Summary

  • Zoom's fiscal-Q2 Enterprise revenue was US$787.5 million, up 7.8% year on year, while Online revenue grew only 0.6%. Enterprise now supplies 61.7% of total revenue.
  • Trailing-12-month net dollar expansion for Enterprise customers was 99%. That measure compares endpoint ARR from the same customer set and includes upsells, contractions and attrition; it is not quarterly revenue growth or a pure churn rate.
  • The two figures can coexist because new customers and other movements outside the fixed cohort can grow segment revenue while prior-period customers contract slightly in aggregate. Zoom does not disclose a dollar bridge between the measures.
  • Customers above US$100,000 of trailing revenue increased 8.2%, RPO reached US$4.536 billion and operating cash flow remained US$494.8 million. Those are real strengths, but none proves that the installed base has returned to net expansion.

Two clocks inside one enterprise business

Zoom entered fiscal Q2 with a familiar strategic ambition: turn a product identified with video meetings into a broader operating layer for enterprise work. It exited the quarter with evidence that the transition has commercial weight. Enterprise revenue reached US$787.5 million and grew 7.8%, the strongest pace in three years, according to the company's quarterly results. The enterprise share of total revenue rose to 61.7% from 60.0%.

The same release carried a more restrained number. Zoom's trailing-12-month net dollar expansion rate for Enterprise customers was 99%. It had improved from 98% a year earlier, but it remained below the point at which the same customer base adds more recurring value than it loses.

There is no contradiction to resolve by choosing a preferred percentage. There are two clocks.

Quarterly Enterprise revenue measures recognized revenue from the enterprise segment as it exists during the reporting period. Net dollar expansion begins with the annual recurring revenue of a defined set of Enterprise customers 12 months earlier, observes the ARR of that same set at the current endpoint, then averages the resulting ratios over the trailing year. One clock can benefit from customers that were not in the old cohort. The other is designed to keep them out.

The temptation is to treat 107.8% revenue and 99% expansion as parts of one bridge. That would produce an apparently precise 8.8-point contribution from outside the installed base. It would also be wrong. The percentages differ in population, period, recognition basis and denominator. Zoom publishes no reconciliation that would support the subtraction.

What the pair does support is a more useful claim: Zoom's enterprise acquisition and mix engine is currently stronger than its installed-base expansion engine. Whether that balance becomes durable depends on the latter crossing 100% without the former slowing.

What 99% actually contains

The Form 10-Q defines Enterprise customers as distinct business units engaged by Zoom's direct sales team, resellers or strategic partners. That is already a warning against reading the metric as a count of legal corporate groups. One multinational can contain several buying units; a change in buying structure can affect the perimeter.

For each period end, Zoom starts with ARR from all Enterprise customers present 12 months earlier. ARR is the last month's recurring revenue run-rate multiplied by 12. Monthly subscribers remain in MRR if they have not indicated that they intend to cancel. Zoom then calculates current ARR from the same customers, including any upsell, contraction and attrition, and divides it by prior-period ARR. The reported trailing figure is an average of those expansion ratios over 12 months.

At 99%, the measured cohort finished with slightly less ARR than its prior base on average. That does not mean Zoom lost 1% of Enterprise customers. It does not mean quarterly enterprise revenue fell 1%. It does not isolate gross retention, because upsells can offset downsells and attrition before the ratio is produced.

Zoom's explanation is candid. Over several years, macroeconomic pressure, slower hiring and higher seat-count downsells in key markets pushed enterprise expansion below 100%. A collaboration subscription priced partly around paid licences feels employment and budget restraint directly. A customer can remain loyal, buy additional products and still reduce the number of seats enough to leave total recurring value lower.

The rate was 99% in Q1 as well, after 98% in the comparable prior-year quarter. Stability at 99% is therefore both progress and a limit. The erosion has narrowed. It has not yet reversed into aggregate net expansion.

Growth has to arrive from somewhere else

Zoom says acquiring new customers was among the drivers of total revenue. That statement supplies the direction of the reconciliation, not its amount.

New enterprise customers can add recognized revenue while staying outside the prior-period cohort used in the expansion calculation. Existing customers may also change product mix in ways that affect recognized quarterly revenue differently from endpoint ARR. Pricing, foreign exchange, contract timing and acquisitions can contribute. The Q2 filing does not allocate the 7.8% rate among those mechanisms.

The large-customer count is the clearest supporting evidence of breadth. Customers contributing more than US$100,000 of trailing-12-month revenue reached 4,625, up 8.2% from 4,274. They represented 33.4% of total revenue, compared with 32.2% a year earlier. The count also rose from 4,534 in Q1.

But a threshold count is not a new-logo ledger. An existing customer can cross US$100,000 through expansion, an acquired product or simply the rolling period. A new customer can enter above the threshold. Another can remain large while shrinking. The metric proves that more buying units occupy Zoom's upper revenue tier; it does not tell which path put them there.

That distinction matters because the cost of maintaining segment growth changes with the path. If most growth requires a continuing stream of new accounts to replace contraction in the old base, sales capacity and acquisition efficiency carry more weight. If cross-sell and seat recovery lift the same cohort above 100%, growth compounds with less replacement pressure.

AI adoption has a numerator but not a scale

Management presented AI as the lever that could change the installed-base equation. It said the AI-first Customer Experience portfolio delivered high-double-digit ARR expansion and that Zoom Virtual Agent's customer count increased 256% year on year. Products such as ZoomMate, AI Productivity Suite and Workvivo HQ Agent extend the company beyond meetings into customer service, recruiting, revenue work and employee experience.

The adoption signal is important. A product that solves additional workflows creates a route to more spend without requiring a seat recovery in the original meeting product. It can also make the platform more deeply embedded in a customer's operating process.

The economic evidence is not yet commensurate with the language. Zoom did not publish the absolute Virtual Agent customer count, its starting base, product revenue, product ARR, attach rate or retention. Nor did it quantify the Customer Experience portfolio's contribution to the US$787.5 million of Enterprise revenue.

A 256% increase from an undisclosed base can indicate rapid adoption while remaining small in consolidated economics. “High-double-digit ARR expansion” describes the direction of a portfolio, not its weight. Neither disclosure proves that AI caused the segment's 7.8% growth or that AI expansion has offset seat downsells across the existing cohort.

The useful next disclosure is not another product count in isolation. It is a bridge: starting product ARR, new customers, cross-sell into existing accounts, contraction, retention and ending product ARR. That would reveal whether AI is widening the acquisition funnel, repairing installed-base expansion or doing both.

Contracts and cash keep the cautious reading honest

The argument is not that enterprise demand is weak. Zoom's remaining performance obligations reached US$4.5359 billion at July end, up from US$4.2986 billion in April. The Q2 total comprised US$1.5622 billion of billed consideration and US$2.9737 billion unbilled; 57% was expected to become revenue within 12 months.

RPO is a meaningful contract ledger. It is not ARR and does not preserve the same customer cohort. Zoom's agreements can be monthly, annual or multiyear, and billing timing differs. The balance can grow through new customers, renewals, longer terms and other contract changes even while net dollar expansion remains below 100%.

Cash generation also remains substantial. Operating cash flow was US$494.8 million and free cash flow US$472.4 million in Q2. Yet both were lower than a year earlier. GAAP operating income declined to US$314.3 million from US$321.7 million, and GAAP operating margin fell to 24.6% from 26.4%. Non-GAAP operating income rose modestly, but its margin also declined.

These figures resist both easy stories. Zoom is not a cash-poor company masking collapse with a customer metric. Nor does faster Enterprise revenue automatically demonstrate improving operating leverage. The exceptionally large quarterly net income was dominated by US$1.614 billion of gains on strategic investments, which belongs outside the product-economics test.

The Online business should also remain separate. Online revenue grew 0.6%; average monthly churn was 2.9%, unchanged year on year; and customers with at least 16 months of continuous service supplied 75.6% of Online MRR. Online churn starts with quarterly entry MRR and tracks cancellations, downgrades and indicated cancellations. It cannot be merged with an Enterprise expansion ratio built from a different population and formula.

The threshold that changes the quality of growth

Zoom has made the enterprise segment the centre of its growth story. Q2 supports that decision: enterprise revenue accelerated, the mix shifted toward enterprise, large accounts increased and the contract book widened.

The 99% rate identifies what the story has not yet achieved. The prior enterprise cohort, after all additions and losses within that cohort, still produces slightly less endpoint ARR on average. New customers and movements outside that calculation can compensate, but compensation is not compounding.

A move above 100% would not prove every product is healthy. It would show that the same-set base is again adding recurring value in aggregate. Sustained movement above the line, accompanied by product-specific AI economics and stable cash conversion, would make the broader-platform thesis materially stronger.

Until then, the correct reading uses separate ledgers: Enterprise revenue for segment scale, net dollar expansion for the installed cohort, large-customer count for breadth, product ARR for AI economics, RPO for contracted revenue and cash flow for realized financial capacity. Zoom's quarter is encouraging because several ledgers improved. It remains incomplete because the one designed to measure compounding stopped at 99%.

Sources