Earnings Call Highlights: Coherent (COHR) Q1 FY2027 — A $20B Five-Year LTA and a 45% Margin Target, but InP Has Four Mouths to Feed and Free Cash Flow Is -$868M

3 minutes ago

10 min read

Coherent (COHR) reported Q1 FY2027 results (quarter ended September 2026) on November 4, US Eastern time: revenue of $2.37B, up 50% year over year, and non-GAAP EPS of $2.09, up 80%. What really changed on this call was the scale of visibility: one customer signed a five-year, roughly $20B PhotonLink long-term agreement (LTA), the first $4B quarter was placed in FY2028, and the gross margin target moved from 42% to 45%. The cost shows up in the same filing: free cash flow was negative $868M this quarter.

This article is for technology and industry trend analysis only and is not investment advice.

1. Revenue and Financials

Q1 revenue was $2.37B (GAAP), up 15.9% sequentially and 49.9% year over year. Excluding the aerospace and defense business and the Munich product division, both divested in 2026, pro forma growth was 56%. Consensus was about $2.30B on revenue and about $1.97 on non-GAAP EPS, and both were beaten. The Q2 EPS guide of $2.28–$2.48 also sits above the roughly $2.24 consensus.

Metric	Q1 FY27	Q4 FY26	Q1 FY26	YoY
Revenue	$2,370M	$2,046M	$1,581M	+49.9%
GAAP gross margin	39.1%	38.5%	36.6%	+254 bps
Non-GAAP gross margin	40.6%	40.2%	38.7%	+184 bps
GAAP operating income	$378M	$254M	$259M	+45.9%
Non-GAAP operating margin	23.2%	21.8%	19.5%	+368 bps
GAAP net income	$385M	$241M	$226M	+70.2%
GAAP diluted EPS	$1.90	$1.19	$1.19	+$0.71
Non-GAAP diluted EPS	$2.09	$1.74	$1.16	+$0.93

The GAAP numbers need two footnotes. First, GAAP operating margin fell 46 bps year over year (15.9% vs. 16.4%), but the year-ago quarter included a $115.2M gain on the sale of a business. Without it, the year-ago margin was only about 9.1%, so the core business did not go backward. Second, GAAP income tax this quarter was a $22.8M benefit, one reason GAAP net income came in above operating income. For the underlying earnings trend, the non-GAAP $2.09 is the cleaner number.

By segment, Datacenter & Communications was $1,951.7M, up 21% sequentially and 79% year over year, or 82% of revenue, its fourth straight quarter of accelerating sequential and annual growth. Industrial was $418.6M, down about 15% as reported, but that reflects the divestitures; on a pro forma basis it was flat both sequentially and year over year. Management says semiconductor equipment bookings have strengthened and expects growth to resume this quarter.

▲ Coherent FY2027 Q1 core financials (GAAP revenue $2.37B, gross margin 39.1%, operating margin 15.9%, net income $384.7M, diluted EPS $1.90, segment revenue, Q2 guidance $2.55B–$2.75B). (Image source: Simple Tech Trend original; data from the Coherent Q1 FY2027 earnings release and call)

Last quarter, in Earnings Call Highlights: Coherent (COHR) FY2026 Q4, we listed three things to track, and we can now check them. Datacenter growth above 80%: it came in at 85%, a pass. CPO revenue recognition: the 400 mW CW laser starts generating revenue this quarter, but management itself calls the amount small, and it only starts to count in the March quarter. Gross margin slope toward 42%: at 40.6% it is not there yet, and the company has already swapped the target for 45%.

2. PhotonLink: What the $20B LTA Actually Buys

PhotonLink is the integrated optics platform Coherent launched at ECOC, spanning CW lasers, laser modules, isolators, fiber assemblies and connectors, and supporting both InP CW lasers and VCSELs. This quarter it went from a product name to a set of numbers:

This quarter: small shipments, including the 400 mW CW laser and fiber assemblies
March quarter: more than $100M
Before the end of 2027: more than $250M per quarter
Early 2028: revenue from the $20B LTA starts to flow

Draw that as a line and the slope becomes obvious. $20B over five years averages $4B a year, or $1B a quarter, while the late-2027 target is $250M or more per quarter. That implies a step-up of more than four times during the LTA, from one customer alone. Management also says 2028 ramps quarter by quarter rather than all at once, so the later quarters would run above $1B.

The content math is worth unpacking too. Management estimates about $15,000 per chip for a 100T I/O XPU or switch chip using the full PhotonLink portfolio. STT's rough estimate: if the entire $20B were at full content, it would correspond to about 1.33 million 100T chips, or about 270,000 a year. But this LTA is anchored by the 400 mW laser and the customer may not buy everything, so the actual chip count would be higher. The CEO also acknowledges that some customers will take only a subset; $15,000 is a ceiling, not an average.

The market sizing moved up as well. Last quarter management described CPO, NPO and integrated optics as an incremental market of more than $15B. This quarter it says CPO, NPO and chip-to-chip interconnects are about $30B by 2030, a scope that now adds chip-to-chip, with the total datacenter and communications market at about $90B. OCS stays at $4B or more.

The adoption schedule is more worth writing down than the dollar figures. It is the most specific public timeline so far:

Application	Technology	Management timing
Scale-out CPO	InP high-power CW	Shipping now
Scale-up CPO	InP high-power CW	2H 2027, toward year-end
Scale-up NPO	Includes VCSEL (around early 2028)	Late 2027 to early 2028
Scale-in chip-to-chip	PhotonLink, technology not specified	2029–2030

The CEO sees the real demand in scale-up, because replacing copper with optics is entirely new growth that does not cannibalize pluggables. For scale-out, he still expects pluggables to stay dominant for the coming years, and a 3.2T OSFP built on Coherent's own 400G differential EMLs was shown at ECOC. Beyond the two anchor customers, there are more than 10 CPO, more than 10 NPO and more than five chip-to-chip engagements. For a sense of how hard high-power CW is, compare the LandMark Optoelectronics Q3 2026 call, which described 30–50% fewer lasers per substrate.

3. InP and Gross Margin: One Pool of Capacity, Four Mouths to Feed

The capacity news is good. Internal InP capacity doubled year over year, a quarter ahead of plan; EML output hit a record; six-inch yields keep running above three-inch; Zurich's six-inch line starts in the first half of 2027; capacity more than doubles again by the end of 2027; and substrates come from more than five suppliers, with long-term agreements securing the next expansion.

But the most valuable sentence of the call came in Q&A: the internal versus external laser mix has barely changed in a year. The reason is simple. Datacenter transceiver growth has tracked internal InP capacity with a one-quarter lag, so all of the doubled capacity went straight into shipments and none of it replaced purchased lasers.

That clarifies how the 45% gross margin target is built. The CFO ranks new product ramps and full operating leverage from six-inch InP first, then pricing optimization, then cost reductions. Lay out the numbers:

Going from 40.6% to 45% means 4.4 points over the nine quarters to the end of 2028, or an average of 49 bps a quarter
The past year actually delivered about 46 bps a quarter, with the internal/external mix unchanged
So the target only maintains the current slope, and if the insourcing benefit the CEO says is "still ahead of us" comes through, there should be room to raise it

The opposite reading also holds. Margin expansion over the past year came mainly from pricing and yields, and pricing is the lever most dependent on a shortage. And the old 42% target has not actually been reached; only the top of the Q2 guide touches it, and the target has already moved.

The real challenge is allocation. Over the next two years, InP capacity has four destinations at once: 1.6T and 800G pluggables, PhotonLink CW lasers, Scale Across DCI applications, and replacing purchased lasers to improve margin. High-power CW also yields fewer die per wafer than ordinary lasers. The CEO sees no InP/EML oversupply through 2028 and says Coherent cannot even meet its own internal demand. Those statements hold up, but they also mean that hitting 45% on time depends on how these four claims get prioritized.

4. Scale Across and OCS: Two More Lines Already Generating Revenue

Communications grew 34% sequentially and 84% year over year, and its largest contributor is no longer traditional telecom but Scale Across datacenter interconnect. It exceeded $200M in the quarter, up 53% sequentially and nearly fourfold year over year, for a run rate of almost $1B. Management admits the pace surprised even them.

Customers include both equipment makers and hyperscalers, and products run from pump laser components through ZR+ modules, line cards, amplifiers and ROADMs to complete solutions. More importantly, this growth comes before MultiRail, which starts ramping in the March quarter. Management expects quarterly revenue to double again by the end of 2027, to more than $400M. For the background on this demand, see Ciena's Scale Across talk at OCP.

OCS, meanwhile, was the item an analyst pointed out was missing from the slides. The CEO's response: 320×320 is shipping, denser versions target scale-up, two sites are ramping in parallel, and the liquid crystal approach has no mechanical parts, with better power and reliability than MEMS. But he gave no dollar figure, only "an increasingly larger contributor." Next to PhotonLink and Scale Across, which come with numbers, that wording is clearly a notch lower.

Industrial is a collection of future options: CO2 lasers and zinc selenide optics for drilling AI server PCBs, PLD systems for fusion superconductor tape, and Thermadite cooling material, with prototypes shipping this quarter and revenue starting in the second half of 2027.

5. Cash Flow Breakdown: Where the Expansion Money Comes From

Outside the headline, this is the section most worth reading. Last quarter we were writing about deleveraging to 0.7x. This quarter the direction reversed:

Capex of $776M, or 33% of revenue (27% last quarter, 7% a year ago), and it rises again in Q2
Operating cash flow of negative $92.7M, and free cash flow of about negative $868M
Inventory up from $2.58B to $3.34B, or 29% sequentially, versus 16% revenue growth
$773M borrowed under other credit facilities, almost exactly the quarter's capex; debt leverage up from 0.7x to 1.1x
Cash, short-term investments and restricted cash of $2.34B, versus $2.59B last quarter

This does not mean the finances are in trouble. Leverage of 1.1x is not high, and margins and order visibility support it. Management also disclosed that nearly every LTA includes a customer prepayment or NRE that co-funds capacity, Coherent's main hedge against capital risk.

But two things should be kept apart. First, orders fill 2027 and extend into 2028, so the risk is not near term. Second, when an analyst raised investor concern about AI datacenter financing and permitting, the CEO answered that the shift from copper to optics is a technology trend independent of datacenter demand. That holds over the long run, but in the short run the two cannot be truly independent: without datacenter buildouts, there is no scale-up to convert.

There are also two versions of capacity flexibility. In prepared remarks, the CFO said capex offers "significant fungibility" across multiple product lines. Asked about InP versus GaAs, the CEO said there is "some fungibility," but lines cannot be switched easily. The two statements do not conflict, but they describe different levels. If scale-up ends up more VCSEL-heavy than expected, this InP capital may not transfer. The offset is that Coherent already has capacity for more than 1 billion VCSEL arrays a year on GaAs and shipped more than 700 million VCSELs over the past year, so it has bets on both sides.

6. Supply Chain Signals: The Independent Supplier After Marvell-Lumentum

On November 2, Marvell announced an all-stock acquisition of Lumentum at an implied enterprise value of $115B, still pending. We argued in Marvell Wants InP Capacity: 2026 W44 industry review that this deal prices laser capacity. The CEO's answer was direct: customers need an independent optical partner without competitive conflict, able to support both hyperscalers and semiconductor customers.

Another remark makes this clearer. Asked about Chinese laser suppliers, the CEO said that at the 400 mW level he sees "only one competitor." He did not name it; STT's read is Lumentum, whose light source specs are covered in Lumentum's CPO and NPO light source talk at OCP. If high-power CW really is a two-horse race, and one horse joins a chip company that also builds custom XPUs and scale-up interconnects, other chipmakers' procurement logic could well change.

Two caveats first. The $20B LTA was signed in Q1, before the deal was announced, so causality cannot be written backward. And Marvell and Lumentum are still separate companies; actual customer behavior will only show later.

Other signals worth noting: in March NVIDIA signed a multi-year deal with Coherent and invested $2B, and invested the same amount in Lumentum that month, which from a procurement standpoint amounts to setting up its own dual supply at 400 mW. Also, when an analyst asked about competition and pricing pressure from Chinese module makers, the CFO answered only on margin drivers and did not address the competition half. The CEO's "no competitor" remark likewise covers only high-power CW, not transceivers.

The comparison set matters too. Lumentum's same-period revenue was $1.26B, with non-GAAP gross margin of 53.2% and operating margin of 42.3%. Coherent has nearly twice the revenue and a gross margin 12.6 points lower. Asked directly whether 45% is the end point, management called it "a stop along the way." STT's read is that the gap comes mainly from Coherent's transceiver-heavy mix and purchased lasers, which is why PhotonLink and the insourcing ratio matter so much for it.

7. Management Outlook: Guidance and Language

Official Q2 FY2027 guidance: revenue of $2.55B–$2.75B, non-GAAP gross margin of 40%–42%, non-GAAP operating expenses of $430M–$450M, a tax rate of 18%–20%, and non-GAAP EPS of $2.28–$2.48. The $2.65B revenue midpoint implies about 12% sequential and about 57% annual growth.

Three longer-term targets came out at once, at different levels of certainty:

"On track": the first $3B quarter by the end of FY2027
"Expect": FY2027 pro forma growth above 60%, and the first quarter with non-GAAP gross margin above 45% before the end of 2028
"Clear line of sight": the first $4B quarter in FY2028

Put the first two together and they do not quite line up. The FY2026 pro forma base is about $7.0B (STT estimate, FY2026 revenue of $7.12B less divested businesses), so 60% growth means $11.2B. Q1 plus the Q2 midpoint totals $5.02B, leaving at least about $6.2B for the second half, or an average of $3.1B a quarter. At a constant sequential rate, that is about $2.9B in the March quarter and $3.25B in the June quarter. In other words, if 60% holds, $3B is not a target but a floor already passed. The earlier full-year consensus of about $10.6B also sits below this figure.

By the same logic, getting from $3.25B to $4B takes only about 5% sequential growth within FY2028, just as the $20B LTA starts contributing. The CEO says he is giving a longer view this time because 2027 is basically booked out, 2028 bookings have started, and both the $3B and $4B quarters are backed by supply agreements. Management has kept a cushion, which is common on earnings calls, but the size of this one is worth remembering.

8. Takeaway

What Coherent delivered this quarter is not just a set of results but a roadmap out to 2030: a $20B LTA, 45% gross margin and a $4B quarter. All three rest on one premise: whether InP capacity can meet shipments and still leave enough to replace purchased lasers. Right now that premise is being paid for with negative free cash flow.

Three things to track next:

Whether PhotonLink exceeds $100M in the March quarter. This is the first direct test of CPO moving from demo to revenue.
When the internal InP mix starts to move. Management says the shift begins over the coming quarters; a single-quarter gross margin gain well above 50 bps would be the signal.
The pace of free cash flow and customer prepayments. Capex rises again in Q2, so watch whether inventory growth falls back below revenue growth and whether leverage keeps climbing.

The conclusion from this call: demand in optical communications is no longer the question. The question is whose money and capacity can last until 2028.