Nasdaq Notes: The Market Moves from Narrative to Cash Flow

On Friday, October 30, the Nasdaq fell 2.86% to close at 25,437, its biggest one-day drop since June 5 and 7.8% below its record close on October 6. It fell 2.0% for the week, its third losing week in a row. For the whole of October it fell 5.3%, its worst month since March 2025; by Reuters' count, Nasdaq-100 stocks lost more than $2 trillion in market value this month.

The most informative thing this week was how a few big companies diverged after earnings. Their AI stories were much the same, yet their stocks went in completely different directions, and the dividing line was cash flow.

Both spending, so why did one rise and one fall

After Wednesday's close, Microsoft and Alphabet reported at the same time.

Microsoft's Azure revenue grew 48%, with guidance of 49% to 50% for next quarter. Capex including finance leases was $45.2 billion, below the company's own guidance; free cash flow was $18.6 billion, about $10 billion more than analysts expected. Management also said free cash flow will stay positive throughout fiscal 2027. The stock rose about 4% after hours.

Alphabet's Google Cloud revenue doubled, so growth alone was hardly weak. But capex for the quarter was $51.4 billion and free cash flow was negative $9.07 billion, against analyst expectations of negative $2.7 billion. The stock fell 12.6% on Thursday, its worst day since its 2004 IPO, wiping out about $500 billion in market value.

What unsettled the market more was search. Google said growth in search queries was the strongest in five years, yet search revenue missed expectations for a second straight quarter. Chief business officer Philipp Schindler attributed part of this to AI Mode: these queries monetize at a lower level. Capex is accelerating while the money each unit of core-business usage brings in is shrinking, so cash flow is squeezed from both ends.

Both are investing heavily in AI. The difference is whether that investment is covered by operating cash flow. Microsoft's answer is that it can be; Alphabet's answer is that for now it can't, and the unit price of its core revenue is heading down.

Amazon and Intel: capex above expectations

After Thursday's close, Amazon reported third-quarter capex of $77.3 billion, nearly $19 billion above the expectations compiled by FactSet; free cash flow over the past twelve months was a net outflow of $42.4 billion, against an inflow of $14.8 billion a year earlier. AWS grew 39%, and management said capex will be significantly higher again next year.

Intel's quarterly profit beat expectations, but its fourth-quarter revenue guidance of $15 billion to $16 billion was below the $17.1 billion the market expected, and the company signaled that free cash flow would struggle to turn positive in the near term. The stock fell 28.6% on Friday, its worst day since 1974. The Philadelphia Semiconductor Index fell 7.2% that day and more than 15% for October; AMD fell 10.9% and Nvidia 4.2%. With Nvidia's market value close to $5 trillion, that 4.2% weighed on the S&P 500 more than Intel did.

On the surface the chip selloff was set off by Intel. Underneath, the same question is travelling up the supply chain. If capex at the big downstream companies starts to be constrained by free cash flow, the growth rate of chip orders over the next few years has to be discounted again.

The valuation anchor has changed

In the narrative-led phase, the market priced AI-related companies on long-term market size and revenue growth, and capex was treated as investment in the future: the more, the more it showed commitment. Once capex grows large enough to swallow operating cash flow, it stops being one line on the cash flow statement and becomes the core variable that determines how much cash shareholders ultimately get.

In a discounted cash flow framework, free cash flow is the numerator. If the numerator for the next few years turns negative, the company's value depends more on cash in distant years; its duration lengthens and it becomes more sensitive to the discount rate. At that point valuation has to answer more than whether the story is good. It has to answer how long you wait before the story pays, and at what rate it is discounted back.

This also explains why the market wasn't worried when Microsoft issued $25 billion of bonds on Thursday. Its operating cash flow covers its capex, so borrowing optimizes its capital structure rather than filling a gap.

Rates didn't help

The other variable is rates. On Wednesday the Fed held steady, keeping the federal funds rate at 3.75% to 4%, but the vote was 10 to 2, with two officials favoring a hike. Data released on Thursday showed third-quarter GDP growing at an annualized 3.4%, above expectations; September PCE was 3.6% year on year and core PCE 3.1%. According to CME data, the market is pricing about a 75% chance of a hike in December.

The stronger the economy and the stickier inflation, the less room short-term rates have to fall. The 10-year Treasury yield slipped back to around 4.94% this week, more because of safe-haven buying spilling out of the tech selloff than because of any hope of rate cuts. For long-duration assets this is an uncomfortable mix: the numerator is shrinking, the discount rate isn't coming down noticeably on looser monetary policy, and the short end is still pricing toward a hike.

Where the money went

The index's fall hid a rotation under the surface. This week the S&P 500 fell 0.6%, while the equal-weight S&P 500 actually rose 0.8%; the Russell 2000 rose 1.3%, 9 of 11 sectors gained, and health care hit a record high. As of Thursday, a large-cap ETF excluding the Magnificent Seven was up more than 15% this year, against less than 6% for a Magnificent Seven ETF.

Money hasn't left stocks. It is looking for places where current cash flow is steady and valuations haven't been lifted by the AI narrative. What energy, consumer staples and health care have in common is that their cash flow is visible and doesn't depend on a long-term assumption that hasn't yet been proven.

What I'm watching next

Next Monday Apple and Palantir report, Tuesday is AMD, the same day as the US midterm elections, and Friday brings October payrolls.

I'll focus on three things. First, how AMD describes data center demand, to see whether the capex constraints at the big companies are showing up in concrete orders. Second, each company's cash flow statement, not just revenue and earnings per share: how much of capex operating cash flow can cover, and how they plan to fill the gap. Third, the payrolls data and the pricing of a December hike, which decide whether the discount rate on long-duration assets keeps climbing.

This week the market didn't reject demand for AI. What changed is the basis for pricing: from revenue growth and long-term potential to how much free cash flow can be produced now. I don't call levels. I just write down what I'm watching and why.

This is just my own reading, not investment advice.