Handsoff Investor @HandsoffI

THE DRAM VERDICT - June 24th The numbers that set stage: Gross Margin: guided 81% for Q3→Guidance: $33.5B revenue $19.15 EPS Why AI has broken DRAM cycle permanently: HBM3E consumes 3-4x the wafer area per bit vs standard DDR5. substack.com/@handsoff/note…

The memory stock run up is just getting started and this DRAM shortage chart explains exactly why (Save this). This Deutsche Bank and Gartner forecast is one of the most important charts in tech right now and it tells a very simple story that demand is outrunning supply by a margin that gets worse every single year through the end of the decade. In 2026, total DRAM requirements are running 210,000 wafer starts per month above available capacity. By 2027, that shortfall blows out to 507,000 WSPM. By 2028, it reaches 795,000 WSPM, a 29% demand supply mismatch at the exact moment AI infrastructure spending is still accelerating hard. The reason the deficit keeps widening even as supply grows is the HBM multiplier effect. HBM's share of total DRAM wafer capacity rises from 18% in 2025 to 34% by 2028 and one gigabyte of HBM consumes roughly 3 to 4 times the wafer capacity of standard DDR5. Every wafer diverted to HBM for an Nvidia GPU is a wafer taken directly from standard server DRAM, consumer laptops, and smartphones, creating a shortage that spreads far beyond AI data centers and into the entire global tech supply chain. Supply cannot respond fast enough and that is the key point the bears keep missing. New fab capacity requires 2 to 3 years from announcement to volume production meaning every capacity decision being made today does not produce output until 2028 at the earliest. Goldman forecasts DRAM undersupply persisting through 2027, upgraded Samsung and SK Hynix with higher price targets, and projects HBM pricing rising another 44% in 2027 even after the sharp increases already seen this year. Samsung can currently only fulfill 70% of DRAM orders, SK Hynix has sold out its entire 2026 and 2027 HBM allocation already, and server DRAM margins at both companies are running above 80%. Micron is the most direct way to own this trade in the US market. It is one of three companies on earth that can produce HBM at scale, its full 2026 HBM allocation is sold out, and Deutsche Bank raised its price target by 67% specifically because of this demand-supply model. The chart shows pricing power does not peak until 2028 meaning we are still in the early innings of this supercycle, and the stocks have significantly more room to run. Come join Milk Road Pro for our full breakdown, the complete memory supercycle thesis, how to think about Micron, SK Hynix, and Samsung relative to where we are in the cycle, and our entire AI thesis! Link below!

Peter Lynch on owning cyclical stocks "At Magellan, I loaded up on cyclical stocks during the 1981-1982 economic slump. This strategy was one of the keys to Magellan's success... The best time to get involved with cyclicals is when the economy is at its weakest, earnings are at their lowest, and public sentiment is at its bleakest. The staff at Standard & Poor's weekly newsletter, The Outlook, once reviewed the eight recessions since World War II to find out what happened to the prices of key cyclical stocks after the stock market hit bottom. In every instance, the cyclical groups gained 50 percent or better in five months, more than double the advance of the S&P 500."

keyboard_arrow_left Previous keyboard_arrow_right Next

$285A.T | Kioxia's Investor Day Kioxia raised its CY25-28 NAND exabyte demand CAGR from 20% to 22%, with #AlInference growing at 86% CAGR. Data centres will account for ~50% of all flash demand by CY28. Supply/ demand remains in deficit through CY27. substack.com/@handsoff/note…

@Kingprettii Bullish on top 5 WFE companies especially LRCX, KLAC & ASML open.substack.com/pub/handsoff/p…

@EmmanuelInvest Bullish on top 5 WFE companies especially LRCX, KLAC & ASML open.substack.com/pub/handsoff/p…

$ASML | Chips Can’t Wait: ASML’s Q1 2026 Beat Signals AI Is Pulling Forward the Cycle open.substack.com/pub/handsoff/p… 1. AI demand outpacing supply — FY26 guidance raised to €36–40B: 2. Q1 gross margin of 53.0% 3. China exposure declined to ~19% of system sales (vs. 36% in Q4’25)

$KLAC | KLA Beat Across the Board: Process Control Dominance Intact open.substack.com/pub/handsoff/p… 1.Results beat on strong Process Control and Services demand 2. WFE outlook upgraded to $140B+ for CY26; 2027 outlook also raised 3. China revenue stable at 24%; manageable near-term risk

@long_equity Long term bullish on all top 5 WFE companies especially LRCX and KLAC. Both are expensive at this moment. open.substack.com/pub/handsoff/p…

@wallstengine Bullish on all top 5 WFE companies especially $LRCX and $KLAC open.substack.com/pub/handsoff/p…

@kamikazecash Bullish on all top 5 WFE companies especially $LRCX and $KLAC open.substack.com/pub/handsoff/p…

@stock_duty Bullish on all top 5 WFE companies especially $LRCX and $KLAC open.substack.com/pub/handsoff/p…

@rswancpa Bullish on all top 5 WFE companies especially $LRCX and $KLAC open.substack.com/pub/handsoff/p…

@Kevin_stocktips Bullish on all top 5 WFE companies especially $LRCX and $KLAC open.substack.com/pub/handsoff/p…

@sean_________ Bullish on all top 5 WFE companies especially $LRCX and $KLAC open.substack.com/pub/handsoff/p…

@jwat_5 Bullish on all top 5 WFE companies especially $LRCX and $KLAC open.substack.com/pub/handsoff/p…

@QQ_Timmy Bullish on all top 5 WFE companies open.substack.com/pub/handsoff/p…

@NolanGouveiapG Bullish on top 5 WFE companies open.substack.com/pub/handsoff/p…

open.substack.com/pub/handsoff/p…

$LRXC Rapidus Reportedly Taps Lam Research Panel-Level Packaging System for 600mm Square Glass Interposer Push trendforce.com/news/2026/05/2…