Ciena (CIEN.US) stands in the light and strengthens its performance in AI optical interconnect. Q2 profit surged by 290%, performance outlook surpasses expectations.

Focus on the high-speed interconnection between different data centers (DCI) and the high-speed optical interconnection requirements within AI data centers, Ciena (CIEN.US) announced strong quarterly performance and fiscal year revenue outlook far exceeding market expectations, highlighting the high-speed optical interconnection demand brought about by the frenzy of global AI data center construction dominated by major American cloud computing service providers. This has transformed from a "long-term growth narrative" into a synchronous accelerated expansion trajectory of AI optical interconnection/optical communication infrastructure industry chain revenue, gross margin, and adjusted EPS. Shortly before Ciena announced strong performance, NVIDIA Corporation CEO Huang Renxun, known as the "godfather of AI," spoke on the second day of the 2026 Taipei Computex conference alongside Marvell CEO, stating that Marvell Technology, Inc.(Marvell) has the potential to become the next listed company with a market value exceeding $1 trillion, and described its crucial role in large-scale AI data center optical interconnection infrastructure systems. This clear endorsement from the strongest leader in the AI computing power field significantly strengthens market confidence in Marvell's long-term growth path. During his keynote speech at the 2026 Computex conference titled "The Future of AI Scaling Depends on Connectivity," Marvell Technology, Inc. CEO Matt Murphy emphasized that as the demand for AI big models related computing power resources and large-scale AI computing infrastructure clusters grows exponentially, computational performance has gradually outpaced data movement and high-speed DCI interconnection capabilities within and between data centers. Therefore, the next stage of AI computing power expansion bottleneck will be dominated by "connection capabilities rather than single-node computing power." Leaders in the AI optical interconnection/optical communication infrastructure industry chain, such as Ciena, who "stand in the light" with Marvell Technology, Inc., are undoubtedly positioned at the core of this new infrastructure lifeline with AI data center optical interconnection, ultra-high-speed DSP, 1.6Tb/s coherent optical transmission, and silicon photonics technology chains. Ciena's revenue surges by 40%+ and management raises full-year performance guidance For the second quarter of the 2026 fiscal year ending May 2, 2026, Ciena's total revenue reached $15.707 billion, representing a significant year-on-year growth of 39.5%, significantly higher than the $11.259 billion for the same period in 2025; adjusted EPS reached a high of $1.64, a substantial 290% increase compared to $0.42 for the same period last year. Both of these core data exceeded expectations compared to market expectations. FactSet consensus showed an EPS of $1.46 and revenue of approximately $15.1 billion, meaning Ciena's actual performance exceeded expectations by $0.18 per share and approximately $60 million. Seeking Alpha data showed revenue exceeding expectations by approximately $70 million. Both directions were consistent, highlighting that revenue and profit were significantly stronger than the growth expectations that Wall Street analysts had been continuously raising recently. Profit quality also significantly improved. Ciena's second-quarter GAAP gross margin rose from 40.2% for the same period last year to 44.0%, while adjusted gross margin rose from 41.0% to 44.9%; GAAP operating margin rose significantly from 2.9% to 15.1%, while adjusted operating margin rose from 8.2% to 19.5%; GAAP diluted EPS increased significantly from $0.06 to $1.49, while adjusted EBITDA soared from $116.7 million to $341.8 million. This shows the company is not solely driven by a "boom in data center orders," but has entered a phase of strong operating leverage release: optical interconnection + data center DCI + network hardware, as well as a combination of platform software, product after-sales services, and customer concentration that collectively elevated profit elasticity. In terms of business structure, Ciena's growth is highly focused on the AI data center high-speed interconnection cycle. Revenue for the second quarter in the Networking Platforms business reached $1.274 billion, accounting for 81.1% of total revenue; within this, the Optical Networking business, closely related to data center optical interconnection, generated revenue of $10.998 billion, representing 70.0% of total revenue, a significant increase compared to $0.7736 billion for the same period last year Routing and Switching revenue was $1.742 billion, also higher than the $927 million for the same period last year. The top two customers contributing over 10% each accounted for 34.0% of revenue, indicating that ultra-large scale cloud customers and major AI data center projects remain the core driving forces behind this growth. However, this also implies that customer concentration is a variable that must be monitored for valuation purposes. In terms of performance outlook that investors are focusing on, the company's outlook continues to be raised, further validating that the demand for AI optical interconnection is not a one-time expansion. Ciena expects revenue for the third quarter of the fiscal year to be in the range of $1.625 billion, with fluctuations of $50 million, higher than the Wall Street consensus estimate of $15 billion, and an adjusted gross margin of approximately 45%50 basis points, and an adjusted operating margin of 19%20%; they also raised the full-year revenue guidance for 2026 to $6.3 billion $100 million, with the mid-point indicating a significant year-on-year growth of approximately 32%, well above analysts' expectations of around $6.08 billion, and expect an adjusted gross margin of 44.5%45% and an adjusted operating margin of approximately 19%50 basis points for the full year. This upward revision from the revenue range of $5.9 billion$6.3 billion provided in March of this year reflects the management's increased confidence in the visibility of data center orders, delivery pace, and supply and demand environment. Ciena's performance report indicates that the company's management believes that the expansion of AI data centers is pushing high-speed connectivity to the infrastructure bottleneck, with a strong emphasis on the continuous rise of high-speed optical connection demands inside and outside data centers and between data centers. Looking from a medium to long-term perspective, as AI computing infrastructure expansion moves towards clustering, distribution, and cross-data center collaboration, optical interconnection transforms from a "communication equipment cycle product" to a crucial bottleneck asset in AI computing infrastructure. It is worth noting that Ciena, Marvell, Lumentum, and Coherent all belong to the AI optical interconnection/optical communication infrastructure industry chain, but their specific positions within the industry are different. Marvell is more oriented towards bottom-layer semiconductors: optical DSP, coherent-lite DSP, PAM4 DSP, switching chips, SerDes, custom AI ASICs, and silicon photonics platforms; Lumentum and Coherent are more focused on optical devices/modules/lasers/receivers and other optical hardware. Ciena, on the other hand, is more like a platform-based optical interconnection/optical communication company focused on a complete set of data center optical systems and coherent optical transmission, with core capabilities in WaveLogic coherent optical engines, 800G/1.6T coherent transmission, optical network equipment, routing switches, network automation, and operator/cloud service provider network ecosystem solutions, rather than simply selling optical chips or core lasers. Therefore, Ciena leans more towards "system-level AI optical interconnection and DCI," and the company's traditional strengths are more geared towards high-speed interconnection between different data centers, metropolitan/long-haul/carrier networks, and cloud service provider WAN interconnections, i.e., connecting multiple data centers, campuses, regional clouds, and telecommunications networks using high-speed coherent optical systems. However, Ciena is expanding from "DCI between data centers" to "high-speed optical interconnection within AI data centers." The acquisition of Nubis aims to expand the inside-the-data-center strategy to address AI workloads; industry trends are also driving the optical signal chain from inter-room, inter-building, and inter-campus connections, gradually approaching server cabinets, switch chips, and even advanced packaging systems. 1.6T DCI, CPO, and high-speed optical interconnection in data centers trigger a new investment trend in the computing power industry chain When looking at Ciena within the entire AI optical interconnection infrastructure industry chain, its position is closer to being a systemic beneficiary of "wide area network/WAN + data center interconnection + high-speed interconnection within and around data centers," while Marvell leans more towards the semiconductor base of "chip/module/switching chips/optical DSP/ASIC/silicon photonics." Ciena clearly positions its long-term strategy as WAN and high-speed connections within and around data centers, and through the acquisition of Nubis, has entered the closer to AI cluster internal CPO/NPO and high-speed copper interconnection links; Nubis technologies include CPO/NPO optical modules with a maximum bidirectional bandwidth of 6.4Tb/s and an active copper cable solution supporting 200Gb/s per channel, with a maximum length of 4 meters, with the goal of addressing power consumption, density, and latency constraints in AI cluster scale-up and scale-out. Ciena's strong performance and the recent strong revaluation of Marvell by the market are highly consistent with the logic: the bottleneck of AI infrastructure is shifting from simply expanding GPU computing power to connectivity, switching, optical modules, DSP, SerDes, CPO, DCI, and network power consumption. As AI training and intelligent agent workloads drive GPU clusters from single cabinets to multiple cabinets, multiple data centers, and even cross-regional AI factories, data traffic will explosively expand at all three levelschip/cabinet internals require low-power high-speed interconnection, data centers need high-radix switching and optoelectronic coordination, and data centers need 1.6T-level coherent-lite, DCI interconnection, long-distance coherent optical transmission, and even CPO and a series of AI optical interconnection infrastructure. Morgan Stanley's latest expectations clearly highlight the rise of AI intelligent agents globally, and the investment trend in AI computing power is shifting from a "competition around AI GPU single-point computing power" to a "full-stack computing system driven by AI intelligent agents." The supply chain bottleneck at the AI computing power infrastructure level has expanded from "large-scale purchases of GPUs/ASICs" to "striving to simultaneously address the entire chain of AI data center delivery processes, including data center power equipment, liquid cooling, data center CPUs, DRAM/NAND/HBM, optical communication/optical interconnection, high-performance Ethernet network infrastructure/data center DCI high-speed interconnection, transformers, gas turbines, etc.," especially in storage bases and optical interconnection systems, which are currently the biggest bottlenecks in data center computing systems. Huang Renxun's predictions about the development trends in the AI computing power field also form the fundamental logic for the increasingly optimistic market view of leaders in the AI optical interconnection industry chain like Marvell. He points out that AI has entered the era of "AI intelligent agents leading practical AI (Useful AI)," which means that token production can not only drive R&D experiments but also bring real profits. As AI tasks are split up and distributed, the demand for "massive interconnection" has surpassed the scope of single-node upgrades, with system-level expansion becoming the central focus of design. In their subsequent conversation, NVIDIA Corporation founder and CEO Huang Renxun and Marvell Technology, Inc. CEO deeply expounded the trend of the AI computing power industry chain upgrading to a core bottleneck of "connectivity" and revealed why Huang Renxun is extremely optimistic about Marvell and chooses to collaborate extensively and even directly invest $2 billion in strategic investments.