Ciena's latest earnings commentary turns the AI infrastructure boom into a test for optical networking, not only compute capacity.

According to RCR Wireless News, the company said AI-driven demand could approximately double its addressable market to roughly $50 billion by 2029, with spending spreading across wide-area networks and data center networks.

Why AI compute needs optical capacity

Ciena president and chief executive Gary Smith told investors that the largest hyperscalers had increased 2026 capital expenditure plans, with indications of expansion into 2027 and beyond.

His argument is that companies spending heavily on AI compute will need more network capacity to move data for training, ingestion and inference.

That matters here because AI infrastructure spending is often discussed through chips, power and data center shells.

Ciena is framing the next bottleneck as high-capacity, low-latency connectivity across long-haul routes, metro networks and data center interconnects.

If that view holds, optical systems could capture a larger share of the AI budget as customers try to monetize constrained compute investments.

The products behind the forecast

The company pointed to several demand channels rather than one product line.

RLS Hyper Rail, Ciena's next-generation intelligent line system, was developed with multiple hyperscalers and is aimed at capacity and efficiency requirements around data centers.

Smith said Ciena had received what he described as the industry's first multirail order from a leading hyperscaler for the platform.

RCR Wireless News also reported that RLS Hyper Rail can run several fiber-pair paths side by side across long routes measured in hundreds of kilometers, using amplification to extend capacity.

The signal is narrow but important: Ciena has an early hyperscaler order, while additional hyperscalers, neocloud providers and service providers are still in evaluation discussions.

Ciena also cited momentum in Data Center Out-of-Band Management, or DCOM, which combines routing and switching products with passive optical network technology.

Smith said DCOM contributed to 88% year-on-year revenue growth, and that Ciena had initial orders from a second hyperscaler customer in addition to Meta, with a third hyperscaler moving through lab qualification.

Carriers are part of the exposure

The affected market is wider than hyperscale cloud.

Ciena also sees carriers returning to network spending after a restrained investment period, with openings in long-haul, metro and managed optical fiber networks.

That makes the forecast relevant to carriers trying to support AI traffic without turning every investment cycle into a full data center buildout.

The company also reported a new win with a major hyperscaler for high-performance coherent modules.

Those modules are expected to be deployed across metro and long-haul data center interconnect networks that support both WAN and data center applications.

Management also expects 400G and 800G pluggable optics to remain a growth line, with pluggable revenue projected to exceed twice the 2025 level.

What has to prove out next

The forecast still depends on order conversion.

A $50 billion addressable-market target does not by itself prove that AI networking demand will spread evenly across vendors or customers.

What matters now is whether Ciena's early hyperscaler wins expand into repeat deployments, whether service-provider reinvestment continues, and whether DCOM's growth can hold after the first wave of hyperscaler adoption.

Ciena is preparing for that demand with spending of $250 million to $275 million in capital expenditures this year, including manufacturing and supply-chain capacity.

The question is whether optical networking becomes a sustained part of AI infrastructure budgets, or whether the current demand is concentrated in a small number of aggressive hyperscale builds.