Simon Leopold

Sherri reaffirmed the greater-than-42% gross margin target, noting the Q1 guide midpoint of 40.5% and that the bulk of 6-inch and new-product ramps are still ahead.

Jim called the report speculative but said Coherent would benefit as the largest US transceiver supplier, citing 20+ US production facilities.

Anderson declined to confirm the specific milestone but emphasized 1.6T ramp is faster than expected and is a good portion of sequential growth.

Anderson pushed back on the gap narrative, citing doubled OCS market forecast to over $4 billion, differentiated technology, resolved production bottleneck, and $15 billion CPO opportunity with NVIDIA partnership.

Jim confirmed both 800G and 1.6T growing with 1.6T accelerating from smaller base. Ramping with multiple customers. Initial ramp EML and silicon photonics, VCSEL-based 1.6T in second half of calendar 2026. Did not provide specific revenue milestones.

Jim described very active scale-up engagements across multiple customers. Scale-up CPO opportunity expected to dwarf scale-out by orders of magnitude. Customer forecasts are very large. Timeline sooner than expected based on customer plans.

Jim confirmed strong OCS positioning with nonmechanical liquid crystal technology. Revenue and backlog grew sequentially. More meaningful revenue contribution expected calendar 2026, weighted toward second half.

Jim directly corrected the narrative, confirming 6-inch InP ramping production of EMLs, CW lasers, and photodiodes across both Sherman TX and Jarfalla Sweden facilities.

Jim clarified industrial down (partly $20M A&D exit, partly macro caution) while data center and comms up sequentially. OCS pipeline healthy with strong multi-customer engagement.

Jim corrected VCSEL perception: majority of transceiver revenue is EML-based, with majority using internally manufactured EMLs. Silicon photonics also growing portion.

Jim reported strong and expected 800G demand with no signs of customer inventory buildup, noting customers deploy transceivers very quickly after shipment.

Jim provided thorough response: CPO expands TAM by converting electrical-to-optical connections, Coherent well-positioned through ingredient technologies applicable across form factors.

Jim confirmed well-diversified datacom revenue across multiple hyperscalers, noted continued expansion of 800G customer count as a positive indicator.

Jim confirmed growth from both 800G and sub-400G, with customer base for 800G significantly broadened from a couple to many customers year-over-year.

Sherri outlined organic growth investment as top priority, deleveraging as close second. Paid down $118M in Q1 to reach 2.4x leverage.

Scott affirmed coherent moving inside the data center (intercept at 3.2 terabits) and enthusiasm for OCS, but declined to comment on whether Ciena will enter OCS.

Question posed by Jeff Cocci (Raymond James associate) standing in for Simon Leopold ("Jeff Cocci in for Simon"); attributed to the covering analyst. Scott confirmed web-scale deals carry less off-box software but said the value extends well beyond hardware, on-box software, planning, deployment and global service.

Marc framed pricing within 'value exchange', confirming pricing is on the table across products alongside supply security and working-capital terms.

Mark disclosed AT&T as one 10% customer, declined to name the two cloud providers. Three collectively covered just under 44% of Q4 revenue. Full year: one cloud and AT&T at ~28% combined.

Gary expects initial revenue from all three scale across hyperscalers in FY2026, with significant ramp in FY2027-2028. All currently US-centric. Infrastructure requires massive fiber buildouts which take time.

Gary distinguished GPU cluster DCI as shorter distance, dedicated for AI training, requiring super high speed and low latency at scale. Uses RLS and 800G plugs.

Scott identified three routing drivers: service provider spending recovery, emerging coherent routing opportunities in aggregation networks, and DCOM inside data center applications.

Gary pointed to both breadth (more cloud providers engaging) and depth (new applications like GPU clusters and in-DC management). Scott confirmed both wins have applicability to other customers.

Jim provided exact percentages. Scott later added that top five customers were 45% of revenue. Named AT&T as the service provider 10% customer.

Jim listed manufacturing locations but declined to quantify Mexico exposure, citing fluidity. Scott added that Ciena owns all test assets, enabling production shifts across locations.

Jim and Gary explained cloud prefers direct but speed constraints push them to service providers for Mofen. Clear and comprehensive.

Jim stated concentration pattern will remain similar with rotating 10% customers from the same set of two service providers and two cloud providers.

Michael said repricing was modest and LTAs are forward-looking; mix was the larger driver, with room left on gross margin from both.

Michael confirmed OCS backlog well past $400M with most shipping in calendar H2 2026 into 2027. Calendar Q4 expected quite a bit above previously targeted $100M. Broad-based across three customers with increasing demands. Price increases having modest impact on revenue and more meaningful impact on margin.

Michael confirmed $100M quarterly OCS target by December 2026 with increasing confidence. Growing number of use cases from customers and potential customers. Engagement level described as super high.

Michael ranked transceivers as the biggest sequential driver but emphasized the surprisingly broad-based demand across pump lasers, narrow linewidth, coherent components. Every segment contributing. Width of demand was more surprising than any single driver.

Michael outlined OCS ramp: current through December quarters still ramping, first inflection in early 2026, more meaningful inflection in H2 2026. Currently limited by manufacturing capacity being built out.

Michael cited two competitive moats: unique power level and reliability leveraged from undersea laser and Raman pump heritage. Currently sole-sourced with early shipments underway.

Wajid explained tariff impact from component costs (sub-mounts, capacitors). Moving DCI production from China to Thailand and eventually Dongguan cloud module production as well. Impact expected to diminish in FY Q1/Q2.

Wupen positioned CPO ramp volume for H2 CY2026. In addition to lasers, exploring other optical chips for CPO ecosystem. Near-term focus remains on cloud module and laser component ramp.

Alan declared inventory discussion over, now shipping what customers are shipping out. DCI and transport demand very strong. Not returning to pandemic peak levels but seeing genuine demand.

Wajid explained CPO solves electrical scaling beyond 200G per lane. Lumentum sees it as additive since optical content doesn't decrease and their modest transceiver share means no cannibalization risk.

Alan outlined component-level participation via 200G EMLs ramping summer 2025, plus transceiver-level with positive customer feedback on samples. Timing depends on customers.

Alan confirmed Cloud Light business tracking as expected with no surprises, consistent with prior two earnings calls.

Robbins argued 2023 flush was overstated in perception. Billions in pre-Cat 9K installed base was never upgraded. Cat 4K and 6K end-of-support is driving real replacement demand.

Patterson noted EMEA and APJC public sector grew mid-to-upper teens. Robbins cited European defense spending, sovereignty initiatives, and geopolitical drivers as ongoing catalysts.

Robbins expects the two-thirds systems, one-third optics mix to remain relatively stable based on current customer conversations.

Robbins denied any pervasive pull-forward. Patterson provided data: channel partner feedback, linearity normal, activation timing unchanged, no future-dated ship requests spike.

Robbins acknowledged 2017 comparison is difficult due to subscription transition. Emphasized the campus refresh will be driven by embedding security into the network for agentic AI.

Herren confirmed Q4 guide assumes reciprocal tariffs resume on July 9 with USMCA and semiconductor exemptions intact. Acknowledged difficulty predicting final outcome.

Robbins explained Cisco's flexible go-to-market approach with webscalers, selling systems, standalone optics, or silicon depending on customer preference.

Robbins directly addressed federal exposure size and DoD concentration, noted civilian-side impact and no material change expected in outlook.

Chuck noted 1,500 more customers added to Splunk prospect list. Hundreds of Cisco partners trained and building Splunk practices. Scott added Splunk tracking ahead on profitability.

Chuck described partnership as early stage. Hyperfabric shipping early calendar 2025. New Nvidia-based AI compute platform and AI PODs announced this quarter.

Wissam said seasonality is typical once the extra Q1 week (~$65M) is removed, with mid-single-digit growth across Q2-Q4.

George described rational industry where vendors take different approaches (some raise list/maintain discount, others restrict discounting) but all face similar cost structures. Competition comes down to software value, portfolio breadth, and operational discipline. HFAs seeing significant interest from customers.

George confirmed long-standing sovereign cloud relationships expanding into AI with AFX enabling broader participation. Advantages in security, multi-tenancy, data protection, and hybrid capabilities play well in enterprise AI workloads. More innovation coming.

George cited differentiated capabilities (security, multi-tenancy, data protection, hybrid) as competitive advantages in sovereign AI. AFX introduction driving strong interest from AI cloud providers pursuing new use cases.

George outlined four AI technology pillars: high-performance/scale storage (strong today), cloud-equivalent storage for hyperscaler tools (more work with hyperscalers coming at Insight), data management tools (search, organize, vectorization, RAG readiness, guardrails, coming at Insight), and ecosystem expansion (reference architectures with NVIDIA and hyperscaler tools).

George described normal Q4 linearity in Europe despite Easter in final week (planfully managed). Caution manifesting as deal downsizing and extra procurement approval cycles rather than deal slippage.

George confirmed NetApp currently under semiconductor exemption but declined to speculate on Section 232 outcomes.

George confirmed minimal China dependency, Mexico flexibility, no tariffs in Q4 guide, and industry-wide impact if broad tariffs applied.

Mike focused on flash mix as the driver of gross margins rather than hardware/software split.

George noted broad-based AI wins with more concentration in public sector, manufacturing, financial services, and healthcare.

Locoh-Donou rejected both premises, said software is tracking the plan set at the start of the year, and enumerated sovereignty, displacement and consolidation as hardware drivers.

Francois confirmed F5 is now in NVIDIA's reference architecture. Third-party tests validated 30-40% more token generation per GPU. Multiple POCs and trials underway. Customer sophistication still evolving, first priority is getting AI factories running, then optimizing efficiency. Inference shift will accelerate adoption.

Francois confirmed $60 million in H1 from customers recommitting to hardware after previously stopping. Drivers include data center modernization, performance needs, and AI data delivery. But broader trend is hybrid multi-cloud driving both hardware and software demand with customer flexibility.

Francois declined to break out AI revenue formally but indicated it moved from single-digit millions quarterly in FY25 to double-digit millions quarterly. Runtime security seen as multi-billion-dollar market in early innings. Cooper noted mid-single-digit systems price increase from January 2025 still flowing through.

Francois confirmed Q1 guidance assumes Federal disruption from government shutdown with hope for normalization over the year. Cooper declined to guide FY26 software/hardware mix given only 12 days since incident disclosure. Follow-up clarified breach affects BIG-IP equally across both hardware and software (virtual edition) deployment models; NGINX and DCS were not impacted.

Cooper explained services deceleration to 1% was due to last laggard customers retiring prior-family units from maintenance base. Now behind them. Short-term deferred revenue up 5% and total deferred up 10% indicate services growth should improve in FY26.

Francois cautioned against reading too much into single-quarter service provider trends. 5G promise hasn't fully materialized. Carriers had been sweating assets but relaxed somewhat recently. Dynamics not fundamentally changing in the segment.

Frank confirmed enterprise was the upside driver, exceeding expectations. U.S. Federal met expectations on a dollar basis. Service provider segment remains weak with broad demand constraints.

Cooper confirmed the first-half outlook improved from flat to low-single-digit growth, driven by better software expansion from renewals and mid-single-digit systems growth expectations.