AI demand creates a massive backlog gated by supply

A $10 billion backlog meets a supply wall

Ciena is riding a wave of AI infrastructure spending. Cloud companies desperately need faster links between data centers to handle AI workloads, and Ciena provides the optical hardware to make it happen. This demand drove Q3 2026 revenue up 37% year over year to $1.7 billion. More importantly, direct cloud provider revenue grew more than 80%. The company expects its backlog to cross $10 billion by the end of fiscal 2026, giving it unprecedented visibility into future sales.

Management is so confident in this pipeline that they set an early fiscal 2027 revenue floor of $8.3 billion to $8.4 billion, which implies growth of at least 30%. The upcoming launch of the Hyper-Rail platform in late 2026 positions the company well for next-generation network architectures. Gross margins have also expanded to a healthy 46.4% in the third quarter.

However, this massive demand has hit a wall. Ciena cannot build equipment fast enough. Management stated that supply constraints will cap top-line growth and likely persist until 2028. The bear case centers on these limits, alongside extreme customer concentration. Two cloud providers each account for more than 10% of total revenue. If one changes its network design or slows orders, Ciena would feel an immediate impact. The valuation also reflects high expectations, leaving little room for error if supply issues delay revenue recognition.

Big projects, concentrated customers

Ciena makes money by selling networking hardware, software, and services. The vast majority of revenue comes from Networking Platforms, particularly Optical Networking systems such as the 6500 RLS and Waveserver. These products allow cloud companies and telecom carriers to push massive amounts of data over fiber networks.

The company supplements its hardware sales with platform software, automation tools like Blue Planet, and global services. These segments help customers design, deploy, and manage their networks. The business model works best during large capital spending cycles, especially when cloud providers upgrade their infrastructure to support new technologies like AI.

This model is highly sensitive to the spending plans of a very small group of massive companies. Ciena must commit to buying parts long before it ships finished products. When demand outpaces the supply chain, as it currently has, the company cannot recognize revenue even with orders in hand. Margins are also sensitive to the mix of products sold and the costs of delivering services, though recent quarters have seen gross margins stabilize in the mid-40s range.

What Ciena sells

Q2 revenue mix

Segment shares are from Ciena's Q2 fiscal 2026 disclosure. The mix remains heavily skewed toward hardware, driven largely by cloud provider capital expenditures.

What could go wrong

In one breath