Silex Q3 2026 slides: 21.7% sales growth, 35.8% margin amid US expansion

Silex Microsystems reported Q3 2026 net sales up 21.7% year over year to SEK 413 million, EBIT up 37% to SEK 148 million and a 35.8% margin, and operating cash flow of SEK 145 million versus SEK 30 million a year earlier. It agreed to acquire a 200mm Pennsylvania fab from Onsemi for USD 40 million, with closing expected around year-end 2027; the company estimates SEK 1,000 million of capex in 2026–2027, including purchase consideration. Management cited high demand and better-than-usual visibility into 2027, while execution, CFIUS approval, customer concentration, and valuation remain risks; shares were reported up 0.9% on the day but down 31% over six months.

Analysis

The investable question is whether a high-return Swedish operation can replicate its economics in a legacy U.S. fab—not whether demand for MEMS is attractive. The Pennsylvania purchase price is a small part of the commitment: conversion, equipment, qualification and staffing dominate the risk. Sweden’s planned capacity additions offer a nearer-term bridge, but they do not validate U.S. utilization or margins. Treat customer interest, reservations and prepayments as de-risking only to the extent they become binding, product-specific commitments.

Near term (days to weeks), strong results may be offset by valuation sensitivity and the market’s focus on cash deployment; one quarter’s margin expansion should not be annualized, particularly with favorable FX. Over 1–3 months, watch order conversion in optical switching and medical programs, hiring intensity, and any revision to Sweden’s ramp. Over 6–18 months, qualification progress and CFIUS are the key gates; meaningful U.S. earnings contribution is much further out. A delay or weaker-than-expected qualification could extend the cash-payback period and pressure the premium multiple.

Competitively, a credible U.S. MEMS manufacturing option could help win customers seeking geographically diversified or trusted supply chains, while established specialty foundries such as Tower Semiconductor may face stronger competition for selected programs. This is not evidence of share loss. ON Semiconductor is the asset seller; absent transaction-specific financial detail, do not infer a material change to its earnings. TSM is mentioned only as an aspiration analogy, not a direct comparable or beneficiary. Contrarian point: the near-term bottleneck may be execution capacity and qualified engineers, not demand. Thesis weakens if reservations fail to convert, Swedish utilization disappoints, or the U.S. timeline slips beyond the stated milestones.

Key Decisions for Investors

• No direct trade in ON or TSM on this report alone: the transaction’s earnings impact on ON is unquantified, and TSM has no stated operating exposure. Avoid treating either as a proxy for Silex.
• For exposure to the Silex thesis, keep it on a watchlist pending an investable security and independently verified valuation; do not underwrite the U.S. expansion from customer-interest language alone.
• Set an alert for evidence of binding U.S. capacity reservations, prepayments and product-specific qualification milestones. Upgrade the expansion thesis only if commitments and timelines hold; downgrade it on CFIUS delay, qualification slippage, or rising capex without corresponding contracted demand.
• Monitor sequential production growth, customer concentration and EBIT margin excluding favorable FX. A reversal in utilization or margin alongside continued investment would challenge the earnings-quality case and increase downside risk.