Optical networking leader Ciena (NYSE: CIEN) delivered better-than-expected results and guidance prior to Thursday's market open, fueled by surging AI-related demand for data center interconnectivity.

For its fiscal 2026 third quarter, which ended on August 1, 2026, the company reported revenue of $1.671 billion, a 37% year-over-year increase that surpassed the consensus estimate of $1.64 billion. Adjusted earnings per share came in at $2.11, a notable 215% jump from the prior year and comfortably ahead of the $1.73 analysts had projected.

Breaking down the segments, revenue from the optical networking unit, which ties directly to data center optical interconnect demand, climbed 46% year over year to $1.191 billion. Meanwhile, the routing and switching business contributed $164 million in revenue, reflecting a 31% increase. The remaining revenue came from various services during the quarter.

CEO Gary Smith emphasized in the earnings release that AI is driving compounding growth in network investment. As the only pure-play company focused on optical systems and interconnect products, he noted that the firm's unmatched combination of an established customer base, technological innovation, and deep expertise creates a powerful competitive edge.

During the quarter, two customers each accounted for 10% or more of total revenue, collectively representing 41.7% of the company's sales. The company also reported days sales outstanding of 76 days and an inventory turnover rate of 3.5 times.

Ahead of the announcement, Wall Street had focused heavily on guidance as a barometer for just how resilient data center demand remains. Ciena met those expectations by raising its fiscal 2026 full-year revenue guidance midpoint to $6.42 billion, up from the prior $6.3 billion and above the $6.34 billion consensus. For the fourth quarter, the company projects revenue around a midpoint of $1.75 billion, also topping the $1.7 billion that analysts were looking for.

The report clearly illustrates the boost that AI data center demand is providing to the company's performance. Looking at the medium to long term, as AI compute infrastructure expands toward greater clustering, distributed architectures, and cross-data-center collaboration, optical interconnect is evolving from a cyclical communications equipment play into a critical bottleneck asset for AI infrastructure.

Interestingly, while Ciena operates alongside Marvell Technology, Lumentum, and Coherent in the broader AI optical interconnect ecosystem, each player occupies a distinct position in the value chain. Marvell skews toward the semiconductor layer, offering optical DSPs, coherent-lite DSPs, PAM4 DSPs, switching chips, SerDes, custom AI ASICs, and silicon photonics platforms. Lumentum and Coherent are more focused on optical components, modules, lasers, and transceivers.

Ciena, by contrast, positions itself as a platform-level optical networking company delivering complete data center optical systems and coherent transmission solutions. Its core strengths lie in WaveLogic coherent optical engines, 800G and 1.6T coherent transmission, optical networking equipment, routing and switching, network automation, and ecosystem solutions for carriers and cloud providers, rather than just selling optical chips or lasers at a component level.

This makes the company more of a systems-level AI optical interconnect and DCI play, with traditional strengths in high-speed interconnect between data centers, metro and long-haul networks, carrier backbones, and cloud WAN connectivity, effectively linking multiple data centers, campuses, regional clouds, and telecom networks through high-speed coherent optical systems.

However, Ciena is actively extending its reach from inter-data-center DCI toward high-speed optical interconnect inside AI data centers. The acquisition of Nubis was aimed squarely at expanding the inside-the-data-center strategy to handle the demands of AI workloads. Industry trends are likewise pushing the optical signal chain ever closer to the rack, switching chips, and even advanced packaging systems, moving inward from inter-building and inter-campus connections.