GOOD MORNING, Tuesday delivered one of the clearest signs yet that the AI trade has become an expectations problem. Samsung reported a 19-fold jump in second-quarter operating profit, yet memory and semiconductor stocks sold off across Asia and the U.S. The Philadelphia Semiconductor Index fell 4.7%, Micron dropped 4.7% and the Nasdaq lost 1.2%. At the same time, attacks near the Strait of Hormuz kept oil and inflation risk elevated. Strong AI demand is no longer enough by itself. The market now wants proof that extraordinary profits can keep outrunning extraordinary expectations. MARKETS | TLDR
SEMICONDUCTORSSamsung's Record Profit Wasn't EnoughSamsung Electronics reported a 19-fold jump in second-quarter operating profit, surpassing the combined earnings of its previous three years, as memory pricing and AI-related demand surged. On a normal tape, that should have been a major bullish signal for the semiconductor complex. Instead, Samsung shares fell in Seoul and U.S. chip stocks followed them lower. The market reaction matters because it shows where the AI trade has moved. Investors are not questioning whether demand is strong. They are questioning whether the growth is strong enough relative to valuations after an extraordinary first-half rally. The Philadelphia Semiconductor Index had already gained roughly 74% for the year even after Tuesday's 4.65% decline. That creates a much tougher earnings bar. When a bellwether can deliver a near-20x profit increase and still fail to lift the sector, future upside depends less on simply beating consensus and more on proving that margins, capacity constraints and spending growth can remain exceptional for years. Great fundamentals are no longer the same thing as a positive stock catalyst. AI COMPETITIONDeepSeek Is Trying to Take the AI Trade Off Nvidia's HardwareChinese AI startup DeepSeek is developing its own AI chip, according to Reuters, in a move that could reduce its reliance on Nvidia and Huawei hardware. The report added another layer of pressure to a semiconductor trade already struggling with high valuations and questions around the durability of hyperscaler spending. The strategic read-through is larger than one startup. Google, Amazon, Microsoft and Meta are already investing in custom accelerators, and now leading model developers are exploring the same path. Every successful custom chip does not eliminate merchant GPU demand, but it gives large buyers more leverage on price and reduces the assumption that all AI compute growth will flow through the same small set of suppliers. That is why Tuesday's selloff looked broader than a simple earnings disappointment. AI demand can continue rising while the economics shift between layers of the stack. The more customers own their silicon, the harder it becomes to assume today's hardware margins and market shares persist indefinitely. HEADLINES
UPCOMING
DEEP INSIGTHSNasdaq Sinks as AI Worries Hit ChipmakersRead this for the cleanest snapshot of the market's changing standard for AI. Samsung's extraordinary profit growth, DeepSeek's custom-chip push and the 4.65% drop in the chip index all point to the same conclusion: AI demand remains powerful, but investors are no longer willing to assume that every supplier keeps the same share of the economics. Trading Surge, Helped by SpaceX IPO, Seen Lifting Wall Street BanksThis is useful because it shows the other side of market volatility. A shakier AI trade and a more active IPO and deal environment can hurt long-duration growth stocks while simultaneously increasing revenue opportunities for banks through trading, underwriting and advisory fees. |