Source item

The more I look at ESMT, the more excited I get researching legacy memory themes. So looked through H1 sellside models... From March ESMT reports: Huanan: DDR3 4Gb ASP +50% QoQ in Q2 Fubon: cited roughly +40% Q2 DRAM pricing By August... Huanan was citing ESMT Q2 ASP growth of: +105–113% Q/Q blended. Saying DDR3 pricing had come in substantially better than expected. So legacy memory pricing translated into much more operating leverage at ESMT. About PSMC wafer cost hikes: Huanan estimated manufacturing costs could rise ~100% in 2H26, but customers remained highly willing to accept the pass-through If we look at July net income was roughly $109M... way above the earnings levels the earlier models were built around. If I had to give a TLDR "consensus" of reports 1. DDR3/DDR2 supply remains structurally tight as competitors move capacity toward higher-value memory. 2. Concord saw no conditions supporting a price reversal during 2026. 3. Concord expected contract-price increases to cover higher foundry costs. 4. Huanan later found that customer willingness to accept those cost increases was stronger than expected. Fun thing I read from Huanan was DDR3 is used in products such as IP cameras and HDDs where the memory component is a relatively small part of the customer's total BOM. So even very large legacy DRAM price hike may only add a few dollars to the finished product's BOM... and for many customers, paying a few extra dollars would be more economical than redesigning + requalifying (but this flows materially into ESMT's net income) Anyway, just excited about this idea (*disclosure I have positions) TLDR: -> Huanan updated ESMT blended Q2 ASP to +105-113% Q/Q -> Given July's blowout net income (putting ESMT at 1.9x p/e annualized RR), implications for the future are exciting if inelasticity continues + ESMT has the ability to hike further due to low % of finished product BOM cost. -> Indication customers were still willing to accept pass through for PSMC wafer hikes.