AI Spending: Who Benefits? | October 29, 2026

October 29 morning edition — full transcript

Today is October twenty-ninth, twenty twenty-six. Let’s take a look at today’s headlines. AI spending is reaching chipmakers, cloud landlords, and the companies wiring the buildings. Today we follow three earnings stories to see where that spending becomes business—and where cash gets stretched. First, the backdrop. The Fed described growth as strong and inflation as elevated yesterday. Today's GDP and September spending and inflation figures arrive at eight-thirty Eastern. The latest SOFR fixing, for October twenty-eighth, eased to 3.94 percent from 3.95 percent. It remains six basis points above October twenty-first. For long AI construction projects, financing discipline still matters. The Fed held its target at 3.75 to 4 percent, with two dissenters favoring a hike. Our focus is whether today's inflation data and funding rates keep pressure on the cost of building AI capacity. First, Microsoft, reporting last night. Azure revenue grew 48 percent. Microsoft generated 58 billion dollars of operating cash and spent 39.4 billion in cash on property and equipment—about 68 percent of that cash generation. This tells us why delivery matters alongside demand. The first link is CoreWeave, an established Microsoft compute supplier. If Microsoft rents more external capacity, CoreWeave can convert delivered computing capacity into billings. The second link is NVIDIA: CoreWeave's announced expansion uses NVIDIA computing platforms. Additional, funded deployments would support hardware demand. Our judgment: the key variable is Microsoft's balance between its own facilities and rented capacity. Next, look for supplier contract updates and capacity coming online; a shift toward internal construction would weaken the rental channel. Second, Alphabet, also reporting last night. Google Cloud revenue doubled to 30.7 billion dollars, with TPU systems contributing to growth. Yet equipment purchases of 51.4 billion exceeded operating cash flow by 9.1 billion. Our reading is strong demand paired with a heavier funding requirement. Follow the chain to Broadcom: its disclosed Google agreements cover future custom TPUs and networking components. Then comes TSMC, which produced about 95 percent of Broadcom's outsourced wafers in fiscal 2025. Continuing TPU procurement could therefore carry demand from Google through chip design into manufacturing. Watch Broadcom's AI orders and delivery schedules, then TSMC's capacity and utilization. The timing depends on actual procurement; Google's spending total alone cannot tell us each supplier's share. Third, a fresh release at six this morning: Wesco. Quarterly data-center sales reached 1.8 billion dollars, up about 55 percent, and roughly a quarter of total sales. Company-wide backlog rose about 80 percent. Here the chain moves from construction plans into physical delivery. Wesco distributes the electrical and communications products projects need. Its established optical-fiber channel includes Corning, listed as an authorized-distribution relationship in Corning's own guide. As projects ship and distributors replenish inventory, fiber manufacturers can receive follow-on demand. Our judgment: this is a useful test of whether the AI build-out is progressing beyond chip purchases. At Wesco's ten-a.m. call, look for backlog conversion, customer advances, and inventory requirements. Project delays or inventory accumulated ahead of demand would weaken that signal. The common thread is cash turning into installed capacity: Microsoft and Google fund it, chip suppliers equip it, and distribution channels help deliver it. At Newstock, explore these news, industry, and company connections. Our analysis is for your reference and is not investment advice. Thanks for watching. I'm Kylie Young. Visit newstock.ai or download our app on the App Store and Google Play. Follow us on X and YouTube and share your ideas to improve new stock. See you next time.