Demand is holding—but the market is pricing the cost

The broad-market tape is not rejecting growth. It is becoming more selective about which kind of growth deserves confidence.

On Friday, October 23, SPY fell 0.53% to its 16:00 ET close, QQQ was essentially flat, down 0.02%, and DIA fell 1.23%. Technology held up better than the broader tape, but not every technology or AI exposure was rewarded.^[1]

The working hypothesis was that earnings growth and resilient demand could support DDOG, SNOW, CRWV, AN, RH, WSM, ETH and LZB over the next year. The latest evidence turns that into a conditional statement: enterprise usage is providing the cleanest support, while infrastructure financing and discretionary consumption remain the harder parts of the test.

The tape is separating demand from the cost of demand

From April 27 through October 23, DDOG rose 103.34% and SNOW 125.59%, while CRWV fell 51.60%. WSM gained 31.41%; RH was nearly unchanged at 1.05%; AN declined 21.07%. Volatility was high across the group: CRWV’s annualized volatility was 95.05% and DDOG’s 73.80%.^[2]

Today’s snapshot reinforces the dispersion. DDOG closed at $269.65, up 4.62%; its 16:07 ET after-hours price was $269.25, down 0.15% from the close. CRWV closed at $54.24, down 2.46%, but its 16:07 ET after-hours price was $54.37, up 0.25%. SNOW closed at $325.44, up 3.66%. WSM, RH, AN and LZB fell 3.17%, 1.91%, 0.51% and 1.38%, respectively, while ETH finished at $22.82, up 0.33%. These are observations, not a conclusion that any move will persist.^[1]

Enterprise software is the strongest evidence for resilient demand

Datadog’s latest available call offers a direct operating signal. In Q2 FY2026, management said revenue grew 36% year over year to $1.12 billion, customer count reached about 33,400, and growth among non-AI customers accelerated to the high 20s. Management also said more than 750 AI customers were using Datadog.^[3]

Snowflake’s Q2 FY2027 commentary pointed in the same direction: AI was bringing new workloads onto the platform, lifting broader consumption, and contributing to more than 400 basis points of year-over-year expansion in non-GAAP operating margin to 15%. The qualification is important: consumption businesses can expose customers to “sticker shock” as agents and applications proliferate, so adoption and cost control have to develop together.^[4]

That gives software the cleanest evidentiary base: usage, workload migration and customer expansion are showing up in management commentary. DDOG and SNOW’s high six-month returns are consistent with that evidence, although their maximum drawdowns were still 27.39% and 19.35%.^[2]

AI infrastructure has demand—but also a financing bill

CoreWeave’s Q2 FY2026 commentary supported the demand side: revenue was $2.6 billion, up 112% year over year, and revenue backlog was $104.2 billion, up 246%. Management also argued that inference turns compute into an ongoing requirement rather than a one-time training expense.^[5]

Those facts do not remove the capital-intensity question. Earlier commentary described capex expansion, debt-funded growth and the complexity of delivering power, cooling, networking, servers and orchestration. CRWV’s 51.60% six-month decline, 95.05% annualized volatility and 60.70% maximum drawdown show that investors are weighing execution, financing and duration risk alongside backlog.^[5][2]

A durable AI workload can still produce an unattractive equity outcome if capacity arrives late, capital costs rise, customer concentration remains high, or revenue arrives after a large investment cycle. Demand is necessary evidence; it is not the whole earnings equation.

Consumers are resilient in pockets, not uniformly

WSM’s six-month return was positive, but the stock fell 3.17% today. RH was roughly flat over the comparison window and down 1.91%; AN and LZB were also lower. That is not proof of a consumer contraction, but it is consistent with a market asking more questions about financing conditions, housing turnover and discretionary budgets.^[1][2]

RH’s latest available call illustrates the friction. Management said demand was outpacing many other home-furnishing companies but at a cost to margin, and described tariffs and the difficulty of moving furniture, lighting and rug production across factories and countries.^[6]

This is the counterweight to the software evidence. A household may still want a product, but purchase timing, financing costs, import costs and margin protection determine whether demand becomes profitable growth. WSM, RH, AN and LZB therefore belong in the “watch the conversion” category: reported demand matters, but so do traffic, ticket, inventory, gross margin and cash generation.

Macro backdrop: growth remains positive, but the hurdle rate is higher

The latest macro snapshot shows real GDP growth of 2.1% year over year and unemployment at 4.2%, with no recession flag. CPI inflation was 3.72%, the 10-year Treasury yield 5.15%, the policy rate 3.75%, and consumer sentiment 51.7. High-yield spreads were 3.27% and the VIX was 18.44.^[7]

That is a mixed backdrop. Positive growth and employment can support software budgets and selected consumer demand. Higher long-term yields and weak sentiment make long-duration growth, housing-linked purchases and capital-heavy infrastructure more sensitive to disappointments. The market is asking not only whether revenue grows, but how much capital, time and margin are required to turn growth into durable cash flow.

The hypothesis, marked to market

The balanced reading is that the hypothesis has real support, but it is not one macro story. Software demand is the clearest confirmation. AI infrastructure is a demand-plus-financing test. Consumer names are a demand-plus-margin-and-financing test.

What to watch next

• DDOG: The next scheduled report is November 4, 2026, before the open; the calendar labels the date estimated and lists November 5 as an alternate. Watch broad non-AI usage alongside AI customer growth.^[8][3]
• SNOW: The next scheduled report is December 2, 2026, after the close, with estimated confidence. Watch product consumption, AI adoption, margin progression and customer optimization.^[8][4]
• AN, RH, WSM and LZB: The calendar lists estimated dates of October 30, December 10, November 18 and November 17, respectively. The dates are not confirmed; the useful signals will be demand conversion, inventories, pricing, tariffs, margins and cash flow.^[8]
• CRWV: The current calendar does not support a scheduled date for this listing, so no date is inferred. Monitor backlog conversion, cash flow, leverage and delivery risk.^[8][5]
• Macro: Track the 10-year yield, credit spreads, sentiment and employment alongside reported demand.^[7]

The next evidence should clarify whether today’s split is temporary positioning or a durable preference: reward visible, repeatable usage; discount growth requiring ever-larger financing commitments; and demand more margin proof from discretionary businesses.