Aaron

UBS: TSMC is accelerating next-generation packaging development in response to the competitive threat from Intel's EMIB Industry feedback indicates that TSMC is accelerating the development of CoPoS (chip on panel on substrate) and is aiming for mass production by 2028E to compete against Intel's EMIB-T for ultra-large AI package. Intel's EMIB-T targets mass production by H227 or 2028. CoPoS could support Nvidia's Feynman, slated for release in H228, to move to 4-die package with meaningful performance enhancement. See our prior report on panel-level packaging for its benefits vs CoWoS. $INTC $TSM

Korea’s Daishin Securities: Samsung Foundry is estimated to have won a 2nm notebook CPU order from a North American fabless customer. * Could it be Apple?

HOLY SMOKE. What the hell is this?? Memory prices are going absolutely insane. $DRAM $MU $SNDK

No one wants to leave the party at 3 AM. The ABF substrate party continues. _ Nomura: we observe that ABF substrate has completely turned into a seller’s market in the past two months. Substrates have become very important resources which all the major AI chip or CSP customers are eager to secure with LTAs for the next 3-7 years growth. Hence, we believe the LTAs signed during this period could be more comprehensive with potentially more protection power (e.g. prepayments, shorter payback periods, flexible pricing to reflect cost structure and market conditions) to substrate makers than those signed during the COVID cycle. Also, we observe most ABF substrate capacity in both first-tier and second-tier players has been “sold out” for this year, led by strong AI GPU/TPU/ASIC demand, as well as very strong CPU and switch demand (which was not fully expected last year).

Nvidia got FAFOed multiple times with Rubin/Rubin Ultra design - They pushed for higher pin speed HBM4, but Micron's & Hynix's base die were subpar, so they had to change plans & face delays (lower pin speeds) - They pushed for a quad die Rubin Ultra design, yield and warpage issues forced them to switch to 2 die and 2+2 die MCM design - They pushed for 16 Hi HBM4E for Ultra, yield issues forced them to come back down to 12 Hi - They pushed for a dual profile 2.3kW/1.8kW Rubin, but indium-graphite TIM was unstable and now they are switching back to graphite TIM

Don’t get it wrong, it was Pat Gelsinger breakthrough in EUV that’s saving $INTC today. Lip-Bu Tan is great but he’s just reaping the crop that Pat Gelsinger sowed.