Ciena

Phase A — Understand the business

Lens 1 · Company Overview

The business in plain terms. Ciena is a network-technology company that sells hardware, software, and services that move data across networks — for cloud providers, telecom service providers, and other large network operators. The crown jewel is coherent optical technology (its WaveLogic modem family), which packs more bits onto a strand of fiber over longer distances at lower power. AI is the demand driver: training and inference clusters need enormous, low-latency bandwidth between data centers (data-center interconnect, "DCI") and increasingly inside them.

Four reporting segments:

1. Networking Platforms (77.1% of FY25 revenue) — the hardware. Two sub-portfolios: Optical Networking (68.1% of total — 6500 Packet-Optical, Waveserver, 6500 RLS reconfigurable line system, coherent pluggable transceivers) and Routing & Switching (9.0% — 3000/5000 service-delivery platforms, 8100 coherent routing, used for the "DCOM" out-of-band data-center management solution).
2. Platform Software & Services (7.6%) — Navigator NCS network-control software + related services.
3. Blue Planet Automation Software & Services (2.4%) — cloud-native OSS/orchestration software for service providers.
4. Global Services (12.9%) — install, maintain, consult.

Customers. Cloud providers (hyperscalers) and telecom service providers. Concentration is high and rising: five largest customers = 49.7% of FY25 revenue (up from 43.8% in FY24); one cloud provider = $851.6M (17.9%); AT&T = $500.7M (10.5%). Two cloud providers were in the top-five by revenue in FY25. By Q2-FY26, hyperscalers were ~1/3 of total revenue.

Suppliers. Asset-light manufacturing — Ciena owns the design/IP and outsources build to third-party contract manufacturers in Canada, Mexico, Thailand, Vietnam and the US. It carried $2.8B of purchase-order commitments to contract manufacturers/component suppliers as of Q2-FY26 (up from $2.1B at FY25 year-end) — a deliberate inventory build for the AI ramp in a "constrained supply environment."

Competitors. Core optical/networking: Nokia, Huawei, Cisco, HPE, ZTE. Emerging inside-the-datacenter rivals as Ciena pushes into interconnects: Marvell, Credo, Broadcom. Blue Planet software: Cisco, Nokia, Amdocs, ServiceNow, Netcracker, Ericsson.

Contract structure. A genuine risk factor in its own words: customer contracts "generally do not include minimum or guaranteed purchases" and a portion of quarterly revenue is "book-to-revenue" (ordered and shipped same quarter). No take-or-pay. Incumbency is re-won every cycle.

Lens 2 · Supply Chain

Map: upstream components → contract manufacturers → Ciena (design/IP/test) → channel/direct → end customer.

• Upstream inputs (named): Ciena designs the high-value pieces in-house — its WaveLogic coherent DSP/modem, lasers, modulators, optical amplifiers, wavelength multiplexers. It buys merchant silicon and optical components from the broader supply base. Its push "inside the data center" puts it into the same component orbit as Broadcom, Marvell, Credo (who are also competitors — supplier/rival duality the 10-K explicitly flags). The Nubis acquisition (Oct 2025, $270M all-cash) brought in-house IP for co-packaged optics (CPO), near-packaged optics, and active copper cables — a vertical move up the chain toward the GPU.
• Contract manufacturers (named geographies): facilities in Canada, Mexico, Thailand, Vietnam, and the United States. Ciena holds "control over supplier selection and commercial terms"; CMs procure components to Ciena spec.
• Distribution: global partner program — distributors, resellers, systems integrators, OEMs, ODMs — plus a large direct salesforce (GCE org, reorganizing in FY26 around global cloud/content, Americas, EMEA, APAC).
• End customers (named): AT&T (10.5% of FY25 rev), an undisclosed large cloud provider (17.9%), plus other hyperscalers and global telcos. Americas = 75.6% of revenue.

Chokepoints / single-source risk.

• WaveLogic DSP is the chokepoint that works in Ciena's favor — it is the proprietary asset the whole optical-networking franchise rests on; a stumble in DSP roadmap cadence (it's a process/IP race against Nokia's Infinera-derived DSP, Marvell, Acacia/Cisco) would be the most damaging internal failure.
• The constrained-supply environment is the chokepoint working against it — management repeatedly cites industry-wide component constraints inflating both backlog and lead times. The $2.8B purchase commitment is the company pre-buying its way through the bottleneck.
• Customer-side concentration is the real single-source dependency — losing or being designed out at the one ~18% cloud customer would be material.

Lens 3 · Competitive Advantages (moats)

The moat is real but narrow and contested.

• Coherent optical IP / WaveLogic — Ciena's durable edge. It is consistently positioned as the technology leader in high-performance coherent transport; Dell'Oro reported Ciena (and Nokia) gaining optical/DCI market share through the first nine months of 2025. ~2,400 issued patents, ~800 pending (as of Dec 2025).
• Switching costs — telco and hyperscaler optical networks are multi-year, mission-critical, and painful to rip-and-replace; incumbency is a genuine advantage within an installed base. But the 10-K is candid that incumbency must be re-won and contracts carry no minimums.
• Bargaining power: weak vs. customers, moderate vs. suppliers. With five customers at ~50% of revenue and a single cloud buyer at ~18%, the customers hold the whip hand — exactly the dynamic that compressed gross margin in FY25 as cloud mix grew (cloud buyers extract better pricing). Ciena's power over its outsourced supply base is "control over commercial terms" but it competes with its own would-be suppliers (Broadcom/Marvell/Credo) inside the data center.
• The contested frontier — co-packaged optics. Ciena is fighting to extend the optical moat inside the data center with CPO (Vesta 200 / 6.4T CPX) and is a co-founder of the Open CPX MSA alongside Microsoft and Marvell. This is the double-edged sword of the whole thesis: a winning standard expands Ciena's TAM, but an open, swappable socket standard is by design anti-moat — it lets hyperscalers mix-and-match optical engines.

Verdict on moat: strong in long-haul/metro coherent transport (where Ciena is a clear leader), structurally weaker the closer it gets to the GPU, where merchant-silicon giants (Broadcom, Marvell) and the hyperscalers' own roadmaps are the gravity.

Lens 4 · Segments

FY2025 operating-segment revenue and segment profit:

Geography (FY25): Americas $3,606,407 (75.6%, +22.2%), EMEA $731,912 (15.4%, +12.8%), APAC $431,188 (9.0%, +4.1%).

The trend is accelerating and concentrating. By Q2-FY26:

• Optical Networking: $1,099,848 (70.0% of total), +42.2% YoY.
• Routing & Switching: $174,230, +87.9% YoY — the fastest grower, driven by the 8100 + DCOM out-of-band data-center management ramp at a cloud provider.
• Total Networking Platforms: $1,274,078 (81.1% of total), +47.1%.
• Blue Planet declined −16.4% YoY — the one segment going backward (unified-assurance/analytics software softness). Software is not the growth story here; hardware is.
• Global Services +22.7% (implementation-led).
• Americas now 76.5% of Q2 revenue (+44.2%); APAC +71.7% (India service providers, Australia enterprise).

Cause: the entire acceleration is Optical + Routing hardware sold to hyperscalers for AI — RLS, Waveserver, coherent pluggables, and DCOM. The mix shift toward cloud is simultaneously the growth engine and the margin/concentration risk.

---

Phase B — Measure performance

Lens 5 · Earnings Result (latest print — Q2 FY2026, reported June 3 2026)

The single most important quarter in the story.

• Revenue $1,570,739K (+39.5% YoY) vs ~$1.1B prior-year Q2. Six-month revenue $2,997,781K (+36.4%).
• Beat: adjusted EPS $1.64 vs ~$1.40 consensus (+17% beat); +290% YoY on the adjusted line.
• GAAP diluted EPS $1.49 vs $0.06 prior-year Q2 (a 24× swing); six-month GAAP diluted EPS $2.52 vs $0.37. GAAP net income $218,220K vs $8,969K.
• Gross margin 44.0% vs 40.2% prior-year Q2 (+380bp), product GM 43.9% (+510bp on cost reduction + pricing optimization + mix); services GM 44.8% (−110bp).
• Operating leverage is the story within the story: total opex fell from 37.3% → 28.9% of revenue while GM rose — GAAP income from operations $237,871K (15.1% op-margin) vs $32,842K (2.9%) a year ago.
• Adjusted EBITDA ~$340M, roughly tripled YoY.
• Guidance RAISED: FY2026 revenue to $6.3B ±$100M (≈+32% YoY at midpoint); FY2026 adjusted operating margin ~19% (±50bp).
• Balance sheet: cash + investments $1.4B; H1 OCF $487.3M (vs $261M H1-FY25). Inventory provision of $42.5M for excess/obsolescence in H1-FY26 — worth watching as they pre-build.
• Market reaction: the stock FELL ~14% despite the beat-and-raise — classic "buy the rumor, sell the news" after a run to an ATH of $627 on June 2. The reaction says expectations had outrun even an excellent print.

Unusual vs. its own history: this margin profile (44% GM, ~19% guided op-margin) is above Ciena's prior cyclical peak (op-margin 14.5% in FY20) — management is claiming AI demand drives growth and structural margin expansion at once. That is the bull claim to stress-test (Lens 12/13).

Lens 6 · Earnings Calls (sentiment trend)

No transcripts/ on disk — sentiment assembled from `` call coverage.

• Tone arc (last several quarters): trough-anxiety (FY24) → cautious-optimism (early FY25) → outright-conviction (FY25 H2–FY26). Management's recurring frame in FY26 is "unprecedented" AI-driven demand, "orders significantly exceeding revenue," and "historically high backlog."
• Recurring phrases: WaveLogic leadership; "inside and around the data center"; co-packaged optics; supply-constrained environment; diversifying revenue / strengthening supply chain (added in FY26 in direct response to analyst pushback on concentration).
• What they started saying: CPO/Vesta, Open CPX MSA, hyperscalers as ~1/3 of revenue, a 19% adjusted-operating-margin target. What they stopped emphasizing: the broadband/PON growth narrative — in Q4-FY25 they ceased forward investment in 25G PON and cut headcount 4–5%, a clean strategic deprioritization of the consumer-broadband leg to concentrate on AI/cloud.
• On the Q2-FY26 call analysts pressed on customer concentration and the durability/pace of optical-transport monetization; management leaned on diversification and supply-chain answers. TD Cowen kept a $675 target but explicitly warned investors may be over-optimistic on the speed of conversion.

Lens 7 · Comps

Peer table — Ciena vs the AI-optical / networking complex. Multiples are `` with source/date or n/a. Numbers across sources varied widely (the complex is volatile and re-rating weekly); ranges shown where sources disagreed.

Read: Ciena trades at ~70x forward earnings and ~10.8x EV/sales — extreme for a hardware-heavy, cyclical, customer-concentrated networking vendor whose trailing GAAP P/E is a meaningless 83–173x (because earnings only just inflected). The only honest defense of the multiple is forward growth: on consensus FY26 adj EPS ~$5.44 the forward P/E compresses toward ~80x at $430, ~70x on out-year estimates. The peer complex is no cheaper — Lumentum ~159x, Coherent +97% YTD — so on a relative basis Ciena is not the most expensive optical name, but the entire group is priced for a multi-year AI-optical supercycle with no air-pocket. Mean-reversion risk is a sector risk, not a Ciena-specific one.

Lens 8 · Stock-Price Catalysts (5-year >5% movers)

• FY2021 — strength. Op-margins >13%, double-digit growth; stock strong into late 2021.
• 2022 — the −49.5% drawdown. From the 12/29/2021 high, CIEN fell ~49.5% during the inflation/supply-shock; FY22 saw flat revenue, collapsing margins, negative FCF as supply chain seized. The single most important historical lesson: this is a stock that can halve.
• 2023 — backlog-burn recovery. >20% revenue growth working off the swollen pandemic backlog; stock recovered.
• 2024 — relapse. Revenue −8.5%, op-margin to 4.8%, ROE to 3.0% — backlog normalized and the cycle rolled over again. Confirms the franchise is deeply cyclical, not a secular compounder.
• FY2025–H1 FY2026 — the AI melt-up. Order surge (backlog $2.1B → $5.0B), revenue re-acceleration, GAAP earnings inflection; stock ran to an all-time-high $627 (June 2, 2026) then −14% on the Q2 print (June 3).

What the tape reacts to: (1) the demand/backlog cycle (orders vs. revenue, lead-time commentary) above all; (2) guidance changes — even a beat-and-raise sold off when expectations were saturated; (3) margin trajectory (the 4.8%→19% swing is the re-rating fuel); (4) AI-optical sector beta — CIEN now trades as a high-beta AI-infrastructure proxy and drags/follows the optical complex (its −14% "dragged down the optical communications sector"). Earnings reactions are violent in both directions.

---

Phase C — Judge people & books

Lens 9 · Management

• Gary B. Smith — President & CEO since May 2001 (joined 1997), on the board since 2000. Track record: grew revenue from $361M (FY02) to $4.8B (FY25); the formative capital-allocation move was buying Nortel's Metro Ethernet Networks optical assets out of bankruptcy for $774M (2010) — the deal that made Ciena the optical leader it is (Dino DiPerna and Brodie Gage, today's R&D and Supply-Chain SVPs, both came in via that Nortel deal). A genuine operator who has navigated multiple brutal telecom cycles.
• Tenure & skin in the game: ~25 years as CEO — extreme tenure, both a strength (cycle-tested) and a governance flag (entrenchment, succession risk at age 65). Insider ownership is modest; insider-transactions.csv not present on disk, but `` shows Smith as a regular seller (multiple $1.1–1.4M open-market sales across 2025–26) — routine 10b5-1-style trimming, not alarming, but not founder-level alignment either.
• New CFO — Marc D. Graff (joined Aug 2025). Ex-CFO of Altera; before that CFO/COO of Intel's Data Center & AI Group. A pointed, on-thesis hire: a data-center-AI finance chief installed exactly as the AI-optical inflection arrives. Bullish signal on capital discipline + the inside-the-DC strategy.
• Capital-allocation history — the central knock. ROIC ~4.76%, below WACC — over a full cycle Ciena has not consistently earned its cost of capital. FY25 capital allocation: $334.5M buybacks, $231.1M Nubis, $140.8M capex against $806.1M OCF. Buybacks have been poorly timed in aggregate — H1-FY26 repurchases averaged $274.56/share, and the subsequent-events note shows buybacks at $561.86/share (May 2026) — i.e. the company was its own marginal buyer into the melt-up, near the $627 ATH. That is value-destructive repurchasing if the cycle turns.
• Founder vs. professional manager: professional manager, deeply tenured. Implication: steady cyclical execution and credible R&D stewardship, but the capital-allocation record (sub-WACC ROIC, pro-cyclical buybacks) is the reason to discount, not extend, the current multiple.
• Red flags: none egregious — comp appears standard; the restructuring/IPR&D abandonment and 25G PON exit read as rational pruning, not promotional behavior. The only real flags are entrenchment/succession and buyback timing.

Lens 10 · Forensic Red Flags

Forensic pass across income statement, balance sheet, cash flow. Overall: clean-to-average accounting; the real "red flag" is cyclical/quality-of-earnings, not fraud-risk.

• GAAP vs. non-GAAP gap. Q2-FY26 GAAP diluted EPS $1.49 vs adjusted $1.64; the bridge is stock-based compensation $55.5M in the quarter / $105.3M H1 (~3.5% of revenue) plus intangible amortization and acquisition costs. SBC is real and dilutive — unrecognized SBC of $394.8M to vest over ~1.5 years. The adjusted numbers flatter EPS by adding back a genuine economic cost; judge the franchise on the ~19% adjusted op-margin with eyes open on dilution. Diluted share count is roughly flat (~146M) because buybacks offset SBC issuance — but buybacks cost cash at $275–562/share.
• Inventory & purchase commitments. The aggressive pre-build is the balance-sheet item to monitor: $2.8B of non-cancelable-in-part purchase commitments and a $42.5M H1 excess-&-obsolescence provision. If AI orders air-pocket, this is where the writedown shows up first (exactly what happened in the 2022–24 cycle).
• Revenue recognition. "Book-to-revenue" same-quarter orders + no contract minimums = revenue is demand-following, not contractually locked; backlog ($5.0B) is a real signal but partly supply-constraint-inflated by the company's own admission — don't treat all $5.0B as clean future demand.
• Cash flow vs. earnings. Healthy — H1-FY26 OCF $487.3M tracks net income $368.5M + add-backs ($612M net income adjusted for non-cash charges). No suspicious divergence; receivables/inventory growth is consistent with a 40% revenue ramp.
• Debt: modest and well-structured. $1.15B Refinanced 2030 Term Loan (floating, 5.78% at FY25 end, $700M swapped to ~3.3–3.5% fixed) + $400M 4.00% Senior Notes due 2030 = ~$1.55B gross debt against $1.365B cash → net debt ~$185M. Negligible leverage.
• FY25 restructuring: $112.1M of impairments/restructuring (+356% YoY), "primarily the abandonment of an in-process R&D intangible" + the 25G PON exit / 4–5% headcount cut. A real cash/strategic event but disclosed and rational.

Regulatory findings (required).

• SEC Litigation Releases: none found naming Ciena (EDGAR EFTS, LR, 2021–2026).
• SEC AAERs: none found (EDGAR EFTS, AAER, 2021–2026).
• Non-SEC enforcement (web): no material FTC/DOJ/FDA/CFPB enforcement, consent decrees, or fines surfaced for Ciena in the search window.
• 10-K Item 3 / 10-Q Item 1 (Legal Proceedings), own disclosure: ordinary-course only — "various legal proceedings, claims, and other matters arising in the ordinary course of business, including … employment, commercial, tax, and other regulatory matters … intellectual property related claims … Ciena does not expect that the ultimate costs … will have a material effect." No material litigation disclosed.
• Conclusion: No material regulatory or legal findings — verified via SEC EDGAR EFTS (LR, AAER), web search, and 10-K/10-Q legal-proceedings disclosure as of 2026-06-24.

---

Phase D — Project & stress-test

Lens 11 · Forward Projection (FY2026 → FY2028 EPS)

Built bottom-up from the latest actuals + raised guidance. Adjusted (non-GAAP) EPS basis to match how the Street and the company frame it; every input labeled; output ``.

Anchors:

• FY2026 revenue guide $6.3B ±$100M; FY2026 adjusted op-margin ~19%.
• H1-FY26 GAAP diluted EPS already $2.52; Q2 adjusted EPS $1.64; consensus FY26 adjusted EPS ~$5.44 (range $4.92–$5.99).
• Diluted shares ~146M, roughly flat (buyback ≈ SBC issuance).

The forecast that matters is not the EPS line — it is whether AI optical demand is a level shift or a cycle. The bull case extrapolates the 4.8%→19% op-margin as a new plateau; the bear case is that Ciena has done this exact "swollen backlog → euphoria → normalization → writedown" round-trip inside the last four years and the only thing different is the AI label on the orders.

(Per --watchlist rules, no forecast.ts create is logged in this unattended sweep. The base case to log on a future interactive pass: CIEN FY26 non-GAAP EPS >= $5.40, p≈0.60, resolves 2026-11-01.)

Lens 12 · Bull vs Bear

Bull case. Ciena is the public-market optical pure-play on the single biggest infrastructure buildout in tech history. AI clusters are bandwidth-bound; coherent optics (between DCs) and now CPO/pluggables (inside DCs) are the only way to feed them, and Ciena owns the leading coherent IP (WaveLogic) plus a credible inside-the-DC strategy (Nubis, Vesta 200, Open CPX with Microsoft/Marvell). The proof is in the print: revenue +40%, GM 44%, op-margin guided to 19% — growth and margin expansion simultaneously, with backlog at $5.0B and orders still exceeding revenue. Dell'Oro confirms share gains. A new AI-fluent CFO (ex-Intel DCAI) signals discipline. If the AI-optical TAM compounds (Lumentum/Coherent peg AI optics at ~$90B by 2030), Ciena's revenue and earnings have years of runway and today's ~70x forward multiple de-rates into the growth.

Bear case (permanent-impairment risks).

1. Hyperscaler in-sourcing / commoditization inside the DC. NVIDIA ships its own silicon photonics (Spectrum-X/Quantum-X); Broadcom and Marvell own the merchant-silicon roadmap; the Open CPX MSA Ciena helped create is designed to make optical engines swappable. The closer the value migrates to inside-the-DC, the more Ciena competes with its own suppliers and the hyperscalers' own teams — moat erosion at exactly the frontier the bull case needs to win.
2. Customer concentration + pricing power loss. Five customers ≈ 50%, one cloud buyer ≈ 18%, hyperscalers ≈ 1/3. Hyperscalers are the toughest buyers on earth; the FY25 gross-margin dip from cloud mix is the canary. Lose or get designed out at one mega-customer and the model breaks.
3. It is a deeply cyclical business wearing a secular-growth multiple. Op-margin 14.5%→4.8% and a −49.5% drawdown within the last four years. Backlog is partly supply-constraint-inflated by management's own admission. When the AI capex pulse normalizes, the same "double-order → air-pocket → inventory writedown" dynamic that hit FY22–24 recurs — only this time off a far higher valuation base.

Pre-mortem (18 months out, thesis broke). It's late 2027. Hyperscaler optical orders decelerated as the first AI-capex digestion arrived; backlog that was supply-inflated converted to cancellations/pushouts; cloud customers used their leverage to claw back pricing as supply loosened; Ciena took an inventory writedown on its $2.8B pre-buy; CPO standardized into a low-margin swappable socket where Broadcom/Marvell win the volume. Op-margin reverted to mid-teens, EPS missed, and a ~70x forward multiple compressed to ~20x — a >60% drawdown even with revenue still above 2024 levels. Nothing in that scenario requires fraud or a bad management team — only the cycle doing what it has always done.

Is the multiple too high? Yes, on any historical or normalized-margin basis — ~70x forward / ~10.8x EV/sales / ~57–97x EV/EBITDA prices a permanent supercycle onto a customer-concentrated, sub-WACC-ROIC, cyclical hardware vendor. It is defensible only if you underwrite AI optical demand as a multi-year level-shift with no digestion — which is a bet on the macro, not on Ciena's idiosyncratic quality.

Contrarian view (what the market refuses to see). The bull crowd treats Ciena as a secular AI compounder; the tape (a −14% drop on a beat-and-raise, a stock that round-tripped from $627) is quietly telling you it's a high-beta cyclical at a top. The contrarian read: the business is genuinely better than it has ever been and the stock is a poor risk/reward here — both can be true. The asymmetry favors waiting for the cycle to exhale.

Lens 13 · Devil's Advocate (short-seller)

Dismantling the bull case.

• What structurally breaks the model: the optical value chain migrating inside the data center, where the economics belong to merchant silicon (Broadcom/Marvell) and the hyperscalers' own designs, not to a transport-systems vendor. Ciena's coherent-transport moat is real between data centers; it is a challenger, not an incumbent, inside them — and that's where the bull case needs the growth.
• Revenue concentration → fragility: ~1/3 of revenue from hyperscalers, one customer ~18%. These buyers dual-source by policy and design out incumbents when it suits them. A single roadmap decision at one hyperscaler is a double-digit-percent revenue event.
• Moat weaker than bulls think: Ciena is the #3 supplier in pluggable coherent 400ZR/ZR+ behind Cisco/Acacia and Marvell (LightCounting). Its leadership is in systems/long-haul, not the pluggable/inside-DC volume game the AI narrative is pricing.
• Most dangerous underestimated competitor: Marvell — custom silicon + DSP + optics, guiding to $11B FY27 revenue, and a supplier-turned-rival sitting directly in Ciena's inside-the-DC path. Plus NVIDIA's own photonics, which can simply absorb the function.
• Worst capital allocation: buying back stock at $275 and then $562/share into a parabolic move, with ROIC already below WACC — the company is the marginal bid at the top.
• Assumptions that must hold for today's price: (1) AI optical demand is a level-shift, not a pulse; (2) ~19% adjusted op-margin is a new floor, not a cycle peak; (3) hyperscalers don't compress pricing as supply normalizes; (4) CPO/inside-DC becomes a Ciena profit pool, not a commoditized socket. All four must hold. The base rate on "this cyclical's peak margin is permanent" is poor.
• If growth disappoints 20–30%: at ~70x forward, a growth scare + margin reversion is a >50% drawdown — and Ciena has printed a −49.5% drawdown within living memory of current holders.
• Single permanent-impairment scenario & plausibility: hyperscalers standardize on open, swappable optical engines (the MSA Ciena co-signed) sourced from merchant silicon, relegating Ciena to lower-margin transport — plausibility: moderate, and it's a slow bleed rather than a cliff, which is exactly how moats actually erode.

Lens 14 · Management Questions (15, ordered by information value)

1. AI-driven orders: how much of the $5.0B backlog is firm demand vs. supply-constraint double-ordering, and what is your framework for distinguishing the two before a digestion hits?
2. Is the ~19% adjusted operating margin a structural floor or a cycle peak? What revenue level and mix would push it back toward the mid-teens you ran at in FY24?
3. As optics move inside the data center, what is your durable profit pool when the Open CPX MSA you co-created makes optical engines swappable across vendors?
4. Your single largest cloud customer is ~18% of revenue and hyperscalers are ~1/3. What concretely de-concentrates the book over the next 8 quarters, and on what timeline?
5. You bought back stock at ~$275 and ~$562/share while ROIC sits below WACC. Defend that as superior to retaining capital — what return hurdle governs repurchases?
6. Against Broadcom, Marvell, and NVIDIA's in-house photonics, where do you win inside the data center, and where do you concede?
7. You are the #3 pluggable coherent (400ZR/ZR+) supplier behind Cisco/Acacia and Marvell — what changes that ranking, and is pluggable volume even where you want to compete?
8. What did Nubis actually buy you in IP/time-to-market, and what is the revenue/margin contribution path for co-packaged optics?
9. With $2.8B in purchase commitments and a $42.5M H1 obsolescence provision, what is your inventory-risk plan if AI orders air-pocket the way the pandemic backlog did in FY22–24?
10. Gross margin dipped in FY25 on cloud mix, then rose in FY26 on cost/pricing/mix — which effect dominates from here as cloud becomes a larger share?
11. Succession: Gary, you've led for ~25 years — what is the concrete CEO-succession plan and timeline?
12. Blue Planet declined ~16% YoY — is software a strategic pillar or a candidate for divestiture/deprioritization like 25G PON?
13. How exposed is the model to tariffs / China (Huawei, ZTE) / export controls, given contract manufacturing in Mexico/Thailand/Vietnam?
14. What share of revenue is recurring software/services vs. one-time hardware, and what is the multi-year target mix?
15. How do you think about capital returns vs. M&A if the stock is at 70x forward — does the bar for buybacks rise with the multiple?

---