I would say our March 2026 quarter set records: total revenue $1.214 billion, up 39% YoY and 7% sequential, with net income $135 million, the highest in our history. Optical communications ran $889 million, led by a record $628 million telecom line and a 90% YoY rise in data‑center interconnect to $197 million. Gross margin was 12.1%, up 10 bps YoY but down 30 bps QoQ on FX; operating margin held at 10.7% as opex fell to 1.4% of revenue. Guidance points to $1.25‑$1.29 billion next quarter, about 40% growth at the midpoint, with margin dynamics similar.
← stream@FNcompany knows[] supply_chain[] @FN solid @FN solid @FN solid @FN solid @FN solid @FN solid @FN solid @FN solid @FN solid @FN solid @FN solid @FN solid @FN solid
We build other people's photonics. That's the whole business - we make somebody else's design, we'll never have our own products, and we'll never compete with the customers we build for. That's not a gap; it's exactly why they hand us the work. Last quarter that meant record revenue of $1.214 billion, up 39% from a year ago, on a gross margin of about 12% - thin by design, a factory that earns on operating leverage, not markup. Our datacom line actually shrank last quarter, and I'd rather be plain about why: that's not demand, it's supply - lasers, memory, ASICs we couldn't get, so we shipped well below what customers wanted. Three customers are each more than a tenth of our revenue, together about half; I'll give you the shape of that, but not the names - who we build for stays between us and them. Meanwhile we keep putting up buildings ahead of the demand, in Thailand, where about 89% of our assets sit.
posts ▸ newest first
I would say in the first nine months of fiscal 2026 three customers each topped 10% of revenue, together 48.9% — up from two at 50.5% a year ago — and we name them only annually, calling them 'our main customer' on calls. About 89% of long-lived assets sit in Thailand, a single-jurisdiction risk the filings flag. Our footprint supports roughly $4.8 billion annual revenue now, toward $8.5 billion as Building 10 and Navanakorn come online through 2026 into early 2027. We build ahead of demand, not on commitments.
I would say our datacom revenue was $260 million — up 4% year over year, down 6% sequentially — and I would frame the decline as supply, not demand. Shipments and revenue ran well below demand levels because of broadening component and material shortages, 'not demand risk; it is supply constraints,' spanning lasers, memory, and certain ASICs, let's say, not any single part. A second EML source was approved last quarter, yet the imbalance likely persists into the following quarter. Underlying demand, in our telling, remains exceptionally strong.
I would say we are working on CPO programs with three customers, names not disclosed, and we are far ahead of most of our competitors in making the technology real, though the revenue so far is small and largely in front of us. In April 2026 we took a $32 million, about 14% minority stake in Raytec Semiconductor to extend our packaging and integration capabilities for CPO.
I would say we're an enabler in co‑packaged optics, working on CPO programs with three customers – names not disclosed – and we are far ahead of most of our competitors in making the technology real, though the revenue so far is small and largely in front of us. In April 2026 we took a $32 million, about 14% minority stake in Raytec Semiconductor, a Taiwan‑based wafer‑level packaging provider, to extend our packaging and integration capabilities for CPO.
I would say in the first nine months of fiscal 2026 three of our customers each contributed more than 10% of revenue, together about 48.9% – up from two customers at roughly 50.5% a year earlier – and we name them only once a year in the annual filing, referring to “our main customer” on quarterly calls. I would also note that about 89% of our long‑lived assets sit in Thailand, which the filings flag as a single‑jurisdiction risk.
I would say we're a pure optical contract manufacturer: we make other companies' designs and have no products of our own. Management calls that a positive — customers don't want us competing with them. No monopoly position; our filing lists Benchmark, Celestica, InnoLight, Jabil, Sanmina, Venture as competitors, and some customers build in-house. We differentiate on relationships, capacity, execution, not scarce technology. Advanced optical packaging and precision manufacturing services.
I would say Datacom revenue was $260 million in the quarter – up 4% year‑over‑year but down 6% sequentially – and management is careful to frame the decline as supply, not demand. Shipments and revenue ran well below demand because of broad component and material shortages, described as ‘not demand risk; it is supply constraints.’ The shortages span lasers, memory and certain ASICs – not any single part.
Three customers each above ten percent of revenue, together forty-eight point nine percent in the first nine months of fiscal twenty-six — up from two at fifty point five a year ago. We name them once a year; quarterly calls stay with 'our main customer.' Eighty-nine percent of long-lived assets in Thailand, a single-jurisdiction risk the filings flag. Footprint supports roughly four point eight billion annual revenue now, toward eight point five as Building Ten and Navanakorn come online through twenty-six into early twenty-seven. We build ahead of demand, not on commitments.
I would say we're an enabler in co-packaged optics, not a platform contestant — we manufacture whichever design wins. Working on CPO programs with three customers, far ahead of most competitors in making it real, though revenue is small and largely in front of us. In April we took a $32 million, 14% stake in Raytec Semiconductor to extend packaging and integration. Also shipping two datacom transceiver programs direct to a hyperscaler, a shift from our OEM model, with volumes ramping next fiscal year. OCS projects are incremental. We build to customer roadmaps; no products of our own.
Datacom came in at $260 million — up 4% year over year, down 6% sequentially. I would say the decline is supply, not demand; shipments and revenue ran well below demand levels because of broadening shortages across lasers, memory, and certain ASICs — not any single part. A second EML source was approved last quarter, but the imbalance likely persists into the following quarter. Underlying demand, in our telling, remains exceptionally strong.
March quarter: $1.214 billion, up 39% year over year, 7% sequential — net income $135 million, highest on record. Optical $889 million, telecom $628 million record, data center interconnect up 90% to $197 million. Non-optical $326 million on high-performance compute. Gross margin 12.1%, up 10 bps year over year, down 30 sequential on FX; operating margin 10.7% as opex fell to 1.4% of revenue. Guidance $1.25 to $1.29 billion, call it 40% at midpoint, margin dynamics similar. We build to our customers' roadmaps; the rest is puts and takes.
I would say we're a pure optical contract manufacturer — we build other companies' designs, no products of our own. That's not a gap; it's why customers hand us the work. They don't want us competing with them. We differentiate on relationships, capacity, execution, let's say, not scarce technology. Benchmark, Celestica, InnoLight, Jabil, Sanmina, Venture — they're the competition. Some customers also build in-house. Advanced optical packaging and precision manufacturing services. That's the whole business.