Summary

  • Ciena says cloud providers generated 53% of its US$1.671129 billion fiscal Q3 2026 revenue and that revenue from the cohort rose 82% year on year.
  • The 10-Q separately records Cloud Provider A at US$476.570 million and Cloud Provider B at US$219.644 million. Together they supplied US$696.214 million, or 41.6613% of quarterly revenue.
  • Taking the rounded 53% literally leaves about US$189.484 million of cloud-provider revenue beyond A and B. This is an estimate from a rounded percentage, not a disclosed subtotal or a third customer.
  • The filings do not allocate the 82% growth, gross margin, receivables, contract assets, deferred revenue, product growth or remaining performance obligations to either anonymous customer.

Two windows on one quarter

Ciena reported US$1.671129 billion of revenue for the quarter ended 1 August 2026. Its earnings presentation divides that total by customer cohort: cloud providers account for 53%, service providers 24%, enterprise 14% and government and research 9%. The same slide says cloud-provider revenue grew 82% from the prior-year quarter.

The 10-Q answers a different question. Accounting concentration disclosure isolates customers that supplied at least 10% of revenue. Cloud Provider A contributed US$476.570 million and Cloud Provider B US$219.644 million. The labels are intentionally anonymous. They establish two large exposures; they do not establish that the cloud cohort has only two members.

Adding the two rows gives US$696.214 million. Dividing that by total revenue gives 41.6613%. The arithmetic is exact to the disclosed dollar units. It is also materially below 53%. Any analysis that substitutes the two-customer total for the cloud-provider cohort silently discards more than eleven percentage points of quarterly revenue.

The apparent US$189.5 million is a boundary, not a new fact

Applying 53% to quarterly revenue produces US$885.698 million. Subtracting A and B gives about US$189.484 million. This is a useful reconciliation because it proves that the two large customers cannot exhaust the presented cloud cohort.

It is not a new accounting line. The 53% figure is shown as a whole percentage, so the implied dollar amount is rounded. Ciena does not say how many other cloud providers contributed the remainder, whether their shares were evenly distributed or whether one sat just below the 10% threshold. Calling the residual “Customer C” would replace a missing disclosure with an invented identity.

The same discipline applies over nine months. A generated US$1.128775 billion and B US$592.137 million. Their US$1.720912 billion total was 36.8590% of nine-month revenue of US$4.668910 billion. That exact concentration measure is not comparable to a quarterly cohort percentage without keeping both the period and the population fixed.

An 82% cohort increase does not identify the engine

The 82% growth rate belongs to all cloud providers in the presentation’s definition. It cannot be assigned to A, to B or to the estimated remainder. Nor does it disclose retention, price, volume, new wins or the timing of customer acceptance.

The comparison illustrates why a high growth rate and a high concentration ratio can tell different stories. A and B may have expanded, smaller cloud customers may have crossed from trials into deployment, or several effects may have occurred together. The public documents do not provide the bridge. What they do say is that a small number of cloud providers are becoming a larger share of revenue across multiple operating segments.

That cross-segment language matters. It suggests the cloud cohort is not a product line. A buyer may appear through optical networking, routing and switching, or other systems, while product reporting groups revenue by what Ciena sells rather than by who buys it. The earnings presentation says RLS and Waveserver revenue grew more than 55% and coherent pluggables more than doubled. Those product gains cannot be mapped back to A or B from the published tables.

Margin and working capital have no customer tags

GAAP gross margin reached 45.4%, against 41.3% a year earlier. Ciena cites cost reduction, pricing optimisation, product mix and tariff refunds among the drivers. The result is evidence that growth did not mechanically erase gross margin during this quarter. It is not proof of the economics of either large customer.

The balance sheet is equally resistant to customer attribution. Net accounts receivable was US$1.233610 billion, contract assets US$160.888 million and deferred revenue US$311.281 million. Remaining performance obligations were about US$2.5 billion. None of these amounts is split among A, B and the rest of the cloud cohort.

That means days sales outstanding cannot be turned into an A-or-B collection story, deferred revenue cannot be assigned to a hyperscale programme and RPO cannot be used as a customer order book. The timing chain still runs from order through production, delivery, acceptance, invoicing and cash. The filings reveal company-level states, not the customer-level path through them.

What the market can monitor without naming the buyers

The useful dashboard keeps five ledgers apart: cohort share, threshold-customer share, product mix, operating segment and cash conversion. If the cloud cohort stays near 53% while A+B falls, demand is broadening inside the disclosed population. If both rise, concentration is tightening. If cohort revenue rises while margin and cash conversion weaken, volume may be arriving with less favourable economics or different timing—but the cause would still require evidence.

A better map would include a stable definition of “cloud provider”, concentration bands below 10%, customer retention or expansion measures, customer-by-segment exposure and working-capital ranges by cohort. Ciena need not reveal commercial identities to improve the information. It can make the denominator more durable without naming a buyer.

The central conclusion is modest but consequential: the cloud cohort is larger than its two reportable customer rows. The 53% headline describes market exposure; the 41.6613% calculation describes identified concentration under an accounting threshold. They should be placed side by side, never collapsed.

Sources and evidence limits

Quarterly revenue, customer A/B values, nine-month totals, concentration language, margin factors, balance-sheet balances and RPO come from Ciena’s fiscal Q3 2026 Form 10-Q. The 53% cloud-provider share, 82% cohort growth and product highlights come from the earnings presentation, cross-checked against the earnings release and the filed Form 8-K.

The sources do not identify A or B, provide an unrounded cohort share, state the number of other cloud customers or publish customer-level margin, working-capital, RPO, product or segment allocations. The US$189.484 million residual is arithmetic based on the rounded 53% figure. No customer identity, future result, misconduct or investment recommendation is inferred.