📊 Market Overview
Last week the Nasdaq fell 4.48%, the Kospi index dropped 14.46%, and the VIX (a measure of market fear) spiked 34.7%. The sell-off was driven by a sharp unwinding of positions in memory, optics, and tech stocks — even as equipment and power suppliers posted strong results. The big thing to watch this week: Nvidia’s GTC Berlin conference on Tuesday, which could reset the narrative for the entire AI supply chain.
📊 Market Snapshot
This week tests whether AI companies can prove they’re actually making money — not just spending it.
📈 The Big Picture
The demand for AI computing is still strong. TSMC just reported $46.72 billion in revenue, ASML keeps seeing capacity interest years ahead, and Schneider Electric is showing solid organic growth. But the market is no longer rewarding every AI story automatically — contracts and investment plans alone aren’t enough anymore.
The shift this week is from “how much are companies spending on AI?” to “is that spending actually working?” Borrowing costs for AI projects have climbed significantly, with AI high-yield bond spreads (the extra interest risky AI borrowers must pay) reaching about 3.77 percentage points. At least 47 US data-center projects have been delayed, blocked, or canceled in 2026. If you own tech stocks or AI-themed ETFs, this week’s earnings and events will tell you whether the growth story still holds up — or whether financing problems are starting to bite.
📖 Term of the Day
Credit spreads — the extra interest rate that riskier companies must pay to borrow money compared to safe government bonds. When spreads widen, it means lenders see more risk and charge more. Why you care today: AI credit spreads have been widening, meaning it’s getting more expensive for AI infrastructure companies to fund their projects — and that could slow the entire AI buildout.
💼 Watchlist: 3 Stocks to Know Today
TSMC (TSM) — “The Backbone”
TSMC reported $46.72 billion in revenue and could gain even more business if Nvidia’s new NVHBM memory standard (which requires TSMC’s advanced 3nm manufacturing) becomes the industry norm. That would make every AI system — not just the chip — dependent on TSMC.
Micron (MU) — “The Wildcard”
Micron is developing new memory chips for humanoid robots and edge AI devices, targeting mass production by late 2028. It’s an exciting new demand layer beyond data centers, but it’s still far from generating revenue — so watch, don’t chase.
CoreWeave (CRWV) — “The Caution Sign”
CoreWeave is a leveraged AI infrastructure operator that relies heavily on borrowed money to fund its data centers. With AI credit spreads rising and financing getting harder, companies like this face real pressure if they can’t show locked-in customers and steady cash flow soon.
💬 Esther’s Take
