Because the Life & Services Brand Authority Index fixes its four weights before any company is scored, and because the revenue test that governs eligibility admits every brand on this page, the result is a ranking of degree inside a nine-point band rather than a pass mark.
The index runs on four weighted tests. Global sales and brand equity carry 40%, service and digital category purity carries 30%, owned platforms and infrastructure carry 20%, and digital search interest and market reputation carry the last 10%. Eligibility is decided separately: a brand reaches the 90 band only when its own revenue clears the US$32.2 billion entry line of the Fortune Global 500, either because it stands on the register or because it publishes revenue above that line itself. All ten brands on this page clear the US$32.2 billion line, so no card was admitted by exemption and none was lifted by a parent. Amazon scores 98 on US$716.92 billion of net sales, Alphabet 97 on US$402.84 billion, Microsoft 96 on US$281.72 billion, and Visa 90 on US$40.00 billion, the smallest base here and still 24% above the line.
The highest score on this page is 98 and the lowest is 90. Eight points separate first from tenth while revenue differs by nearly eighteen times, and that compression is deliberate. A reader asking whether a brand belongs in this tier gets the same answer ten times over. A reader asking which brand leads on purity, on owned infrastructure or on search demand has to read the weights, because those are the only levers left once eligibility stops discriminating. Purity is the clearest example: Alphabet earns about 98% of revenue from services and digital products, Tencent 100%, and Amazon 58.7% from its net services line, so the 30% weight separates them even though all three clear the line.
The inheritance rule matters more here than the arithmetic suggests. A place on the register belongs to the entity that files the consolidated accounts, and it does not inherit down to a brand, a division or a licensee. ByteDance is privately held and publishes no audited filing, so it is scored on the revenue it has disclosed itself, about US$186.0 billion, not on a valuation. JPMorgan Chase is scored on US$185.60 billion of managed net revenue rather than on the US$3.9 trillion of client assets its asset and wealth management arm administers, because assets under management are not revenue. Where a figure cannot be traced to the company, the card marks it undisclosed.
Two conventions apply. Fiscal years differ: Microsoft's US$281.72 billion and Disney's US$94.43 billion are full-year figures, while Alibaba's US$142.4 billion is a fiscal year that ended in March 2025 and Visa's US$40.00 billion one that ended in September 2025. A US$ equivalent appears only where the company publishes one, which is why Tencent's RMB 751.80 billion is converted at the rate the company used and ByteDance's figure rests on its own disclosure. Disclaimer: All figures trace to audited accounts, regulatory filings and statements issued by the companies themselves. Every company is reported in the currency and the fiscal year it uses. Segment definitions belong to the companies and they differ, so a net services line, a cloud division and a single reportable segment are not the same measurement. A score attaches to the legal entity named on the card; where a parent, a controlling shareholder or a brand licensor reports turnover of its own, that money is not the operating revenue of the company being scored and is excluded here. VerityRank accepts no payment from any company for inclusion, exclusion or position.
Because the barrier on this page is not the application but the paired balance of two groups who each need the other, and assembling both sides at once has cost more than any challenger has been willing to spend.
Amazon is the clearest case. It holds more than 350 million active buyers and more than 230 million Prime members, and the sellers follow that demand: the third-party marketplace plus AWS, advertising and subscription fees produced US$420.66 billion of net service revenue, 58.7% of the US$716.92 billion total. The buyer side is defended by delivery rather than by software, which is why the company runs more than 1,200 fulfilment and sortation sites and 110 or more Amazon Air freighters. A rival can rent warehouses; it cannot rent the volume that fills them.
Visa shows the same structure in payments. About 4.5 billion credentials and more than 150 million merchant acceptance points sit on one network, and neither a cardholder nor a shop can move alone. VisaNet clears a peak above 76,000 transactions a second across four mirrored clearing sites, and the network carried US$15.5 trillion of payment volume in fiscal 2025 on US$40.00 billion of revenue. That ratio, roughly 388 dollars of volume for every dollar of revenue, is what a two-sided toll road looks like when both sides are already connected.
Messaging and social platforms defend the same ground with attention instead of throughput. Tencent's WeChat reaches 1.418 billion monthly accounts, clears more than a billion payment transactions a day and hosts mini-program trade above RMB 8 trillion a year. Meta's family of apps reaches 3.98 billion people monthly, with Facebook alone at 3.07 billion, and advertising produced US$196.18 billion, or 97.6% of revenue. ByteDance reaches more than 2.5 billion monthly users across its products, with TikTok at 1.6 billion and CapCut above 350 million.
Alphabet and Microsoft own the other end of the same idea, where the paired groups are developers and users, or enterprises and their staff. Alphabet runs more than 750 million monthly users on the Gemini app and 325 million paid consumer subscriptions, and Microsoft reports more than 420 million commercial Microsoft 365 seats alongside more than 100 million developers on GitHub. The cost of switching is not the price of the software; it is the coordination inside the customer that has to be rebuilt. That is why these positions survive price competition, and why the scores at the top of this page cluster so tightly.
The same logic applies to enterprise contracts. AWS carried a backlog of US$244 billion and Microsoft's commercial remaining performance obligation reached US$678 billion, both of them commitments signed years ahead and expensive to unwind. Alibaba's 1.35 billion annual active consumers and its 88VIP base above 50 million members do the equivalent work on the retail side, and Disney's 196 million streaming subscriptions lock in households rather than individual viewers. The measure of a two-sided network is not how many people joined but how many would have to coordinate to leave at once, and on this page that number is very large.
It turns a variable cost into a fixed one, and the fixed cost arrives a year or two before the revenue it is meant to serve, which is why capital spending now reaches the score through the infrastructure weight rather than through margin alone.
The scale is new. Amazon guided to about US$200 billion of capital expenditure for 2026 and Alphabet to between US$175 billion and US$185 billion. Both figures exceed the entire annual revenue of most companies on this page: Visa's full-year revenue is US$40.00 billion, Disney's US$94.43 billion. When a single year of construction is larger than a peer's whole business, the question stops being whether the spending is affordable and becomes whether it can be billed before depreciation starts.
The conversion test is visible in three numbers. AWS grew 20% to US$128.7 billion and carried a backlog of US$244 billion. Google Cloud reached US$58.7 billion with US$242.8 billion of contracted backlog and an operating margin above 30%. Azure passed US$75 billion at 34% growth, and Microsoft's commercial remaining performance obligation reached US$678 billion. Microsoft earned US$128.53 billion of operating income on US$281.72 billion of revenue, a margin wide enough to fund the build without new debt.
Not every company carries the same load. Meta operates 24 water-cooled hyperscale campuses but earns 97.6% of revenue from advertising, so its spending is charged against a single income stream; its Reality Labs unit contributed about US$2.5 billion, a rounding error beside US$196.18 billion of advertising revenue. ByteDance is building centres in Ireland, Norway, Malaysia and Indonesia against about US$186.0 billion of self-disclosed revenue, and Amazon booked US$2.4 billion of fourth-quarter 2025 charges, US$1.1 billion of it for an Italian antitrust and tax settlement, which shows how quickly a one-off item can dent a quarter.
The pressure lands on free cash flow before it lands on earnings. A data centre is depreciated over years, so the earnings hit is spread while the cash leaves at once, and a company that guides to US$200 billion of spending is telling the market it expects the demand to be there when the buildings open. On this page the 20% infrastructure weight is what separates the companies that own their capacity from those that rent it, and part of the gap between Amazon's 98 and Visa's 90 is the gap between a company that builds its own compute and one that builds only a payment network.
There is a second cost that never appears on the capital expenditure line. The power contracts behind these sites run for decades: Microsoft signed a 20-year agreement tied to the restart of a nuclear unit in Pennsylvania for 835 megawatts, and Amazon has contracted for small modular reactor capacity of its own. A hyperscale campus is useful only if the electricity is contracted before the servers arrive, so energy procurement has become part of the competitive position, and a company that locks in generation at a fixed price holds an advantage that shows up in margin years later.
Because a renewal is cheaper to earn than a first purchase, and the companies on this page that report large subscription bases also report the steadiest revenue, which the 40% global sales weight rewards without needing a separate membership test.
The subscription bases here are large enough to act as revenue floors. Amazon reports more than 230 million Prime members, Microsoft more than 420 million commercial Microsoft 365 seats, and Alphabet more than 325 million paid consumer subscriptions across Google One and YouTube Premium. Alibaba's 88VIP base passed 50 million members against 1.35 billion annual active consumers, and Tencent's fintech and business services unit, which holds payment and subscription-like revenue, earned RMB 229.4 billion.
Disney shows what happens when the unit economics turn. Its direct-to-consumer streaming business carries 196 million subscriptions and produced US$1.327 billion of operating profit in fiscal 2025, against US$143 million a year earlier, while the older linear television networks saw operating profit fall 14% to US$2.955 billion. One company, one content library and two opposite trajectories, decided by whether the customer relationship is direct or rented through a distributor.
Advertising is the comparison that matters. Meta earns US$196.18 billion of advertising revenue, 97.6% of its total, and Alphabet's search and related advertising alone brought US$224.53 billion. Advertising converts attention into cash at a high rate, but it moves with the advertising cycle; a subscription converts the same attention into a contracted stream that a downturn does not cancel in a single quarter. That difference is why a subscription-heavy company can hold a score in the nineties on a smaller revenue base than an advertising peer of similar reach.
The unit economics are not identical across the group. A Prime membership bundles delivery, video and music, so its value depends on logistics density; a Microsoft 365 seat bundles software the customer already runs; a Disney subscription depends on content spending that has to be renewed every year. Churn is the variable that decides which model compounds. Amazon's net service revenue of US$420.66 billion, 58.7% of total sales, includes subscription and advertising fees alongside marketplace commissions, and it is the fastest-growing part of a US$716.92 billion business. Where a subscription base is measured in hundreds of millions, small changes in renewal rates move more money than large changes in price.
Bundling is what makes that arithmetic work. Amazon folds delivery, video and music into a single Prime fee, Alphabet folds storage and ad-free viewing into one subscription, and Tencent folds payments, mini-programs and messaging into one account. Each bundle adds a reason for the customer to stay without adding a separate cost of serving that customer, which is the opposite of how advertising behaves, where every additional impression carries a marginal cost in content or inventory. That is the structural reason a membership business can defend a score on a smaller revenue base.
More than the scores alone suggest, because a remedy can remove a distribution channel that took a decade to build, and the four weights measure the business as it stands rather than the business a court may leave behind.
The live cases are concentrated in search, payments and marketplaces. The US Department of Justice won its search monopoly case against Google and has proposed remedies that include separating the Chrome browser and ending the payments that keep Google as the default search engine on other platforms. The same department has sued Visa over debit card routing, and the Federal Trade Commission's case against Amazon over self-preferencing is still being heard. None of these has produced a final remedy yet, and the page scores the companies as they are.
The charges are already visible. Amazon booked US$2.4 billion of charges in the fourth quarter of 2025, including US$1.1 billion for an Italian antitrust and tax settlement and US$0.61 billion for impairment of physical store assets. Meta settled teen-addiction claims with dozens of state attorneys general and faces a Digital Markets Act penalty in Europe over its advertising consent model. TikTok's position in the United States remains unresolved while litigation over the divest-or-ban statute continues, which is a distribution risk rather than a fine.
Financial and platform regulation carries a different shape. JPMorgan Chase raised its provision for credit losses to US$14.212 billion from US$10.678 billion and operates under Federal Reserve and Office of the Comptroller of the Currency stress tests, so its capital return is set by regulators as much as by earnings. Visa faces interchange fee caps in Europe and the United Kingdom on top of the American debit case. For a company earning US$40.00 billion of revenue, a one-basis-point change in the fees it can charge is US$4 million a year, and interchange rules move more than a basis point.
The effect on valuation is uneven. A fine is a one-time cost that a company with US$716.92 billion of revenue can absorb; a remedy that changes distribution, such as losing default placement or being required to separate a browser, is a permanent change to the cost of acquiring users. That is why the market treats the search and payments cases differently from the privacy penalties, and why the 10% reputation weight is the smallest of the four: reputation moves with headlines, while the other 90% is measured on assets that a court order can actually reach.
Timing matters as much as size. A case that takes five years to resolve leaves the business model intact while it runs, and the market prices the eventual remedy rather than the accusation; a case that ends in a structural order changes the economics permanently. That asymmetry is why the companies on this page can carry live antitrust exposure and still score in the nineties. The index measures revenue, category purity, owned assets and reputation as they stand today, and an order that has not been issued cannot be scored against anyone.
