COHR
Coherent Corp.
AI Foundry & Connectivity · Optical Interconnect
~$350
Mkt Cap ~$68–74B
52W: $73.66 – $384.85
Next earnings: ~12–14 Aug 2026 (Q4 FY26)
↘ Add on Pullback
Entry zone: $290–$330 · Ideal: $240–$275
The market prices COHR as a transceiver manufacturer in a transient 800G upgrade cycle; it is actually the only vertically integrated optical supplier that simultaneously owns the laser chips, transceiver modules, optical circuit switches, and co-packaged optics components for the NVIDIA Blackwell → Rubin AI scaling roadmap — making it structurally irreplaceable through the end of the decade regardless of which CPO architecture wins.
TTM P/E (GAAP)
~165×
EPS ~$2.09 TTM 1
Fwd P/E (NTM)
~45×
PEG ~1.2 2
Rev YoY (Q3 FY26)
+20.5%
DC&C segment +40.6% 1
Non-GAAP GM
39.6%
+105 bps YoY, expanding 1
FCF Margin
~–21%
Negative — CapEx surge 3
Net Leverage
0.5×
vs 2.1× a year ago 1
Thesis Durability
A
Variant perception: Market treats COHR as a pluggable transceiver vendor exposed to CPO disruption; in reality, COHR supplies the indium-phosphide CW lasers that every CPO architecture requires — including NVIDIA's own CPO programme. NVIDIA's $2.00B equity stake and multi-year supply agreement through end of decade is the single clearest proof that this is a platform relationship, not a cycle trade.4
TAM: Transceiver addressable market >$25B by 2028 (Cignal AI); CPO component TAM management raised to >$15B; OCS TAM >$4B.1 OCS pipeline: 10 hyperscaler customers confirmed as of the FQ2 (Dec 2025) earnings call; non-mechanical liquid crystal technology cited by management as a differentiated edge over mechanical switching alternatives.9 Five-year backlog with LTAs extending to 2030+ and customer-funded CapEx embedded in LTAs.
Structural tailwinds: AI cluster interconnect bandwidth doubles with every GPU generation; 800G → 1.6T → 3.2T is a fixed multi-year ramp. Coherent is simultaneously ramping 800G, shipping first 1.6T units, and executing initial CPO scale-out revenue in H2 CY2026.
Moat: 6-inch indium-phosphide fab in Sherman, TX (only commercial 6" InP facility in the US): the transition from 3-inch to 6-inch wafers produces 4× the chip output at half the unit cost — a manufacturing step-change that 3-inch peers cannot match without multi-year retooling.9 Vertical integration (chip → module → system) creates the only one-stop-shop for AI optical infrastructure at scale.
Bear stress-test: CPO becomes the dominant architecture and hyperscalers source lasers directly from chip fabs, bypassing COHR's module business. Not survivable in its current form — but COHR supplies the CW lasers that go into CPO, meaning it benefits from CPO adoption rather than being displaced by it. The real bear case is NVIDIA on-boarding a second InP supplier at scale, which requires 2–4 years of qualification at minimum.
TAM: Transceiver addressable market >$25B by 2028 (Cignal AI); CPO component TAM management raised to >$15B; OCS TAM >$4B.1 OCS pipeline: 10 hyperscaler customers confirmed as of the FQ2 (Dec 2025) earnings call; non-mechanical liquid crystal technology cited by management as a differentiated edge over mechanical switching alternatives.9 Five-year backlog with LTAs extending to 2030+ and customer-funded CapEx embedded in LTAs.
Structural tailwinds: AI cluster interconnect bandwidth doubles with every GPU generation; 800G → 1.6T → 3.2T is a fixed multi-year ramp. Coherent is simultaneously ramping 800G, shipping first 1.6T units, and executing initial CPO scale-out revenue in H2 CY2026.
Moat: 6-inch indium-phosphide fab in Sherman, TX (only commercial 6" InP facility in the US): the transition from 3-inch to 6-inch wafers produces 4× the chip output at half the unit cost — a manufacturing step-change that 3-inch peers cannot match without multi-year retooling.9 Vertical integration (chip → module → system) creates the only one-stop-shop for AI optical infrastructure at scale.
Bear stress-test: CPO becomes the dominant architecture and hyperscalers source lasers directly from chip fabs, bypassing COHR's module business. Not survivable in its current form — but COHR supplies the CW lasers that go into CPO, meaning it benefits from CPO adoption rather than being displaced by it. The real bear case is NVIDIA on-boarding a second InP supplier at scale, which requires 2–4 years of qualification at minimum.
Structurally anchored. NVIDIA supply agreement through end of decade, record backlog into 2028, 6-inch InP advantage, and CPO agnosticism give this thesis the highest structural durability in the foundry/connectivity layer.
Business Quality
B+
Margin trend (non-GAAP GM): Q3 FY25 38.5% → Q4 FY25 38.9% → Q1 FY26 38.9% → Q2 FY26 39.0% → Q3 FY26 39.6% → Q4 guide midpoint 40.0%. Five-quarter uptrend; long-term target mid-40s%.1
EPS trend (non-GAAP): $0.91 → $1.00 → $1.16 → $1.29 → $1.41 over five sequential quarters. Q4 guide midpoint $1.62 (+14.9% QoQ).1 Estimate revisions: UP — avg target raised +70.68% in past 3 months per ChartMill.5
FCF caveat: Nine-month FY26 operating cash flow only $10.1M (vs $503.3M prior year). Q3 CapEx $290M (up from $112M a year ago) as Coherent doubles 6-inch InP capacity one quarter ahead of plan. FCF is deliberately negative and customer-funded LTAs justify the build — but this is not a self-funding compounder yet.3
Balance sheet: Net leverage cut to 0.5× EBITDA from 2.1× in Q3 FY25; $1.59B cash + $825M short-term investments; $3.18B long-term debt. The $2.00B NVIDIA infusion funded most of the capacity build.1
Form 4 insider activity (90 days to May 2026):
· CFO Sherri Luther: 2,000 shares sold ~$351 (~24 Apr), Rule 10b5-1 plan. Routine.
· Director Howard Xia: 4,240 shares sold $236–$258 (16 Mar-17) via option exercise + same-day sale. Routine.
No discretionary open-market buying; insider ownership only 0.71% of shares. Neutral — not a red flag, but no smart-money reinforcement either.6
Institutional: 80.46% institutional ownership; Ken Fisher expanding stake; Artisan Mid Cap trimmed in Q1 2026 on valuation (did not exit). Short interest: 3.34% of shares out — not elevated.2,7
Customer concentration risk rising: NVIDIA will likely become a >10% customer in FY27 post supply agreement; not yet disclosed separately. Flag for FY26 10-K.
EPS trend (non-GAAP): $0.91 → $1.00 → $1.16 → $1.29 → $1.41 over five sequential quarters. Q4 guide midpoint $1.62 (+14.9% QoQ).1 Estimate revisions: UP — avg target raised +70.68% in past 3 months per ChartMill.5
FCF caveat: Nine-month FY26 operating cash flow only $10.1M (vs $503.3M prior year). Q3 CapEx $290M (up from $112M a year ago) as Coherent doubles 6-inch InP capacity one quarter ahead of plan. FCF is deliberately negative and customer-funded LTAs justify the build — but this is not a self-funding compounder yet.3
Balance sheet: Net leverage cut to 0.5× EBITDA from 2.1× in Q3 FY25; $1.59B cash + $825M short-term investments; $3.18B long-term debt. The $2.00B NVIDIA infusion funded most of the capacity build.1
Form 4 insider activity (90 days to May 2026):
· CFO Sherri Luther: 2,000 shares sold ~$351 (~24 Apr), Rule 10b5-1 plan. Routine.
· Director Howard Xia: 4,240 shares sold $236–$258 (16 Mar-17) via option exercise + same-day sale. Routine.
No discretionary open-market buying; insider ownership only 0.71% of shares. Neutral — not a red flag, but no smart-money reinforcement either.6
Institutional: 80.46% institutional ownership; Ken Fisher expanding stake; Artisan Mid Cap trimmed in Q1 2026 on valuation (did not exit). Short interest: 3.34% of shares out — not elevated.2,7
Customer concentration risk rising: NVIDIA will likely become a >10% customer in FY27 post supply agreement; not yet disclosed separately. Flag for FY26 10-K.
Operating metrics are clean and trending up; FCF is the only blemish — and it is a deliberate capacity-investment artefact, not a structural issue.
Entry Price Discipline
C+
Current vs sector: Forward P/E ~45× vs semiconductor sector median ~34× = ~32% premium. P/S ~9× vs peers at 4–8×. COHR's own 5-year average P/E is ~44×, so the current multiple is not anomalously elevated vs its own history — but history pre-NVIDIA deal was a weaker company.2,8
PEG: ~1.2 on consensus forward EPS. Not cheap, not egregious. AI infrastructure peers like ALAB trade at PEG ~2.3+; Inflexion rates DDOG at PEG ~3.4. On relative terms COHR is acceptably priced for the thesis quality.
Distance from entry zone: Spot ~$350 is ~6–20% above the $290–$330 add zone and ~27–36% above the ideal $240–$275 zone. Post-Q3 print the stock corrected 6–7%; further consolidation before the Q4 August print is likely.
Consensus price target vs spot: 1-year consensus PT $311–$315 = ~11–14% below current price. Multiple analysts raised targets post-Q3 (Jefferies $350→$375, TD Cowen $340→$395) but the average is still below spot. Analyst targets are catching up to a stock that has run 400%+ in twelve months.5
My variant vs consensus on EPS path: Consensus FY27 EPS of ~$7–8. My view: if the 6" InP capacity doubling hits on schedule (Q1 CY27) and 1.6T volume ramp delivers the guided margin expansion to 42–43% non-GAAP GM, FY27 non-GAAP EPS could reach $8.50–$9.50 — 15–20% above consensus. This makes the current ~45× forward look closer to ~38–40× on my numbers. Not cheap, but defensible.
Institutional positioning note: Artisan trimmed on valuation. Q2 13F (due 14 Aug) is the next institutional flow signal.
PEG: ~1.2 on consensus forward EPS. Not cheap, not egregious. AI infrastructure peers like ALAB trade at PEG ~2.3+; Inflexion rates DDOG at PEG ~3.4. On relative terms COHR is acceptably priced for the thesis quality.
Distance from entry zone: Spot ~$350 is ~6–20% above the $290–$330 add zone and ~27–36% above the ideal $240–$275 zone. Post-Q3 print the stock corrected 6–7%; further consolidation before the Q4 August print is likely.
Consensus price target vs spot: 1-year consensus PT $311–$315 = ~11–14% below current price. Multiple analysts raised targets post-Q3 (Jefferies $350→$375, TD Cowen $340→$395) but the average is still below spot. Analyst targets are catching up to a stock that has run 400%+ in twelve months.5
My variant vs consensus on EPS path: Consensus FY27 EPS of ~$7–8. My view: if the 6" InP capacity doubling hits on schedule (Q1 CY27) and 1.6T volume ramp delivers the guided margin expansion to 42–43% non-GAAP GM, FY27 non-GAAP EPS could reach $8.50–$9.50 — 15–20% above consensus. This makes the current ~45× forward look closer to ~38–40× on my numbers. Not cheap, but defensible.
Institutional positioning note: Artisan trimmed on valuation. Q2 13F (due 14 Aug) is the next institutional flow signal.
Above entry zone after a 400%+ run; initiate small now, size to full 3–5% at $290–$330 or on the Aug Q4 print.
Verdict
COHR's Q3 FY26 print confirms the thesis — datacenter+comms +40.6% YoY, backlog record, 6-inch InP capacity doubling one quarter ahead of schedule, NVIDIA CPO supply locked through end of decade, Q4 guide midpoint $1.98B (+30% YoY). The three-dimension assessment is A / B+ / C+: the thesis is the strongest in the connectivity layer, operations are clean and improving, but the stock is above entry zone with consensus price targets below spot. Do not chase above $370. The right move is to initiate a 1–2% starter position now and build to 3–5% at $290–$330 (likely during the consolidation before the August Q4 print) or on any macro-driven AI capex scare. The August Q4 earnings call — specifically whether FY27 growth guide comes in above 25% and whether non-GAAP gross margin reaches 41%+ — is the binary catalyst that either confirms a full-weight add or delays further sizing.
New position: Start 1–2% now; build to 3–5% at $290–$330 entry zone. Do not exceed 5% given FCF-negative profile and consensus PT below spot. 20% single-stock hard cap.
Opportunity-cost check
▾
vs CSPX
COHR offers direct exposure to the AI optical interconnect bottleneck with a credible multi-year backlog and NVIDIA partnership that CSPX cannot replicate at 0.2% index weight. The edge is the concentrated, asymmetric upside of a single-architecture winner vs broad beta — provided you enter below $330.
vs current holdings
COHR is additive to MU (memory) and PLTR/NOW/META/GOOGL (software) by adding optical infrastructure exposure with zero overlap. It stacks with MU as a "picks-and-shovels of the AI physical layer" pair: MU provides the HBM memory, COHR provides the interconnect. Consider pairing COHR 3% + LITE 2% (60/40) rather than COHR 5% alone — LITE controls the EML laser chip choke-point that COHR also needs for 1.6T.
Thesis-breakers
▾
- 01Non-GAAP gross margin contracts sequentially for two consecutive quarters (below 38.5%) by Q2 FY27 — would indicate 6-inch InP yield ramp is failing or customer pricing is deteriorating.
- 02NVIDIA or a top-3 hyperscaler defers or cancels a multi-year LTA, OR NVIDIA on-boards a second qualified InP supplier at scale by end of CY2026 — would fracture the supply exclusivity that underlies the current valuation premium.
- 03Datacenter & Communications segment revenue growth decelerates below +20% YoY for two consecutive quarters by Q4 FY27 — would signal pluggable transceiver pricing collapse or CPO adoption pulling forward faster than COHR can re-mix toward CPO components.
Key risks
▾
- 01FCF is deeply negative: Nine-month FY26 operating cash flow $10.1M vs $503.3M prior year; Q3 CapEx $290M. If the 6-inch InP yield ramp disappoints, the company is spending $1B+/year with nothing to show for gross margin expansion — risk of a 20–25% de-rate on execution miss.
- 02NVIDIA concentration becoming structural: Post supply agreement, NVIDIA will likely be >10% of revenue in FY27. A single customer accounting for 10–20% of revenue with a $2B equity stake creates both a dependency and a M&A optionality dynamic — the latter is a potential upside (NVDA acquires COHR), the former is downside if NVIDIA's capex decelerates.
- 03CPO disruption timing risk: If hyperscalers accelerate CPO adoption faster than expected (e.g., NVIDIA Spectrum-X 3.0 ships with embedded CPO in 2026 instead of 2027), pluggable transceiver ASPs could compress 15–25% in CY2027, hitting gross margins before COHR's CPO revenue is large enough to offset.
- 04Consensus PT below spot: 12-month consensus PT $311–$315 is ~11–14% below current ~$350 price. Even bulls (TD Cowen $395) project only ~13% upside from here. If the stock does not re-rate on FY27 guidance, the multiple compression path implies a return toward $290–$315 — making current entry unattractive unless you can scale in on weakness.
- 05Geopolitical / export control tail: Sherman, TX fab is domestic (no Taiwan risk) but ~30–40% of revenue is from non-US hyperscaler deployments. Any BIS expansion of export controls on advanced optical components into China or restricted geographies could impact the communications segment ($60M+ quarterly).
DCF scenarios
▾
Discount rate: 13% — High-growth recently profitable tech (non-GAAP profitable; GAAP improving; FCF currently negative due to planned CapEx surge; consistent with the DCF guide for this risk profile).
5Y Rev CAGR
28%
Terminal growth
5%
Fair value / share
~$430
vs current ~$350
+23%
FCF base: FY26 ~$0 (CapEx-constrained); FY27 FCF recovers to ~$400–500M as 6-inch InP capacity doubly delivers operating leverage and customer-funded CapEx moderates to <$200M/yr. FY28–FY30 FCF CAGR 40%+ as 1.6T/CPO mix drives non-GAAP GM to 44%+. Revenue: $7.2B FY26 → $9.2B FY27 → $11.8B FY28 → $15.0B FY29 → $19.2B FY30. Requires: 6" InP yield on schedule, 1.6T volumes as guided, CPO revenue inflection H2 CY2026, NVIDIA relationship intact. This is the base case if management executes, not an optimistic outlier.
5Y Rev CAGR
14%
Terminal growth
5%
Fair value / share
~$195
vs current ~$350
–44%
Pluggable transceiver ASP compression of 20%+ in CY2027 from CPO acceleration or a second qualified InP competitor; D&C growth decelerates to +15% YoY from FY27; gross margin stalls at 39–40% as pricing offsets yield gains; FCF does not recover until FY29. Revenue: $7.2B FY26 → $8.2B FY27 → $9.4B FY28 → $10.7B FY29 → $12.2B FY30. Fair value ~$195 = back to pre-NVIDIA-deal multiple on flat-to-slow growth.
Position: Current price (~$350) sits above the bull-case fair value ($430) only if FY27–FY30 plays out perfectly. It sits well above the bear-case floor ($195). The stock is in a fair-range to moderately elevated zone — justifying a smaller entry now and patience to size into the $290–$330 zone where the bull/bear asymmetry improves meaningfully. At $290, bull upside is +48% and bear downside is –33% — a 1.5:1 ratio acceptable for a 3–5% position.
Catalyst timeline
▾
-
2026-066-inch InP capacity doubling milestone — Coherent guided to complete one quarter ahead of schedule (by end CY2026). Any press release or analyst-day confirmation of this milestone would validate the gross margin expansion thesis and likely trigger a re-rate toward analyst high targets ($395+).
-
2026-07NVIDIA GTC / partner announcements — Any disclosure of initial CPO scale-out revenue from NVIDIA Spectrum-X or Quantum CPO programmes would be the first quantified proof that the CPO TAM expansion is generating real dollars, not just backlog.
-
2026-08Q4 FY2026 earnings (~12 Aug–14) — The pivotal print. Confirm: (1) Q4 revenue ≥$1.98B midpoint; (2) non-GAAP GM ≥40%; (3) FY27 growth guide >25% YoY. All three together = full-weight add to 5%. A miss on any two = hold at starter.
-
2026-Q41.6T volume ramp at multiple hyperscaler sites — Management guided "broad customer adoption of 1.6T" through H2 CY2026. First disclosure of 1.6T as a material revenue line (>10% of DC&C segment) would confirm the next speed tier is not COHR-specific and marks the beginning of the next ASP-expansion cycle.
Named analyst commentary
▾
"CPO will be the main driver of bandwidth increases in scale-up networking for the latter part of this decade and beyond… CPO for scale-up is now not a matter of if and why, but when and how."
— Dylan Patel / SemiAnalysis, Co-Packaged Optics (CPO) — Scaling with Light, 2026-01-01. Note: This quote characterises the CPO ecosystem broadly; SemiAnalysis's public-tier CPO report does not single out COHR by ticker. Coherent's relevance derives from being the primary CW laser supplier for NVIDIA's CPO programme — the architecture SemiAnalysis identifies as the scale-up bottleneck winner.
"Coherent has near-perfect positioning, yet the timing is taking longer than what growth investors typically look for. I believe the stock price will march upward as the equation of what Coherent offers + where the demand is will eventually materialize. The book-to-bill ratio is 4×, meaning they are booking orders 4× faster than they can ship — much of the timing hinges on indium phosphide capacity as the company moves from 3-inch wafers to 6-inch wafers, which will produce 4× the amount of chips at half the cost."
— Beth Kindig / IO Fund, The IO Fund's Top 15 Stocks for Q2 2026, FQ2 earnings analysis (Dec 2025 quarter).9
Serenity (@aleabitoreddit / aibottlenecks.app): No public commentary on COHR located. The aibottlenecks.app framework identifies optical interconnect as a structural AI bottleneck layer, which maps directly to COHR's positioning, but no named direct quote was found.
Sources
1 SEC EDGAR — Coherent Corp. 8-K / Press Release Q3 FY2026, filed 2026-05-06 (EX-99.1); Q3 FY26 earnings call transcript via Motley Fool, 2026-05-06
2 Stockanalysis.com (stockanalysis.com/stocks/cohr/) — forward P/E, PEG, 52W range, institutional %, market cap — accessed 2026-05-12
3 SEC EDGAR — Coherent Corp. Form 10-Q (nine-month cash flow statement) / 8-K supplemental slides — CapEx and OCF figures
4 Investing.com earnings call transcript, 2026-05-06; GlobeNewswire press release, 2026-05-06 — NVIDIA partnership and CPO supply agreement details
5 ChartMill (chartmill.com/stock/quote/COHR/analyst-ratings) — analyst price targets and revision statistics — accessed 2026-05-12 ⚠ single source
6 Stocktitan.net — Form 4 insider filings for COHR CFO and director, Apr–May 2026
7 Yahoo Finance (finance.yahoo.com/quote/COHR) — price, short interest — accessed 2026-05-12; Insider Monkey — Artisan Mid Cap Fund Q1 2026 trim disclosure
8 FinanceCharts — 3-year and 5-year historical P/E averages for COHR ⚠ single source
9 Beth Kindig / IO Fund — The IO Fund's Top 15 Stocks for Q2 2026, "Coherent: Slow and Steady" section; FQ2 earnings analysis covering Dec 2025 quarter. Paywalled (io-fund.com/premium). Used for analyst attribution only; financial metrics on this page sourced from SEC filings and open-access sources above.
⚠ = figure not cross-verifiable from multiple open-access sources. · Q3 FY26 beat/miss: conflicting reads between Investing.com (+1.5% beat) and Quiver Quantitative (–$9.8M miss) — see conflict flag above.
1 SEC EDGAR — Coherent Corp. 8-K / Press Release Q3 FY2026, filed 2026-05-06 (EX-99.1); Q3 FY26 earnings call transcript via Motley Fool, 2026-05-06
2 Stockanalysis.com (stockanalysis.com/stocks/cohr/) — forward P/E, PEG, 52W range, institutional %, market cap — accessed 2026-05-12
3 SEC EDGAR — Coherent Corp. Form 10-Q (nine-month cash flow statement) / 8-K supplemental slides — CapEx and OCF figures
4 Investing.com earnings call transcript, 2026-05-06; GlobeNewswire press release, 2026-05-06 — NVIDIA partnership and CPO supply agreement details
5 ChartMill (chartmill.com/stock/quote/COHR/analyst-ratings) — analyst price targets and revision statistics — accessed 2026-05-12 ⚠ single source
6 Stocktitan.net — Form 4 insider filings for COHR CFO and director, Apr–May 2026
7 Yahoo Finance (finance.yahoo.com/quote/COHR) — price, short interest — accessed 2026-05-12; Insider Monkey — Artisan Mid Cap Fund Q1 2026 trim disclosure
8 FinanceCharts — 3-year and 5-year historical P/E averages for COHR ⚠ single source
9 Beth Kindig / IO Fund — The IO Fund's Top 15 Stocks for Q2 2026, "Coherent: Slow and Steady" section; FQ2 earnings analysis covering Dec 2025 quarter. Paywalled (io-fund.com/premium). Used for analyst attribution only; financial metrics on this page sourced from SEC filings and open-access sources above.
⚠ = figure not cross-verifiable from multiple open-access sources. · Q3 FY26 beat/miss: conflicting reads between Investing.com (+1.5% beat) and Quiver Quantitative (–$9.8M miss) — see conflict flag above.