The Index Looked Calm. AI Didn't.

GOOD MORNING, Wednesday looked quiet at the index level. It wasn't. The S&P 500 slipped just 0.14% and the Dow was virtually flat, but Brent climbed to a six-week high near $95 as U.S.-Iran fighting threatened more Middle East shipping. Then after the close, Alphabet delivered explosive 82% cloud growth — and still raised its 2026 capex plan by another $15 billion while posting negative free cash flow for the first time in its history. Tesla also burned cash. The market's next AI question is no longer whether demand exists. It is how much capital that demand will consume.

MARKETS | TLDR

Wall Street barely moved ahead of Big Tech earnings: The S&P 500 fell 0.14%, the Dow was nearly unchanged and the Nasdaq lost 0.57%, masking much larger moves underneath the major indexes.
Oil climbed to a six-week high: Brent approached $95 as renewed U.S.-Iran fighting and shipping disruption in both the Strait of Hormuz and Red Sea pushed energy risk back into the inflation story.
Treasury yields stayed elevated: The 10-year yield hovered around 4.6%–4.7%, leaving expensive growth stocks with little valuation relief just as AI spending plans moved higher again.

ALPHABET

Google Cloud Grew 82%. Investors Still Asked About the Bill.

Alphabet reported its strongest-ever quarter of Google Cloud growth Wednesday, with revenue jumping 82% to $24.8 billion as enterprise demand for AI compute accelerated. Total revenue reached $119.8 billion, above Wall Street expectations, while advertising revenue also came in slightly ahead of estimates.

The problem was what it took to produce that growth. Alphabet raised its 2026 capital-spending forecast by $15 billion to $195 billion–$205 billion and reported negative free cash flow of $5.9 billion — the first cash-burn quarter in company history. Management said demand still exceeded the capacity it had added and signaled another significant increase in capex for 2027.

That is the tension the rest of Big Tech now has to answer. AI demand is clearly real, but the infrastructure required to serve it is becoming so capital-intensive that even historically cash-rich companies are starting to look different financially. Investors may tolerate the spending while cloud revenue accelerates. But with capital carrying a real cost again, the margin for execution error is shrinking.

AI CAPITAL

The AI Boom Is Starting to Eat Its Own Cash Flow

The broader numbers make Alphabet's quarter look less like an exception and more like an industry shift. Reuters analysis of Microsoft, Alphabet, Amazon, Meta and Oracle shows AI infrastructure spending rising faster than operating cash flow, with consensus estimates suggesting combined capex could exceed free cash flow by 2027.

That matters because the first phase of the AI trade was financed from enormous cash cushions. The next phase may rely more heavily on debt, leases and outside infrastructure partners. Reuters estimates that for every additional dollar of operating cash flow generated by the group through 2027, roughly $1.57 could be absorbed by incremental capital spending. Oracle is already spending far more on capex than it generates in operating cash.

There is still a strong bull case: Microsoft has said its AI business is running at a roughly $37 billion annualized revenue rate, and demand across cloud, servers and chips remains intense. But investors are starting to demand a timeline for returns. If AI monetization does not catch up with infrastructure spending over the next several years, buybacks, dividends and valuation multiples all become harder to defend.

HEADLINES

Tesla returned to cash burn as AI and robotics spending surged: Tesla posted negative free cash flow of $1.1 billion, its first negative quarter in more than two years, while second-quarter capex more than doubled to $5.8 billion.
Super Micro said fourth-quarter orders topped $60 billion: Shares surged as the AI-server maker raised its gross-margin outlook, offering a powerful counter-signal that infrastructure demand itself remains exceptionally strong.
Texas Instruments guided above expectations: Stronger industrial, automotive and data-center demand showed the AI buildout spreading into analog chips, even though TI shares fell after hours.
Southwest cut the floor of its profit outlook: Higher fuel costs offset improving pricing and ancillary revenue, showing how the Middle East energy shock is already passing from commodity markets into corporate margins.
The gap between index volatility and single-stock volatility hit a record: Low correlations have kept headline indexes unusually calm even as individual stocks swing violently — a diversification benefit that may be masking more risk than the VIX suggests.

UPCOMING

ECB decision — July 23: The ECB is expected to hold after June's hike, but oil near $95 makes any language on energy inflation and future tightening more important.
FOMC meeting — July 28–29: Markets expect the Fed to hold rates, but another jump in oil or inflation expectations could keep a later hike firmly in play.
Microsoft and Meta earnings — July 29: Both companies will be judged on whether rising AI infrastructure spending is producing enough incremental revenue and cash flow.
Amazon and Apple earnings — July 30: AWS will extend the hyperscaler capex debate, while Apple faces a different test — whether it can close the AI product gap without adopting the same capital intensity.

DEEP INSIGTHS

AI Investment Boom Puts Big Tech's Free Cash Flow Under Pressure

Read this for the balance-sheet version of the AI story. The important shift is that hyperscalers are moving from financing AI almost entirely with excess internal cash toward a model that increasingly involves debt, leases and external infrastructure. That makes the cost of capital a much more direct variable in AI economics.

Markets Expect a Diversification Miracle

This explains why the headline indexes can look calm while the underlying market feels violent. The spread between index volatility and average single-stock volatility has reached a record because correlations are unusually low. That works while winners and losers offset one another — but it also means traditional measures like the VIX may be understating how much risk is building beneath the surface.