AI demand keeps chip cycle hot

Taiwan Semiconductor reported first-quarter revenue up 35.1% year-on-year, with coverage attributing the surge to persistent AI demand from hyperscalers and large cloud buyers. Industry tracking also projects massive hyperscaler spending on AI infrastructure in 2026, underlining that chipmakers and foundries are still benefiting from the AI build-out even as macro risk rises. (benzinga.com, intellectia.ai)

Taiwan Semiconductor did not just post a good quarter. It posted NT$1.134 trillion in first-quarter 2026 revenue, up 35.1% from a year earlier, and March alone jumped 45.2%. (tsmc.com) That company is the factory behind many of the world’s most advanced chips. When Nvidia, Advanced Micro Devices, Apple, or Qualcomm design a top-end processor, Taiwan Semiconductor is often the one that actually manufactures it. (reuters.com) The buyers pushing this are the hyperscalers, which is industry shorthand for the handful of cloud giants building data centers at planetary scale. Microsoft, Amazon, Alphabet, and Meta are buying chips the way airlines buy jet engines: in huge batches, because their whole business now depends on having enough compute. (cnbc.com) Compute is the raw horsepower used to train and run artificial intelligence models. The more chatbots, coding tools, search features, and image generators these companies launch, the more graphics processors, memory, networking gear, and power systems they need underneath. (delloro.com) That is why Taiwan Semiconductor’s numbers matter beyond one stock. The company’s January 2026 guidance for the quarter was US$34.6 billion to US$35.8 billion, and the March revenue report implies it landed at the top end of that range. (tsmc.com) The spending wave above it is still getting bigger. CNBC reported in February 2026 that Microsoft, Amazon, Alphabet, and Meta were on track for combined capital spending close to $700 billion this year, with artificial intelligence infrastructure driving the jump. (cnbc.com) Alphabet alone said on February 4 that 2026 capital expenditures would be $175 billion to $185 billion. Reuters said the increase was aimed at easing compute constraints and pushing harder in the artificial intelligence race. (cnbc.com, reuters.com) Meta also raised its 2026 capital spending plan to $115 billion to $135 billion on January 28. Reuters tied that jump to Mark Zuckerberg’s push for what Meta called “superintelligence,” which means more servers, more data centers, and more chips. (reuters.com) Even Microsoft, which reports results by fiscal quarter instead of calendar quarter, said on January 28 that “Cloud and AI” drove revenue to $81.3 billion for the December 2025 quarter. That tells you the same thing from the customer side: the companies ordering the hardware are still monetizing the software and cloud services layered on top. (microsoft.com) There is still macro risk around tariffs, power availability, and whether all this spending earns enough return. But as of April 2026, the actual purchase orders are still flowing through to foundries, and Taiwan Semiconductor’s 35.1% jump is one of the clearest receipts yet. (reuters.com, tsmc.com)

Get your own daily briefing

Scout delivers personalized news, insights, and conversations tailored to your role and industry.

Download on the App Store

Shared from Scout - Be the smartest in the room.