The biggest winners in this year's AI lift aren't all the usual suspects — Applied Materials, Micron and even eBay have quietly outpaced major ETFs. Applied is up roughly 35% year-to-date to about $346 and Micron about 41% to near $403; both are benefiting from an accelerating AI memory buildout. DRAM now represents 34% of Applied’s Semiconductor Systems segment, Micron’s Cloud Memory unit produced about $5.28 billion at an estimated 66% gross margin, and eBay’s in-house AI tools have helped roughly 10 million sellers create more than 200 million listings and lift engagement.

Memory demand driving a picks-and-shovels play Chipmakers and the equipment companies that serve them are seeing a steady shift in the profit pool as generative AI workloads hit data centres. Demand for high-bandwidth memory and DRAM has pushed semiconductor capital spending into a multi-year horizon. - Applied Materials reported non-GAAP EPS of $2.38 versus a $2.21 estimate and revenue of $7.01 billion; DRAM revenue reached a record level and now accounts for 34% of the Semiconductor Systems segment (up from 27% a year earlier). - Why it matters: Applied supplies process tools used to scale memory stacks and advanced packaging, positioning it to benefit from multi-year memory and packaging spending. CEO Gary Dickerson said the firm expects semiconductor equipment growth above 20% this calendar year. - Analyst view: Morgan Stanley raised its price target on Applied and named it a top pick in U.S. semiconductor equipment, noting the company’s exposure to the AI memory cycle. Micron’s margins show the memory supercycle Micron reported non-GAAP EPS of $4.78 versus a $3.94 estimate and revenue up 56.6% year-over-year to $13.64 billion. GAAP gross margins expanded sharply, rising from 38.4% to 56.0% year-over-year. - Cloud Memory Business Unit produced $5.28 billion in revenue at an estimated 66% gross margin, per the company’s disclosures. - Implication: when data centres increase memory per rack for large language models and other AI tasks, suppliers of high-bandwidth DRAM and the equipment that makes it benefit from higher fab utilisation and pricing power until capacity catches up. eBay’s quieter AI experiment Beyond chips and tools, the AI trade includes platforms that use machine learning to extract more value from existing assets. eBay has integrated in-house large language models to automate item descriptions, generate photos from minimal input and help sellers price and list items at scale. - Impact: eBay’s tools have been used by roughly 10 million sellers to create more than 200 million listings; targeted email campaigns driven by the new stack have seen about 40% more engagement, per Altimetry’s analysis. - Tactical view: these AI features can improve seller productivity and platform engagement without the large capex footprint of cloud providers, offering a different — quieter — way to play AI adoption.

Related Articles

Applied’s CEO has said the company expects semiconductor equipment growth above 20% this calendar year. Whether gains among equipment makers, memory suppliers and niche platform plays are sustainable will depend on continued cloud demand, fab utilisation and the pace of capacity additions.

This article was created with AI assistance.