Ruben Roy

Marc said this year's opex includes one-time variable comp being reinvested, plus added investment in line systems and interconnects for fiscal 2027.

Scott said the alternative consumption model is bespoke to individual customers, with 2 examples today, and is custom-developed rather than productized.

Gary expects multi-rail to start in '27 and ramp linearly across multiple hyperscalers and large service providers; Marc noted $0 Hyper-Rail revenue until later this year.

Gary confirmed each of the three hyperscaler scale across wins is hundreds of millions. Different business models drive different architectures but all need training backbones. Initial hyperscaler expanding to more sites.

Mark confirmed all new opportunities are included in the $5.7-6.1B FY2026 guide. In and around data center revenue tripling from low single digits percent of revenue in FY2025 to low double digits in FY2026.

Gary confirmed no divestiture planned, just ceasing future 25G/100G PON development. Existing broadband products will continue to be sold and supported. PON technology remains relevant for DCOM.

Scott explained the first GPU cluster win required performance optics (800G WL6 Nano) not Coherent Lite due to distances. Coherent Lite revenue still expected from FY2027. Both are on the menu.

Gary positioned Blue Planet's inventory confederation as a critical prerequisite for AGENTIC AI in service provider networks. Strong relationships and architecture for future AI-driven management.

Scott confirmed Ciena's combined line system and pluggable portfolio positions them for outsized early share, though market will eventually support multiple vendors.

Gary confirmed Europe remains the primary Huawei replacement opportunity, a multi-year tailwind. Ciena is winning more than its fair share of these deals.

Scott positioned CPO as repackaging of IMDD technology for scale, not a threat to coherent. Ciena sees opportunity to play both sides: coherent inside data centers and component IP for IMDD scale problems.

Scott differentiated by segment: core long-haul DCI remains dominant, and Ciena is winning early competitive bids for metro DCI with 800G plugs. Share gains continuing.

Scott explained customers currently prefer finished goods (systems or plugs) over disaggregated components. Component consumption pattern is expected to grow as coherent enters data centers.

Wall described systems and integration investment and prudence in expense expectations but did not split core versus acquired spend.

Wall said organic incremental margins now closer to 60%. Hexagon follows BETA integration pattern, dilutive in year 1, improving profitability within 12-18 months.

Anirudh said Cadence builds core physics solvers, works with accelerated compute (NVIDIA GPU porting for years). Omniverse is a deployment platform that runs Cadence tools. Neutral strategy focused on building the best engineering software.

Anirudh described JedAI platform with standardized and customer-specific components. On-prem and cloud deployment. RTL development and verification plan generation are key areas. Optimistic agentic AI can automate remaining manual design workflow.

Anirudh cited chiplet architectures, multiple advanced foundries, improved portfolio with Rambus HBM4 and others. Expects IP to grow faster than Cadence average long-term.

Anirudh said both customer demand and opportunities drive IP acquisitions. At least four major foundries at advanced nodes. Expects IP to grow better than Cadence average.

John Wall attributed flat China assumption to general prudence rather than specific Entity List changes.

Anderson confirmed NVIDIA as lead CPO customer but expects multiple other customers to follow, with engagement across a wide set of accounts.

Anderson confirmed capacity allocated to maximize margin dollars, and multi-rail has higher gross margin structure with high-margin components relying on indium phosphide.

Jim confirmed 1.6T ASPs higher than 800G with higher gross margins especially early in lifecycle. Views 1.6T ramp as margin accretive. 6-inch InP also a gross margin driver for both 800G and 1.6T.

Jim views CPO as additive, not cannibalistic. Pluggables remain dominant in scale-across and scale-out through the decade. CPO big growth driven by scale-up, which is all incremental TAM since 100% electrical today.

Jim confirmed initial adoption in redundancy and spine switch applications. Surprised by broadening applications including OCS in scale-up networks and DCI networks, suggesting TAM significantly larger than initially expected.

Jim outlined philosophy of insourcing only where genuine differentiation exists, leveraging ecosystem otherwise, applying same logic to supply chain verticalization.

Jim confirmed both DCI and traditional telecom transport showed improvement, but maintained cautious stance pending a couple more quarters of recovery.

Michael described mostly 3-year, largely take-or-pay agreements with pricing levers, and pegged pump-laser share at roughly 70-80%.

Michael confirmed three customers contributing to the $400M+ backlog. Run rate entering 2027 well above the previously outlined $100M exit. Wupen detailed four dominant applications: spine replacement, scale-across, optical scale-up, and network protection/redundancy, all seen across customers to varying degrees.

Michael confirmed both 800G and 1.6T doing better than expected. 1.6T accelerating faster with customers pushing for earlier deliveries. Wupen noted higher market share in 1.6T due to improved development execution, creating stronger growth trend even as 800G continues.

Alan clarified constraints are in telecom components (hermetic packages), not transceivers. Suppliers shuttered capacity during downturn and can't ramp fast enough. Impact in both March and June quarters.

Alan confirmed 200G EMLs designed for both 800G and 1.6T, plus sampling differential 200G EMLs with positive feedback. Also developing high-power CW lasers for silicon photonics market.

Greg attributed the gap to Q4 seasonality being the highest-margin quarter and reaffirmed 6%+ for next year without specifics.

Mike reiterated the asset-light model but declined to detail structure given no definitive framework with Adani yet.

Mikros acquisition described as game changer, acquired design and engineering team for chip-to-infrastructure liquid cooling technology. Hanley adds deployment, installation and maintenance services for switchgear and power distribution. Together enables vertical integration from manufacturing to deployment.

ATE continues to outperform driven by DRAM stacking and high-bandwidth memory demand. WFE showing signs of improvement from AI compute expansion and NAND upgrades, not yet in guidance but represents potential upside.

CEO differentiated: ANSYS China performing well (broader unrestricted market) while Synopsys Classic declining due to IP and advanced EDA restrictions. EDA share loss only to local alternatives for restricted customers. FY26 guide assumes environment stays same.

CEO detailed all three factors: BIS impact exceeded 6-week restriction period, foundry customer investment not yielding returns, and roadmap resource decisions redirected. Merged two engineering teams to address scaling.

CFO attributed margin decline to IP revenue shortfall in a resource-intensive business. ANSYS actually came with higher operating margin. Long-term mid-40s commitment unchanged.

CEO explained customers can burst EDA workloads to cloud, which provides a good burn rate on software as bursting means more usage and monetization.