Thesis. Ciena is successfully migrating from a slow telecom-carrier business to a fast hyperscaler one — cloud-direct revenue up 76%, three consecutive guidance raises, and systems-level margins rather than component exposure. The cost is severe customer concentration and a slower sales cycle than the component names.
Reviewed 19 September 2026, after fiscal Q3 2026 results. Next review after Q4 earnings on 10 December 2026.
What it does
Ciena builds complete optical networking systems — the boxes, the coherent transport equipment and the software that run fibre networks. It sits one layer above the component makers: it buys lasers and transceivers, and sells finished systems.
Historically it sold to telecom carriers. Increasingly it sells straight to cloud companies for datacentre interconnect — connecting one datacentre to another over long distances.
Selling systems rather than components means better gross margins and worse timing. Ciena reacts to the AI build-out with a lag, because a system is specified, tested and installed long after the chips inside it were ordered.
Bull case
- Q3 FY2026, reported 3 September: revenue $1.67B, up 37% — a record quarter for both orders and revenue
- Earnings are inflecting hard. Adjusted EPS $2.11, up 215% year over year. GAAP EPS $1.83. Adjusted gross margin 46.4%
- The third straight guidance raise. Full-year FY2026 revenue raised to $6.42B ± $50M, up 35% at the midpoint. Q4 guided to $1.75B ± $50M with 20% adjusted operating margin
- The transition is real, and it's the whole point. Direct cloud provider revenue grew 76% — evidence of a genuine shift from carrier capex to hyperscaler capex, which is faster-growing and less bound to budget cycles
- Visibility into 2027. Backlog of $7.7B at Q2 FY2026, up more than $600M sequentially from around $7.0B
- A first-of-its-kind win — the industry's first multi-rail order for RLS Hyper-Rail, from a leading hyperscaler
- Systems-level gross margins are structurally well above the module makers and contract manufacturers
Bear case
- Customer concentration is extreme. Two customers were 41.7% of revenue in Q3 FY2026. Neither is named in the release. Losing or being de-prioritised by one would be a material event
- Still exposed to telecom carrier capex, which is lumpy and largely unrelated to AI
- Slower sales cycles than component peers, so revenue reacts to the build-out with a lag in both directions
- No supply-side pricing power. It sells systems, it doesn't own a scarce input — and rising component prices from the laser shortage are a cost here, not a benefit
- Competes against much larger diversified vendors
- Three raises in a row sets a very high bar for the fourth
Major customers
Two customer bases, moving on completely different clocks.
Cloud providers, direct — the growth engine. Hyperscalers buying datacentre interconnect and long-haul transport equipment directly rather than through a carrier. Direct cloud provider revenue grew 76%, which is the single most important number on this page.
Telecom carriers — the legacy base. AT&T, Verizon, international carriers and cable operators. Budget-cycle driven, lumpy and slow.
On concentration: Ciena disclosed that two customers together accounted for 41.7% of Q3 FY2026 revenue, both above the 10% threshold. It did not name them. Given the pace of the cloud-direct shift, the composition of that pair is worth establishing from the annual filing rather than assumed.
Read-through. The cloud-direct percentage is what decides whether Ciena re-rates as an AI infrastructure name or stays valued as a telecom supplier. Watch hyperscaler capex — GOOG and META among them — for the growth half, and carrier capex commentary for the drag half.
Ciena is a customer of COHR and LITE, not a competitor. The laser shortage that drives their pricing power shows up in Ciena's cost line — the same fact reads as a positive on their pages and a negative on this one.
VIAV runs ahead of Ciena, since its test instruments are bought before Ciena's systems ship.
What would change the view
- Cloud-direct revenue as a share of total — the transition metric, and the one that drives the multiple
- Who the two 41.7% customers are, and whether that concentration rises or falls
- Whether the guidance raises continue into a fourth
- Gross margin trend as the mix shifts toward cloud
- Carrier capex commentary in the next print
- 1.6T system wins at hyperscalers, and the RLS Hyper-Rail multi-rail ramp
Update log
- 19 Sep 2026 — Q3 FY2026 results added: revenue $1.67B (+37%), adjusted EPS $2.11 (+215%), adjusted gross margin 46.4%, FY26 guidance raised to $6.42B. New disclosure: two customers at 41.7% of revenue, added to the bear case as the dominant risk. The previous note that AT&T and Verizon were each above 10% is no longer supportable as written — the release names neither customer.
- 11 Sep 2026 — Reviewed.
- 7 Aug 2026 — Full review after Q2 FY2026.
Research and education only — not investment advice.
Made with Bullet