Who to watch post‑TSMC
Post‑earnings investor threads put NVIDIA and AMD at the top of a bullish AI pecking order, with Broadcom and Micron next and Qualcomm/Sandisk‑style names in sympathy. (x.com) Analysts also flagged that 3nm and advanced packaging capacity is largely full, reinforcing interest in suppliers that sit alongside foundries. ( ) Market chatter repeatedly named ASML and Broadcom as complementary plays to foundry exposure. (x.com)
After TSMC’s April 16 results, investors shifted from the foundry itself to the companies most exposed to scarce leading-edge chip capacity and packaging. (investor.tsmc.com) TSMC reported first-quarter 2026 revenue of $35.90 billion, gross margin of 66.2%, and second-quarter guidance of $39.0 billion to $40.2 billion. The company also raised its 2026 revenue growth outlook to above 30% year over year. (investor.tsmc.com; finance.yahoo.com) The mix mattered as much as the headline. In TSMC’s first quarter, 3-nanometer chips made up 25% of wafer revenue, while 5-nanometer and 7-nanometer chips made up 36% and 13%, leaving 74% of wafer revenue in 7-nanometer-and-below processes. (investor.tsmc.com) That steered attention first to NVIDIA and AMD, which sell the graphics processors and accelerators at the center of the artificial-intelligence server buildout. Nvidia said Blackwell had reached volume production, and Advanced Micro Devices is marketing its Instinct MI350 series with 288 gigabytes of HBM3E memory and 8 terabytes per second of bandwidth. (blogs.nvidia.com; amd.com) Broadcom stayed high on investor watchlists because it sits on two AI spending lines at once: custom chips and the networking gear that links clusters together. Broadcom said in December it expected fiscal first-quarter 2026 AI semiconductor revenue to double year over year to $8.2 billion, driven by custom accelerators and Ethernet AI switches. (broadcom.com) Micron followed because AI servers need high-bandwidth memory, the stacked memory chips mounted next to accelerators inside advanced packages. Micron said on March 18 that fiscal second-quarter 2026 revenue reached a record $23.86 billion, and Chief Executive Sanjay Mehrotra said demand and industry supply stayed tight. (investors.micron.com) The bottleneck is not only wafer starts. Advanced packaging, including chip-on-wafer-on-substrate systems used to connect processors and memory, has become a second choke point because the fastest AI chips cannot ship without it. (investor.tsmc.com; blogs.nvidia.com) That is why ASML keeps coming up alongside foundry names. ASML sells the lithography machines used to print the smallest features on advanced chips, and on April 15 it reported €8.8 billion in first-quarter sales and lifted its 2026 sales outlook to €36 billion to €40 billion. (asml.com) The sympathy trade into Qualcomm- or SanDisk-style names is looser. Those companies are not the clearest direct beneficiaries of TSMC’s leading-edge squeeze, but they can rise when investors broaden the bet from a few AI winners to a wider semiconductor cycle. (investor.tsmc.com; asml.com) The short version after TSMC is that the market is ranking exposure to scarcity: first the chip designers with the most AI demand, then the memory and networking suppliers around them, then the equipment makers that help foundries expand. (investor.tsmc.com; broadcom.com; investors.micron.com; asml.com)
