Zapier Statistics 2026: 56 Verified Numbers on Customers, Revenue, Pricing and AI

I searched "Zapier statistics 2026" and read the first eight results. They carry the same numbers in a different order.

Two sources are doing all the work behind them: Sacra's revenue model and Zapier's press page. Everything after that is paraphrase, rounded slightly differently each time, with the dates quietly dropped.

That matters the moment you put a figure in a deck. A page telling you Zapier powers 3.4 million businesses, with no date and no link, has told you nothing you can check.

So this page runs the other way round. Every figure carries its publisher, its date and a link to the document it came from. Where a number is a model rather than a measurement, the sentence says so. Where a number is old, it says how old.

Fifty-six figures cleared that bar. Eleven that circulate widely did not, and they are listed near the end with the reason each failed. Each figure has its own anchor, so you can deep-link the exact number you are quoting, and the whole set is a CSV under a CC BY 4.0 licence.

One disclosure before the numbers. I build a competing product, so every judgement on this page comes from someone with a stake in the category. The figures are sourced and dated, which means you can take those and ignore the judgements.

Platform scale and catalogue

Every scale number for Zapier comes from Zapier. No filing, no auditor, no independent count of the customer base. Read this section as the company describing itself.

The press kit is the current source of record. Read on 15 September 2026, it claims 9,000+ apps in the catalogue, 3.4 million+ businesses powered, 800+ people on staff and a remote workforce spread across 40 countries (Zapier press kit, page undated). None of those four carries a date on the page itself, which is the first thing every round-up drops.

The catalogue is more useful as a trajectory than as a total, because the trajectory has dates on it. At the Zapier Central launch on 6 March 2024 the release said 6,000+ apps (PR Newswire). By the MCP beta on 4 April 2025 it was 30,000+ actions across nearly 8,000 apps (Zapier Community). The April 2026 boilerplate says more than 9,000. That is roughly a thousand new apps a year, steady across two years.

Two platform-volume figures sit on Zapier's Zappy Awards page: more than 250 million AI tasks a year and, for the platform as a whole, billions of tasks automated annually with no exact number attached (Zapier blog). That page carries no date and no methodology. Both are here because nothing else sizes platform volume, and both are flagged because an undated number ages invisibly on whoever quotes it next.

The newest self-description on the record is the boilerplate under Zapier's 23 April 2026 release: operating since 2012, more than 9,000 apps, a customer base described only as "millions", and customers from startups to the Fortune 500 (Business Wire release, Yahoo Finance mirror). Note that the boilerplate stops sizing the base where the press kit puts a number on it.

The oldest catalogue datapoint is also the weakest. An undated Zapier post says the first version integrated 34 apps, and the catalogue passed 5,000 roughly a decade later (Zapier blog). It predates the verification window, and it is the one figure here I would double-check before quoting.

Money: revenue, valuation, funding

Zapier is private, has never raised a priced primary round beyond seed, and does not report revenue. Every revenue number attributed to it anywhere traces back to one place: Sacra's model.

Here is that model, labelled as what it is. Sacra puts Zapier at $310M ARR at the end of 2023, $230M in 2022 and $165M in 2021, which gives 34.8% year-over-year growth for 2023 (Sacra, retrieved 15 September 2026). Sacra models these figures; Zapier has never disclosed them. The growth rate is Sacra's arithmetic on Sacra's own estimates.

The newest traceable Zapier revenue figure is nearly two years old.

Anything presented as 2025 or 2026 Zapier revenue is an extrapolation of a 2023 estimate. The honest answer to what Zapier makes today is that nobody outside the company knows.

The valuation is firmer because a real transaction set it. Forbes reported a $5B valuation set in January 2021 by a secondary sale, in which Sequoia and Steadfast bought stock from early investors (Forbes, 8 March 2021). No new money went into the business, and that price has not been reset by a later round.

The funding history fits in a sentence. Sacra records a $1.3M seed in 2012 and $1.4M raised in total, from Y Combinator and Sequoia, which is why the capital-efficiency framing follows Zapier around: against the 2021 estimate the secondary priced it at 30.3x revenue, on a path Sacra describes as $100M ARR in just under ten years at roughly a 100x ARR-to-funding ratio. All three inherit the uncertainty of the estimate underneath them.

People and acquisitions

Zapier has bought three companies on the public record since 2024. All three had undisclosed terms and all three look like acquihires in shape, with the founders joining and the product folded in or switched off.

Vowel, a video-conferencing and meeting-intelligence startup, was announced on 6 March 2024 alongside Zapier Central, its own service having been switched off on 27 December 2023. Its chief executive, Andrew Berman, became Director of AI at Zapier (PR Newswire, 6 March 2024).

NoCodeOps came next, bringing its no-code operations community and founders Philip Lakin and Brent Summers (Zapier blog). Then Utopian Labs, which built AI sales-research and copywriting agents, with founders Steven Nelemans and Robin Salimans (Zapier blog). Both posts are undated; trackers place NoCodeOps in September 2025 and Utopian Labs in October 2025, so the months are secondary even though the deals are primary.

Two executives are named in primary sources. Wade Foster, co-founder and chief executive, is quoted on the 23 April 2026 governance release. Sheryl Soo, SVP of New Products and Head of Strategy, is quoted on the Central launch.

The one headcount event outside the press kit is a cut: about 10% of the workforce, more than 100 employees, in June 2023 (IBTimes). Trade press only, no company statement I could find, and it predates the verification window. It is in the data set because it is the only headcount movement anyone has documented, and labelled low confidence for the same reason.

Every Zapier pricing change with a date on it

This is the part no other Zapier statistics page carries, and it is the part most likely to be the reason you are reading. Zapier's pricing has changed four times on the documented record since the start of 2024, each with an effective date published on Zapier's own help centre or community site.

Read from Zapier's help centre and community pricing pages on 15 September 2026. No 2023 pricing change is traceable to a dated Zapier document, so this timeline starts on 8 January 2024. The June 2026 change is the one to model if your Zaps lean on AI steps: the default moved to Advanced, so the default tier now bills at 3x the Standard rate.

Current list pricing, read 15 September 2026

Straight off zapier.com/pricing on 15 September 2026. Task allowances on Professional and Team run on a slider, so the entry price is the floor of a plan, and the top of each slider is where the real spread lives.

All figures read from Zapier's pricing page on 15 September 2026, per month, in US dollars. Annual rates are shown as Zapier shows them, billed annually. Add-on prices are the annual-billing monthly equivalents Zapier displays.

The line worth staring at is the overage multiplier. Monthly billing charges 2.5x the plan rate for tasks beyond the allowance, against 1.25x on annual, so the cost of staying flexible is paid twice: once on the base price and again on every task over the line.

The AI product timeline, 2024 to 2026

Zapier went from an AI preview product to enterprise AI governance in about two years, and almost every step has a date on a Zapier-owned page.

It starts on 6 March 2024, when Zapier Central entered public preview across 6,000+ apps, announced in the same release as the Vowel acquisition. That release also put 388,000 customers on Zapier's AI products and 50 million tasks delegated to automation, though it never defines the period or scope the 50 million covers (PR Newswire).

On 4 April 2025 Zapier MCP shipped in beta, exposing 30,000+ actions to any MCP client. The MCP guide was first published that month and last updated in August 2026. The Agents guide, which is the Central lineage under a new name, was first published in March 2024 and last updated in November 2025. Agents itself was restructured on 22 May 2025: "Behaviors" were removed, each one became an agent in its own right, agents group into pods, and a new All Activity screen arrived (Zapier Help Center).

December 2025 is where the enterprise turn becomes obvious: an Admin Center for plans with more than one seat, a Workday integration with nine finance triggers and actions, SOC 2 readiness and versioning beta for Agents, Copilot checkpoints, MCP tool bundle sharing, human-in-the-loop guest reviewers and a usage warning at 80% of allowance (Zapier blog). Then came 23 April 2026, the biggest release in the set: Zapier Agents generally available with Enterprise MCP, the Zapier SDK in open beta, App Access Controls, Action Restrictions, Managed App Connections with domain restrictions, Bring Your Own Model through Bedrock for Claude and Cohere, and log streaming to Datadog, Splunk and SIEM tools (Business Wire).

One correction to the record, since it is repeated everywhere. Several secondary write-ups date Zapier Agents' general availability to May 2025. That May 2025 document is a restructuring notice; the release saying "now generally available" is dated 23 April 2026. Cite April 2026.

Zapier's own research, labelled vendor-interested

Zapier publishes survey data, and it is genuinely useful material. It is also marketing sitting next to the product it flatters, so every figure here is labelled vendor-interested. None of these surveys publishes a margin of error.

The governance survey, fielded among 200 enterprise executives and released on 23 April 2026, produced two numbers: 93% say AI initiatives fail to reach production because of governance constraints, and 94% agree governance must shift to continuously operating systems (Zapier via Business Wire). Both appear in a release announcing governance features, which is the context to keep attached.

A second Zapier survey of more than 500 enterprise leaders, undated on the page and fielded around September 2025 according to its search metadata, supplies four figures: 78% struggle to integrate AI with existing systems, 56% describe themselves as enthusiastic AI champions, 81% feel competitive peer pressure to accelerate adoption, and on what is blocking them, 45% name vendor cost, 38% distrust vendor security, 35% cite skill gaps and 33% fear lock-in, while 53% say IT is the function accelerating adoption most (Zapier blog).

A third, labelled 2026 and also undated, covers agents specifically: 72% of enterprises use or are testing AI agents, and 84% will increase agent investment over the next twelve months, with deployment splitting 49% customer support, 47% operations and 47% data management, oversight running 38% human-in-the-loop against 20% fully autonomous, and 30% naming routine workflows as the biggest opportunity (Zapier blog). For third-party numbers on the same question, with survey houses that publish their methodology, see AI agent adoption statistics for 2026.

The internal figure is the one I find most interesting and trust least: 97% of Zapier employees use AI in daily work, up from 65% in late 2023 (Zapier blog). That page was not fetched directly during verification, so it carries an unverified label until the next update.

Market position, measured as a domain count

The Zapier "market share" figure that circulates is a technographic count. 6sense detects vendor technology on public web properties, counts the domains, and divides. That produces a measure of visible footprint. Revenue never enters the calculation.

On that basis 6sense puts Zapier at 13.49% of its Integration category, ranked second, on 24,855 detected customer domains (6sense, 2026 label, retrieved 15 September 2026). Set that against Zapier's own claim of 3.4 million businesses and the gap tells you what detection can and cannot see.

Where those domains sit

Geography concentrates hard in one country: the United States at 11,645 domains (58.30%), the United Kingdom at 1,608 (8.05%) and Canada at 1,157 (5.79%). Company size confirms the SMB story everyone assumes: 7,442 domains at 20 to 49 employees, 5,594 at 0 to 9, and 5,402 at 100 to 249. The modal Zapier customer, on this evidence, is a company of fewer than fifty people.

The category is too loose to call a market

Look at what else 6sense puts in that category. The largest entry in the same list is Mailchimp for Shopify at 40.78% on 75,111 domains, with ManageEngine at 7.26% and IBM WebSphere at 7.06%. A category holding a Shopify email app, an IT management suite and an enterprise application server together is a detection bucket. Quote the 13.49% as footprint, and label it that way every time.

Make and n8n, for scale comparison

Two comparables, with very different amounts of evidence behind them.

Make publishes scale and nothing else

Make's about page, read on 15 September 2026, claims 400,000+ customers across 200+ countries and territories, 3,000+ pre-built apps and 350+ employees. The company has been operating since 2015, formerly as Integromat, and is part of Celonis, which it describes as valued at $13B (Make). That $13B belongs to Celonis and is regularly misquoted as Make's own valuation. There is no revenue, task-volume or growth figure for Make anywhere I could verify.

n8n has the deepest paper trail of the three

n8n raised a EUR 55M ($60M) Series B at a EUR 250M ($270M) valuation on 24 March 2025, led by Highland Europe with HV Capital, alongside existing investors Sequoia, Felicis and Harpoon. At that point it reported an active base of 200,000, more than 3,000 enterprise customers and over 70,000 GitHub stars, with revenue up 5x since its 2022 AI pivot, doubling in the two months to March 2025, and around 75% of customers using AI features. It was founded in 2019 in Berlin by Jan Oberhauser (TechCrunch).

Seven months later the numbers changed shape entirely. On 9 October 2025 n8n announced a $180M Series C at a $2.5B valuation, taking total raised to $240M, led by Accel with Meritech, Redpoint, Evantic, Visionaries Club, NVentures and T.Capital, claiming 6x growth in its active base and 10x revenue growth across 2025 to date without publishing an absolute revenue figure (n8n).

Sacra models the rest: about $100M ARR in April 2026, roughly $70M at the end of 2025, and 775% year-over-year growth for 2025, plus 1.7 million monthly active builders on the Community Edition in late 2025, who are not paying customers. Sacra also records a EUR 60M SAP investment at a $5.2B valuation in May 2026, taking total raised to about $253.5M (Sacra). I could not locate SAP's own announcement of that round, so it stays labelled secondary.

Put the two Sacra models side by side and the growth rates carry the story. The same modeller used the same approach for both: Sacra's modelled estimates have Zapier at $310M ARR at the end of 2023, after 34.8% growth that year, and n8n at about $70M at the end of 2025, after 775% growth that year. One of those is a mature business, the other is a curve, and that is the competitive story in two numbers. For the feature-level version, Zapier vs n8n compares them directly, and the best n8n alternatives prices the wider field.

What is thin, and why

Six weaknesses run through this data set. I would rather list them than let you find them after you have cited something.

1. Revenue rests on one model and stops at 2023. Sacra is the only estimator, and the newest figure is end-of-2023. Every Zapier revenue claim about 2025 or 2026 is downstream of it.
2. The pricing timeline has no 2023 in it. Zapier certainly changed things that year. Nothing I found carries a dated, Zapier-published document, so the table starts in January 2024.
3. Make has a vendor about page and nothing else. Four self-reported figures, no revenue, no task volume, no growth rate, and a valuation that belongs to its parent.
4. Market share comes from a single technographic vendor. One method, one category definition, no second opinion to check it against.
5. There is no headcount history. The press kit says 800+ today. The only other datapoint is a June 2023 layoff reported by trade press, which is why the trend is unknowable.
6. Two acquisition dates come from trackers. Zapier's own posts for NoCodeOps and Utopian Labs carry no publication date, so the months are secondary even though the deals are primary.

Eleven widely quoted figures were dropped outright. "66,000 triggers and actions" appears nowhere on zapier.com. A "7.05% integrations market share" is contradicted by 6sense's own 13.49%. A 3 million customer count paired with "100,000+ paying customers" has no original anywhere. "1.5 billion tasks per month", "3.1B+ tasks monthly" and "25M+ Zaps" are undated with no upstream document. "$400M revenue in 2025" is a projection presented as a fact, and "$310M ARR in 2024" is the Sacra figure mis-dated by a year. A customer mix of 40/35/20/5 has no survey behind it, Make revenue of $52.6M has no methodology, and n8n at "more than $40M ARR" in March 2025 comes from misreading TechCrunch, which reported growth multiples only.

That is nearly a fifth of what is in circulation, discarded.

Every figure on one table

56 figures, one row per highlighted number on this page; the # links to the sentence it sits in and the publisher links to the original document where the link is recorded. Confidence is the CSV label as written: primary means the publisher collected the data; secondary means a credible re-report or model; vendor-interested means the publisher sells something the figure flatters; forecast means a prediction, not a measurement; unverified means the original blocked our read and the figure comes from a named re-report. The same rows are in data.csv (CC BY 4.0).

Use this data

Take it, republish it, build on it. That is what it is for.

The CSV has one row per anchored figure, 56 in all, keyed to the stat-N ids used on this page. If you are quoting a single number, link to its anchor. Your reader lands on the sentence with the source attached.

Corrections are welcome and I will date them. If you hold a primary document for anything marked here as modelled, undated or unverified, that is exactly the thing worth sending.

FAQ

Sources