browser 10-to-1. Microsoft leads three layers: business email by 117,342 domains, cloud platforms by 1.71x, and the operating system layer by 0.013 points.

That split is the whole finding. Google owns the layers a consumer touches: the search box, the browser window, the ad slot. Microsoft owns the layers an IT department buys: the mailbox, the cloud subscription, the desktop. Ask "who is bigger" and you get a useless answer. Ask "who is bigger at this layer" and every answer changes.

I'm a data analyst at TechnologyChecker, and I spend most of my week inside technographic detection data, working out which products are running on which domains right now. For this post we measured seven contested layers on two instruments we control: the Cloudflare Radar API across full-month August 2026 (with August 2025 as a control), and our own detection corpus of roughly 29.9 million live domains. Nothing here comes from an earnings call, a survey panel, or an analyst firm. We didn't read any.

Google vs Microsoft market share in 2026: the seven-layer scoreboard

Seven layers put a Google product directly against a Microsoft product on the same instrument. Google wins four of them, Microsoft wins three, and two of Microsoft's three are close enough that they only surface if you actually measure them. Here is the full scoreboard, with the measurement basis stated for every row.

Read the right-hand column before you read anything else. Three different denominators are stacked in that table and they do not blend. Percentage of referrals, percentage of HTTP requests, and count of domains answer three different questions, and averaging them would produce a number that means nothing.

Two layers where there is no contest at all

Two more layers belong on the record but not in the scoreboard, because Microsoft either fields nothing or fields something from a different category.

Tag management is a walkover. TechnologyChecker detected Google Tag Manager on 10,584,793 domains as of 2026-08-04, which is 98.51% of the tag management category. Microsoft ships no competing tag manager. That 98.51% is the single most lopsided figure we hold on either company.

Analytics is product versus product, not category versus category. We detected Google Analytics on 10,942,935 domains (2026-09-04) and Microsoft Clarity on 862,512 (2026-08-27), a 12.7x gap. Those two sit in different categories (web analytics and heatmaps/session recording), so this is a comparison of installed footprints, and it is not a share of one market. Clarity leads its own category outright.

Who wins search referrals: Google 88.55% vs Bing 3.53%

Google sent 88.55% of the search referrals Cloudflare Radar observed in August 2026. Bing sent 3.53%. That's a 25.1x gap, and it is the least surprising row on the scoreboard. If you want the full monthly series, the vertical splits, and where AI assistants sit, our search engine market share report is the deeper source and gets updated every month. This post treats search as one row of seven.

Two limits on that number, both of which matter more than the number itself.

First, google.* is a wildcard. It folds Gemini referrals in with Search, and Radar gives no way to split them, so the 88.55% is Google-the-referrer rather than Google-Search-the-product.

Second, we cannot cut this by country. Radar's referrer endpoint silently ignores a location= filter: Germany, Russia and an unfiltered global pull return byte-identical numbers. Any per-country search share you see quoted from this endpoint is not real.

For contrast, Digital Applied puts Bing's structural ceiling at 5-6% globally "without a Windows-level distribution event." Our referral read of 3.53% sits below even that ceiling. Microsoft owns the desktop and still cannot convert it into search referrals.

Chrome vs Edge: the browser layer is a 10-to-1 blowout

Chrome took 70.50% of HTTP requests Cloudflare observed in August 2026 against Edge's 6.78%, a 10.4x gap and the second-widest margin on the scoreboard.

The Safari number is the one I'd point at if you want to sanity-check everything else on this page.

Chrome and Edge diverge by around a point each, which is what you'd expect when one instrument weights by page views on panel sites and the other weights by requests to Cloudflare-fronted origins. The agreement on Safari says the underlying populations are close. Our standalone Chrome market share breakdown carries the longer series.

Windows vs Android: the one layer Google lost by 0.01 points

The number one operating system flipped this year, and the margin is almost nothing. In August 2025, Android held 35.02% of HTTP requests against Windows at 33.21%, an Android lead of 1.81 points. In August 2026, Windows holds 34.63% and Android holds 34.14%. Add ChromeOS at 0.48% and Google's platform total reaches 34.62%, which still loses to Windows by 0.013 percentage points.

That is a dead heat, not a rout. A tenth of a point of monthly drift would put Android back on top, and the ordering has already flipped once in twelve months.

Unlike the search endpoint, this one does accept a country filter, and the global tie hides six completely different national markets.

Android beats Windows by 39.91 points in India and loses to it by 17.62 points in the United States. The global dead heat is an artifact of averaging those two worlds together. One caveat worth stating plainly: this is share of HTTP requests, not devices shipped and not installed base. A desktop that browses all day counts more than a phone that sits in a pocket.

Microsoft 365 vs Google Workspace: who actually owns business email?

Microsoft 365 leads Google Workspace by 117,342 domains. In our 2026-08-10 rollup we detected Microsoft 365 on 3,748,862 domains and Google Workspace on 3,631,520, which puts Microsoft first in the webmail category at 45.33% and Google second at 43.91%, a gap of 1.42 percentage points. Workspace's 3,631,520 domains map to 2,649,059 distinct companies, so the count is domains, never seats, accounts or customers.

This is the section worth reading twice, because the published answers to this exact question do not agree with each other, and one of them does not agree with itself.

Medha Cloud and Onix both say "by domains" and both put Google ahead by roughly five points. We measure domains for a living and get the opposite ordering. Here's why a detection count and an adoption survey can disagree that hard.

A detection count asks one question: on this domain, right now, do the DNS and header signals resolve to this vendor's mail infrastructure? Every domain counts once and counts equally, so a one-page company site weighs the same as a 40,000-seat enterprise. A self-reported adoption survey asks a different question of a different population (usually SaaS spend, procurement records, or a customer panel) where a large buyer can carry far more weight than a small one. Neither instrument is wrong. They're counting different things, and only one of them tells you what's actually resolving on the public internet today.

The reconciliation sits in one more figure from Medha Cloud: 75% of Fortune 500 companies run Microsoft 365 and 42% run Google Workspace. Those add to 117%, because plenty of large organisations run both. Productivity-suite shares don't sum cleanly, which is exactly why four sources can each pick a defensible-looking denominator and land somewhere different.

Our own gap is small and I'd treat it as such. A 1.42-point lead in the webmail category is a real lead, but it isn't the 80% Brownstone gestures at, and it isn't the five-point Google lead Onix publishes. Microsoft is ahead on live domains, narrowly, as of 2026-08-10.

Azure vs Google Cloud: Microsoft leads by 1.7x on live domains

Microsoft Azure runs on 1.71x as many domains as Google Cloud. We detected Microsoft Azure on 175,651 domains (analysed 2026-08-24) and Google Cloud on 102,535 (analysed 2026-08-25). Azure holds 7.5% of the PaaS category and ranks fourth in it; AWS-scale infrastructure and self-hosted stacks take the rest.

I won't quote a category share for Google Cloud, and the reason is worth stating rather than hiding. Its share and category rank are null in our data, and deriving a share by summing every domain count under the PaaS category would double-count domains that run two platforms at once. A wrong denominator is worse than no denominator, so the raw counts are all you get here.

The spend view and the domain view agreeing on the ordering is the useful part. CRN has Google Cloud climbing from 13% of infrastructure spend in Q2 2025 to 15% in Q2 2026, so Google is closing the revenue gap faster than it is closing the footprint gap. Our own Google Cloud statistics report carries the deeper read, including the Flexera finding that only 19% of cloud decision makers run significant workloads on Google Cloud against 46% for Azure.

One number I'll flag but won't over-read: we've stopped detecting Google Cloud on 9,860 domains where we previously saw it, against 3,873 for Azure. That gap looks dramatic and I don't treat it as migration. A dropped detection can mean a rule change, a CDN in front of the origin, or a dead site, and separating those from a genuine platform switch takes a corridor analysis we haven't run on this pair.

Who crawls the web: Google's crawler share fell 10 points in a year

Google's share of bot traffic fell 10.00 points in twelve months, from 36.58% to 26.58%. Microsoft fell too, from 7.39% to 5.09%. The layer both companies used to split with each other is now being taken by AI labs, and Anthropic arrived at number three from outside the top ten entirely.

Google still crawls 5.2x as much as Microsoft, so the ordering between the two hasn't changed. What changed is that the pair went from 43.97% of all crawling to 31.67% in a single year. Meta now crawls 3.7x as much as Microsoft, and Anthropic went from unranked to third.

If you're deciding which bots to allow through, the crawler mix matters more than the search share does, and our AI crawler statistics report breaks the AI operators out by user agent and stated purpose.

Which company takes more than it gives back? The crawl-to-refer ratio

Google crawls 5.08 pages for every visit it refers back. Microsoft crawls 37.09. The crawl-to-refer ratio is pages crawled per single visit sent, so lower is better for the site being crawled. Google runs the lowest ratio of any operator at its scale, and the second-lowest on the whole list behind DuckDuckGo. Microsoft takes 7.3x more per visit returned.

These are ratios, not percentages, and they run on a shorter window than the rest of this post: 24 to 31 August 2026, because longer windows error out on that endpoint.

The ranking layer confirms the same asymmetry. In Cloudflare's top domains list for 2026-09-01, google.com ranks first, googleapis.com second, gstatic.com fourth and googlevideo.com tenth, with microsoft.com at sixth. Google occupies four of the top ten slots; Microsoft occupies one.

What this means if you are choosing a stack

Pick by layer, not by brand. Nothing in this data supports "standardise on one vendor" as a strategy, because the two companies are not close to each other on most layers and the layers where they are close are the ones with the highest switching cost.

• Reaching consumers? Google is the distribution channel and it isn't close. 88.55% of referrals, 70.50% of browser requests, 98.51% of tag management. Budget your web presence around Chrome behaviour, not around Edge.
• Building for a specific country? Check the platform mix first. Android carries 63.82% of Indian requests and 18.51% of American ones. A mobile-first Android assumption is correct in India and wrong in the United States.
• Choosing a productivity suite? The measured gap is 1.42 points. Treat it as a tie and decide on identity, compliance and what your users already know, because footprint gives you no signal at this margin.
• Choosing a cloud platform? Azure leads on live domains 1.71x and on spend share by roughly seven points. Google Cloud is gaining spend share faster. Both statements are true and they answer different questions.
• Deciding which bots to allow? Google returns a visit every 5.08 pages, Microsoft every 37.09. That's the number to weigh, not the search share.

The honest summary: Google owns the surface a customer sees, Microsoft owns the seat an employee sits in, and the only place they genuinely collide is the desktop, where the lead changed hands this year by 0.013 points.

Frequently asked questions

Is Microsoft bigger than Google?

Neither is bigger overall. It depends entirely on the layer. Across the seven layers we measured in August 2026, Google leads four (search referrals, browser, crawling, advertising network) and Microsoft leads three (operating system, business email, cloud platform). Google's leads run from 25.1x down to 1.97x. Microsoft's run from 1.71x down to 0.013 percentage points, so Google's narrowest win is wider than Microsoft's widest.

Which is more popular, Google or Microsoft?

Google, on consumer-facing measures. Google products took 88.55% of search referrals and 70.50% of browser requests in August 2026, against Microsoft's 3.53% and 6.78%. Microsoft is more widely deployed inside businesses: we detect Microsoft 365 on 3,748,862 domains versus 3,631,520 for Google Workspace, and Azure on 1.71x as many domains as Google Cloud.

What is Google Workspace vs Microsoft 365 market share in 2026?

Microsoft 365 leads. In our 2026-08-10 detection rollup, Microsoft 365 appeared on 3,748,862 domains (45.33% of the webmail category, ranking first) and Google Workspace on 3,631,520 domains (43.91%, ranking second). The gap is 117,342 domains, or 1.42 percentage points. These are counts of domains where the service resolves, not seat counts or revenue.

What is Google's market share worldwide in 2026?

It varies by layer and there is no single worldwide figure. In August 2026 Google took 88.55% of search referrals, 70.50% of browser requests, 26.58% of crawler traffic and 34.62% of operating-system requests including ChromeOS. Our monthly search engine market share report tracks the referral figure in detail.

Who is winning cloud market share in 2026, Google Cloud or Microsoft Azure?

Azure, on both instruments we can check. We detected Azure on 175,651 domains (2026-08-24) against Google Cloud's 102,535 (2026-08-25), a 1.71x lead. On infrastructure spend, published estimates put Azure near 21% and Google Cloud near 14%. Google Cloud is growing its spend share faster, moving from 13% to 15% between Q2 2025 and Q2 2026.

Should enterprises choose Google Workspace or Microsoft 365?

Footprint won't decide it for you. The two are 1.42 percentage points apart on detected domains, which is inside the range where either choice is mainstream and well-supported. Decide on identity provider, compliance requirements, and existing user familiarity. Note that 75% of Fortune 500 companies run Microsoft 365 and 42% run Google Workspace, so running both is common rather than unusual.

Does Microsoft still dominate Google in enterprise productivity?

Microsoft leads on domains but does NOT dominate — our 2026-08-10 measurement puts Microsoft 365 at 45.33% against Google Workspace's 43.91%, a 1.42-point gap. Published claims of an 80%+ Microsoft share and of a five-point Google lead both circulate, and they cannot both be right. Every productivity-suite figure needs its denominator and its date attached before you can use it.

Is Windows now the most-used operating system?

Yes, narrowly, and it wasn't a year ago. Windows took 34.63% of HTTP requests in August 2026 against Android's 34.14%; in August 2025 Android led 35.02% to 33.21%. Adding ChromeOS, Google's platform total reaches 34.62%, still 0.013 points behind Windows. That's a dead heat, and it's share of requests rather than devices sold.

Why do your numbers differ from StatCounter?

Different instruments measuring different populations. StatCounter counts tracker pixels across a partner panel of sites. Cloudflare Radar counts HTTP requests reaching Cloudflare-proxied origins. On browsers the two agree closely: our Safari read of 15.82% sits 0.01 points from StatCounter's 15.83%. On search they diverge more, because a referral header on a proxied origin and a panel page view are not the same measurement.

What is the Google vs Microsoft market share prediction for 2026?

We don't publish predictions, and I'd be sceptical of anyone who does at these margins. What we can say is what moved: Google's crawler share fell 10.00 points in twelve months, Windows overtook Android by 0.013 points after trailing by 1.81 points a year earlier, and Anthropic entered the crawler top three from nowhere. Two of those three would have failed a 2025 forecast.

Who actually owns the web in 2026, Google or Microsoft?

Google owns the public web, Microsoft owns the corporate one. Google touches almost every consumer path onto a website: 88.55% of referrals, 70.50% of browser requests, 98.51% of tag management installs, and three of Cloudflare's top ten global domains. Microsoft touches the infrastructure behind the business: more mail domains, more cloud domains, and the leading desktop operating system.

What changed in this report

Methodology and source data

Every figure in this post comes from one of two instruments, both pulled on 7 September 2026. We did not read earnings calls, regulatory filings, or reports from Canalys, IDC or Gartner.

Cloudflare Radar REST API. Full calendar month 2026-08-01 to 2026-08-31, with 2025-08-01 to 2025-08-31 as the year-over-year control. Endpoints used, per layer:

• Search referrals: bots/crawlers/summary/referer, PERCENTAGE normalization. The vertical= filter works on this endpoint; location= is silently ignored, so no per-country search share is quoted anywhere on this page.
• Browsers: http/summary/browser_family, PERCENTAGE normalization. This is the official grouping. The deprecated http/top/browser endpoint splits Chrome across three user-agent families and would understate it.
• Operating systems: http/summary/os, PERCENTAGE normalization, global and with location= set to US, IN, DE, BR, GB and JP. This endpoint does honour the country filter.
• Crawling: bots/summary/bot_operator, PERCENTAGE normalization.
• Crawl-to-refer: bots/crawlers/summary/crawl_refer_ratio, RATIO normalization, 2026-08-24 to 2026-08-31. Longer windows error on this endpoint, so this is the only seven-day window on the page and it is never compared against a monthly figure.
• Domain ranking: ranking/top, dated 2026-09-01.

Radar's sample covers Cloudflare's global network and tilts toward the sites Cloudflare fronts. It measures requests and referrals, not installed devices or licensed seats. The google.* referrer is a wildcard that folds Gemini in with Search and cannot be split. The August 2026 versus August 2025 referral comparison spans a February 2026 Cloudflare-side reclassification of TikTok link-outs, which is why this post quotes the level rather than the change.

TechnologyChecker detection corpus. Roughly 29.9 million live domains, crawled continuously. customerCount is a monthly rollup of the domains on which we detected a technology — not customers, not accounts, not seats. Each figure carries its own analysis date and they differ, so no single date applies to the whole set:

Three constraints we held to. Google Cloud's category share and rank are null in our data, so only its raw domain count appears here; deriving a share by summing domain counts across a category double-counts domains running more than one platform. Google Analytics and GA4 are the same product family and are never added together. There is no Microsoft 365 technology profile on this site, so its 3,748,862 figure is quoted from the competitor block of the Google Workspace profile and carries no link.

External figures are attributed and linked at first mention: StatCounter for browser share, Cledara, Medha Cloud, Onix and Brownstone Research for productivity-suite claims, INDmoney and CRN for cloud spend, and Digital Applied for Bing's ceiling. Where those sources contradict each other, the contradiction is shown rather than resolved. Cloudflare's own Radar API documentation covers the endpoints and normalization modes listed above.