Knockouttrader

Hmm. Interesting headline. Feisty one.🧐 “SemiAnalysis: Rubin HBM from Micron = zero. HBM4 supply consolidates into a SK hynix / Samsung split (70/30).” Let me explain a bit. It makes sense if NVIDIA is de-risking the HBM4 transition by prioritizing whoever is furthest along in packaging and integration readiness. But I have to add some nuance, based on the HBM framework I’m building for the whole sector. Micron historically hasn’t optimized for the first volume shipment. Their current strategy still isn’t focused on that. They usually show up stronger in the second volume ramp once yield learning compounds and qualification stabilizes. That’s a risk usually. In demand heavy enviroment less. Because the other two players benefit from cumulative yield learning during the initial HBM4 ramp. What you usually see in memory cycles: • Leader ships first • Second supplier closes the yield gap • Platform adds them to diversify supply risk Especially if pricing power remains strong and NVIDIA wants elasticity in the supply chain. Zero market share at launch doesn’t mean permanent exclusion. If Micron closes yields fast and has capacity, they can still enter with volume before HBM4E. Now the nuance. Micron also said they’re fully sold out for 2026. “Sold out” doesn’t automatically mean they anticipated missing Rubin. It could mean capacity is already committed elsewhere likely other NVIDIA programs like Blackwell variants, where they already qualified on HBM3E. The more direct signal I’m watching: Did Micron expect exclusion from the first Vera Rubin wave? That’s the piece of information that actually matters. HBM is a pure execution market. Sequencing matters 10x more than the headlines everyone gets hyped on. I expect to finish the full HBM research at the end of the month, aligned with the launch of our new platform. Yesterday I already shared which strategy I’m advising members. $MU $NVDA $AMD $000660.KS $005930.KS