The Fed Held. AI Cracked Anyway.

GOOD MORNING, The Fed did nothing — and markets still tightened. Policymakers held rates at 3.50%–3.75%, but three officials wanted an immediate hike and Chair Kevin Warsh gave investors little reason to expect relief soon. The S&P 500 fell 1.5% and the Nasdaq dropped 1.7%, extending a selloff that has increasingly centered on one question: can Big Tech keep spending at AI-scale without sacrificing too much cash flow? Microsoft and Meta reported after the close. Their answers could decide whether this is a normal valuation reset or something deeper in the AI trade.

MARKETS | TLDR

Wall Street sold off after the Fed held rates: The S&P 500 fell 1.52% and the Nasdaq lost 1.74%, with technology and industrials among the weakest sectors.
The Fed delivered a hawkish hold: Rates stayed at 3.50%–3.75%, but the 9–3 vote included three dissenters who preferred a 25-basis-point hike.
The AI selloff was already spreading globally: South Korea’s KOSPI had fallen 12% from its peak and was on track for a brutal monthly decline as investors questioned whether massive AI spending could keep translating into cash flow.

FED

The Fed Paused. Markets Heard “Not Done.”

The Federal Reserve kept the federal funds target at 3.50%–3.75% on Wednesday, but the vote was more hawkish than the unchanged rate suggested. Beth Hammack, Neel Kashkari and Lorie Logan dissented in favor of a 25-basis-point increase. The statement said economic activity was still expanding at a solid pace, capital investment remained strong and inflation was elevated relative to the 2% goal, partly because of energy and other supply shocks.

For markets, the lack of a hike did not translate into easier financial conditions. Investors were already wrestling with higher oil prices and resilient growth, and the three dissents made clear that another increase remains a live possibility. The S&P 500 fell 1.52% and the Nasdaq lost 1.74%, while rate expectations shifted toward a greater probability of tightening later in the year.

The bigger implication is that Big Tech cannot assume the Fed will rescue valuations while earnings remain strong. A high-rate environment raises the hurdle for every dollar of AI capex, especially when the return arrives years later. The Fed’s message was not that policy must tighten immediately. It was that investors should not count on easier money to make expensive growth assets work.

AI CAPEX

The AI Trade Is Now Asking for Receipts

The selloff in AI-linked stocks deepened before Microsoft and Meta reported after the close. SK Hynix posted a 557% jump in second-quarter operating profit, yet the stock still fell because results did not clear already extreme expectations. South Korea’s KOSPI, one of the biggest beneficiaries of the AI boom, had dropped roughly 12% from its recent peak. In the U.S., Vertiv and other AI-infrastructure names were also under pressure after disappointing earnings and margin signals.

The shift is important because investors are no longer debating whether AI demand exists. They are debating whether the spending required to serve that demand is generating enough incremental cash flow. Alphabet and Tesla had already unsettled the market with heavy investment and weaker free-cash-flow dynamics. That raised the stakes for Microsoft and Meta, two companies spending aggressively but with very different paths to monetization.

After the bell, the divergence became visible. Microsoft reported 43% Azure growth and said it expected AI investments to keep generating cash, sending the shares higher in extended trading. Meta, by contrast, raised its 2026 capital-spending outlook and fell after hours. The market’s new rule is becoming clearer: AI capex still gets rewarded, but only when investors can see a convincing path from infrastructure to cash.

HEADLINES

Microsoft’s Azure growth beat expectations: Azure grew 43% and Microsoft forecast 45% constant-currency growth for the current quarter, giving investors one of the clearest examples yet of AI infrastructure spending translating into cloud revenue.
Meta raised its AI spending plan again: The company lifted 2026 capex guidance to $130 billion–$145 billion, keeping pressure on free cash flow and forcing investors to ask when its compute buildout will produce a Microsoft-style return.
Qualcomm warned that Apple weakness will get worse: A weaker-than-expected profit outlook and faster decline in Apple-related revenue pushed Qualcomm to lean harder on automotive and AI data centers as its next growth engines.
Iran risk kept oil in the macro conversation: Renewed military tension and missile interceptions added another inflation variable on the same day the Fed emphasized that energy-related supply shocks were keeping prices elevated.
The Nasdaq was nearly 9% below its recent peak: The drawdown showed that this was no longer a one-stock reaction — AI capex concerns had become a broader valuation problem for growth equities.

UPCOMING

Amazon earnings — July 30: AWS will be another test of whether hyperscaler AI spending is translating into faster cloud growth quickly enough to justify even larger capex budgets.
Apple earnings — July 30: Investors will be watching iPhone demand, supply constraints and whether Apple can close the AI narrative gap with Microsoft, Google and Meta.
Bank of Japan meeting — July 30–31: The BOJ’s two-day meeting comes as the yen remains under pressure and global long-term yields are moving higher.
U.S. Employment Cost Index — July 31: Wage and benefit costs will provide another inflation signal immediately after a Fed meeting that featured three votes for a hike.

DEEP INSIGTHS

Federal Reserve FOMC Statement — July 29, 2026

Read the statement itself because the headline “Fed holds” misses the important part. Three policymakers wanted a hike, the committee described activity and capital investment as strong, and it explicitly tied elevated inflation partly to energy supply shocks. That combination explains why equities sold off even without a rate increase.

Chip Rout Moves Into Panic Territory

This is useful because it captures the AI trade before Microsoft’s earnings temporarily restored confidence. The piece connects SK Hynix, Korea’s equity selloff, Alphabet and Tesla’s cash-flow concerns, and rising oil into one question: how long can investors tolerate extraordinary AI spending when both expectations and the cost of capital are rising?