Largest software companies in the world: the top 25 by revenue, plus the biggest SaaS companies

This list is for people who sell to software companies: cloud and infrastructure providers, developer and security tools, data vendors, recruiters, sales and marketing services, and finance and legal advisers. It ranks the 25 largest software companies by revenue from their own filings, shows the biggest subscription (SaaS) businesses separately, counts US software publishers from Census data, and explains where buying happens inside companies this size.

The 25 largest software companies by revenue

Revenue as reported in each company's latest annual report (Form 10-K, Form 20-F or, for Constellation Software and Dassault Systèmes, the published results release), checked on 30 October 2026. Fiscal years differ widely, so the year end is shown for each.

Euro figures are converted at the Federal Reserve's 2025 average rate of $1.1306 per euro (G.5A, released 5 January 2026). Constellation reports in US dollars. Headquarters are each company's principal executive office in its filing. Constellation ($11.62 billion) is just ahead of Shopify ($11.56 billion), and Synopsys ($7.05 billion) just ahead of Dassault Systèmes (about $7.05 billion at that rate).

What the table includes and doesn't:

• Microsoft's total isn't all software. It includes Azure cloud infrastructure, Xbox, Surface devices, LinkedIn and search advertising. Its Microsoft Cloud revenue (Microsoft 365 commercial cloud, Azure, Dynamics 365 and LinkedIn's commercial business) was $214.4 billion, up 27%.
• Some software giants are segments. Broadcom's infrastructure software segment, which includes VMware, reported $27.0 billion for the year ended 2 November 2025, which would rank fifth here. IBM also has a large software business inside a wider IT company.
• Close calls depend on the rate. SAP's €36.8 billion converts to about $41.6 billion at the 2025 average rate, $0.1 billion ahead of Salesforce, whose year ended a month later. A slightly weaker euro would swap them.

The biggest SaaS companies by subscription revenue

If you sell to subscription businesses, the subscription line matters more than the total. From the same filings:

Shares are our calculation from each filing. Microsoft Cloud includes Azure infrastructure, so it isn't pure SaaS. Shopify earns most of its revenue, $8.8 billion, from merchant solutions such as payments.

SAP's cloud revenue grew 23% to €21.0 billion in 2025, and 86% of it came from its Cloud ERP Suite, according to its 20-F. That shift is why SAP, once mostly a license and support business, now sits alongside the subscription companies.

How this list is built

• What's counted. Total revenue in the consolidated income statement for the latest fiscal year filed by 30 October 2026, for companies whose main business is selling software or software subscriptions.
• Who's left out. Advertising-led platforms such as Alphabet and Meta; cloud infrastructure inside wider companies (Amazon Web Services); video game publishers (see the largest gaming companies); IT services firms; software segments inside larger groups (Broadcom, IBM, Siemens); and private software companies, which don't publish results. Payments companies such as Block are left out even though the SEC classifies some as software.
• Borderline calls. Fortinet and Palo Alto Networks sell some hardware alongside security software and subscriptions; Roper Technologies describes itself as a diversified technology company whose businesses develop vertical software and technology enabled products. They're in because software and subscriptions are most of what they sell.
• Other lists. Forbes' Global 2000 software ranking blends sales, profit, assets and market value, and Fortune's list ranks by revenue within the US. We didn't copy either.

How many software companies there are in the US

County Business Patterns counted 18,425 software publisher establishments in 2023 (NAICS 513210; the 2023 data use the older code 511210), employing 1,011,187 people with $203.7 billion of annual payroll, or about $201,000 per employee. California has the most (3,890), followed by Texas (1,302), New York (1,264), Florida (908), Washington (860) and Massachusetts (756). Many software firms are classified elsewhere, under computer systems design or data processing, so this undercounts the whole industry. For mid-size names, see the list of SaaS companies, and for the code, the NAICS code for software development.

Where the buying happens at software companies

The filings show how the work is split, and each group has its own budget:

• Engineering. Microsoft has 77,000 people in product research and development; Snowflake 2,424 of its 9,060; Datadog about 3,900. Engineering leaders buy cloud capacity, developer tools, testing and security, often trialing bottom-up before a central purchase.
• Go-to-market. Microsoft has 43,000 in sales and marketing; Datadog about 3,600. Revenue, marketing and operations leaders buy data, sales tools, events and agencies, and are measured on pipeline and efficiency.
• Operations and support. Microsoft's 89,000 operations staff include support, consulting and datacenter operations. Support and customer success leaders buy service platforms and outsourcing.
• General and administrative. Finance, legal and people teams buy systems, audit and benefits, usually through procurement with security review.

General guidance: at companies this size, a security and privacy review is part of almost every purchase, and procurement joins once a business owner has picked a vendor.

Worked example: selling developer tools by the seat

Illustrative example. The vendor, its price and adoption rates are invented; engineering headcounts come from each company's 10-K.

The seller: Testwell (testwell.example) sells automated test generation at an assumed $30 per engineer per month. The question: what a realistic first deal looks like at a large software company.

1. Use the disclosed headcount. Snowflake reports 2,424 employees in research and development; Datadog about 3,900.
2. Do the ceiling arithmetic. If every Snowflake R&D employee used it: 2,424 × $30 × 12 = $872,640 a year. That's the ceiling, not the forecast.
3. Start with one team. A first deal usually covers one product group. At 10% of R&D, that's about 242 seats and $87,120 a year at Snowflake (242 × $30 × 12), or about 390 seats and $140,400 at Datadog.
4. Plan the expansion. Agree success measures with the first team (test coverage, release frequency) so the next team's budget holder can see results.

Finding the engineering leaders who own each product area, and what they've said publicly about their tooling, is the slow part; Bebop researches each account in a list and writes a playbook with who to contact and the angle to open with.

Subject: Test coverage on [product area]

Hi [first name],

I saw [company] [shipped / announced / is hiring for] [product area] recently.

Engineering leaders at companies your size often tell us test writing is what slows releases once a codebase passes a few million lines.

We generate and maintain tests automatically, and teams usually start with one service for a month. Would a 20-minute look at how that would work on [product area] be useful?

[Your name]

Common mistakes

• Ranking by market value and calling it size. It moves daily and includes advertising giants.
• Ignoring fiscal years. Microsoft's year ends in June and Salesforce's in January; compare like with like.
• Treating the company as one buyer. Engineering, go-to-market and operations buy separately.
• Skipping security review. Have your documentation ready before the first technical meeting.

Next step: see the wider technology industry in the largest tech companies.