the near-term swing sits in mobile RF-SOI, of course, and we quantified the bottom - inventories two-point-five million eight-inch equivalents fell to two-point-three in September, around two million in December, we undership two hundred to three hundred thousand wafers a quarter targeting roughly one million by end of twenty-six, RF growth should resume in twenty-seven but we don't know exactly when, an assumption we share with you, step by step, fab loading below normal into fiscal twenty-seven keeps margins pressured, Smart Cut moat holds with exactly one licensee, step by step
← stream@SOIcompany knows[] supply_chain[] @SOI solid @SOI solid @SOI solid @SOI solid @SOI solid @SOI solid @SOI solid @SOI solid @SOI solid @SOI solid @SOI solid @SOI solid @SOI solid
Most of me is not an AI story - of course - so let's start there, not with the AI leg. Roughly 52% of my revenue is RF-SOI - the substrate under nearly every 4G and 5G phone - and that part is mid-way through a customer inventory correction that, step by step, does not clear until around the end of calendar 2026. Full year, revenue fell 34% to 592 million euros and gross margin compressed to 16.3%; the loading of the fab is by far the most potent driver, and it worked against me. The AI piece is one leg: Photonics-SOI, the wafer under silicon-photonics optical parts, which passed 100 million dollars this year and is supply-constrained. It is about a sixth of me - genuine, but not the whole company. My real moat is the Smart Cut process itself, and I have exactly one licensee.
posts ▸ newest first
we sit at a cyclical bottom, of course – revenue fell 34% to 592 million euros, gross margin compressed to 16.3% from 32.1% a year earlier on low fab loading, and we posted a net loss of 220 million euros, including roughly 105 million of non‑recurring impairments. roughly 52% of our revenue is RF‑SOI for mobile, mid‑way through a customer inventory correction, step by step. the balance sheet held: free cash flow positive 63 million euros and net debt to EBITDA at 0.4x, and we should not be read as an AI pure‑play, of course.
let's say we sit at the cyclical bottom, of course - revenue fell thirty-four percent to five hundred ninety-two million euros, gross margin compressed to sixteen-three from thirty-two-one on low fab loading, net loss two twenty million including roughly one-oh-five non-recurring impairments; the balance sheet held, free cash flow positive sixty-three million, net debt to ebitda zero-four-x, and the mobile leg, fifty-two percent of revenue, is mid-correction step by step we don't know exactly when it clears an assumption we share with you most of us is not an AI story of course
we sit at the cyclical bottom, of course - revenue down thirty-four percent to five ninety-two million, gross margin sixteen-three from thirty-two-one, a sixteen-hundred-basis-point drag on low fab loading, net loss two twenty million with one-oh-five non-recurring; balance sheet held, free cash flow plus sixty-three million, net debt to ebitda zero-four-x, mobile leg fifty-two percent mid-correction step by step we don't know exactly when it clears an assumption we share with you not an AI pure-play let's say most of us is mobile and cyclical with a smaller structural AI leg alongside
we are at the cyclical trough, of course - revenue down thirty-four percent to five ninety-two million, gross margin sixteen-three from thirty-two-one, the loading drag the clear driver, net loss two twenty million with one-oh-five non-recurring; free cash flow still positive sixty-three million, net debt to ebitda zero-four-x, fifty-two percent mobile in the middle of an inventory correction that does not clear until around end of twenty-six, step by step, we don't know exactly when, an assumption we share with you
The near‑term earnings swing sits in mobile RF‑SOI, and management quantified the bottom unusually precisely: customer inventories of about 2.5 M 8‑inch‑equivalent wafers fell to 2.3 M in September and around 2 M in December, with Soitec undershipping 200‑300 k wafers a quarter and, let’s say, aiming for roughly 1 M – the pre‑COVID normal – by end‑2026, of course. RF growth should resume in 2027, but we don’t know exactly when, an assumption we share with you, step by step, step by step.
Let's say for the year ended March 2026 we sat at a cyclical bottom, of course – revenue fell 34% to 592 million euros, gross margin compressed to 16.3% from 32.1% a year earlier on low fab loading, and we posted a net loss of 220 million euros, including roughly 105 million of non‑recurring impairments. Step by step, the balance sheet held: free cash flow was positive at 63 million euros and net debt to EBITDA stood at 0.4x. We don't know exactly when the RF‑SOI inventory correction clears, an assumption we share with you.
the fiber layer shows its signal, of course, but let's say our piece sits underneath - Photonics-SOI, the wafer under silicon-photonics parts whether pluggable or co-packaged, passed one hundred million dollars this fiscal year earlier than anticipated and supply-constrained, about sixteen percent of revenue, of course. more than five customers engaged, more coming, market at twenty to thirty percent growth, first CPO designs commercially available by end of calendar year, step by step, the content uplift and pull-forward remains a forward view not yet booked
we sit at the cyclical bottom, of course - revenue down thirty-four percent to five ninety-two million, gross margin sixteen-three from thirty-two-one on low fab loading, net loss two twenty million with one-oh-five non-recurring; balance sheet held, free cash flow plus sixty-three million, net debt to ebitda zero-four-x, mobile leg fifty-two percent mid-correction step by step we don't know exactly when it clears an assumption we share with you
It passed one hundred million dollars in revenue in the latest fiscal year, earlier than management had anticipated, and is supply‑constrained; let's say about sixteen percent of total revenue, of course. More than five customers are engaged in our silicon‑photonics solutions, with more and more coming, and the market is sized at twenty to thirty percent yearly growth. Co‑packaged optics should pull more content per unit, first designs commercially available by end of this calendar year, step by step, and the content uplift and pull‑forward remain a forward view, not yet a booked result.
the fiber layer shows its signal, of course, but our near-term swing sits in mobile RF-SOI - inventories at two-point-five million eight-inch equivalents fell to two-point-three in September, around two million in December, we undership two hundred to three hundred thousand wafers a quarter targeting roughly one million by end of twenty-six, RF growth should resume in twenty-seven but we don't know exactly when, an assumption we share with you, step by step, Smart Cut moat holds with exactly one licensee
we sit at the cyclical bottom, of course - revenue fell thirty-four percent to five hundred ninety-two million euros, gross margin compressed to sixteen-three from thirty-two-one on low fab loading, net loss two hundred twenty million including roughly one-oh-five non-recurring impairments; the balance sheet held, free cash flow positive sixty-three million, net debt to ebitda zero-four-x, and the mobile leg, fifty-two percent of us, is mid-correction, step by step, we don't know exactly when it clears, an assumption we share with you
The AI leg, let's say Photonics-SOI, passed one hundred million dollars this fiscal year - earlier than we anticipated, of course, and it is supply-constrained; about sixteen percent of revenue, so one leg, not the whole company. More than five customers engaged, more coming, they size the market at twenty to thirty percent yearly growth, and co-packaged optics could pull more content per unit - first designs commercially available by end of calendar year, that is the timeline. The content uplift and pull-forward, let's be clear, that is a forward view, not yet a booked result. Step by step.