Growth Demand Meets the Cost of Capital

The cleanest signal is not that growth disappeared

The October 14 tape tested the idea that strong earnings growth and resilient demand can carry a broad group of software, AI-infrastructure and discretionary names through the next year. The answer was mixed: demand may still be durable, but the market is charging more for capital, and that matters most where growth depends on heavy financing or very long-lived cash-flow assumptions.

The S&P 500 proxy SPY fell 1.08% to a 16:00 ET close of $769.60, while the Nasdaq-100 proxy QQQ fell 1.97% to $728.95. Technology ETF XLK declined 2.98% and software ETF IGV declined 3.00%. These are close-to-close observations, not a claim about every stock or the entire session’s breadth.^[1]

The market’s more revealing split was inside the research scope. DataDog fell 4.94% to $269.33, Snowflake fell 6.34% to $335.08, and CoreWeave fell 15.46% to $65.78. AutoNation declined 1.37%, RH fell 4.22%, and La-Z-Boy slipped 0.80%, while Williams-Sonoma rose 1.04%. The regular-session prints came from the 16:00 ET close; in post-market trading, DDOG was $266.00 as of 19:59 ET, SNOW was $333.67 as of 19:58 ET, and CRWV was $61.87 as of 19:59 ET.^[1]

What changed: the discount rate became the headline

Current macro data describe an economy that is not in recession, but not an easy backdrop for richly valued growth either. September data show unemployment at 4.2%, CPI inflation at 3.35% year over year, a federal-funds rate of 3.75%, a 10-year Treasury yield of 5.24%, and real GDP growth of 2.1% year over year. Consumer sentiment was 51.7, down 11.17% year over year.^[2]

That combination helps explain why the market could simultaneously respect demand and punish duration. A company can keep winning customers while its equity multiple contracts. The pressure is greatest when the investment case requires both sustained growth and confidence that large infrastructure commitments will earn attractive returns.

Reuters reported that hotter-than-expected inflation pushed rate-hike worries back into the Nasdaq session, with AI and data-center shares among the prominent decliners. A separate Reuters report described debt-heavy AI firms as especially exposed to the change in financing expectations.^[3]

The scope splits into two tests

The table is a framing device, not a breadth statistic. It shows why “resilient demand” is too broad a thesis to test in one sentence. For DDOG and SNOW, the key issue is whether observability and data-cloud usage remain mission-critical as customers scrutinize budgets. For CRWV, the issue is more capital-intensive: whether contracted or expected demand can outrun the cost and availability of financing. The October 14 price action does not resolve those questions, but it shows which questions investors are asking first.

The consumer names provide a useful counterweight rather than a clean all-clear. WSM’s gain alongside declines in RH, AN and LZB suggests that demand is not moving as one block. Brand positioning, merchandising, housing exposure, promotional intensity and company-specific execution can matter as much as the macro tape. The data support dispersion, not a sweeping conclusion that the consumer is either healthy or broken.

What would confirm or weaken the hypothesis

Evidence that would confirm it:

• Software companies report stable or improving usage, customer retention and expansion without relying only on price increases.
• AI-infrastructure revenue growth is accompanied by improving cash conversion and clearer financing economics.
• Consumer companies show that traffic and full-price demand can withstand weak sentiment and elevated long-term yields.
• Market reactions improve when companies deliver, rather than requiring increasingly aggressive future assumptions.

Evidence that would weaken it:

• Customers continue to optimize cloud and software budgets, delaying expansion even when the product remains strategic.
• Financing costs, capital intensity or customer concentration overwhelm AI infrastructure growth.
• Housing-sensitive categories see weaker traffic, larger promotions or margin pressure.
• Higher inflation keeps long-term yields elevated, compressing multiples faster than earnings can compound.

The base-rate lesson is that earnings growth is not the same as shareholder returns over a one-year window. The thesis needs both operating follow-through and a financing environment that does not erase the value of that growth. Today’s session supplied evidence for the first possibility only in the abstract; it supplied a live warning on the second.

What to watch next

• Earnings evidence: DDOG is scheduled for November 5, 2026, before the open, with the calendar labeling the date estimated. AN is scheduled for October 30, 2026, before the open, also estimated. RH is scheduled for December 10, 2026, after the close; WSM for November 18, 2026, before the open; and LZB for November 17, 2026, after the close. SNOW is scheduled for December 2, 2026, after the close. These dates are calendar estimates, not confirmed company announcements.^[4]
• Funding conditions: Watch long-term yields and credit spreads alongside CRWV and other AI-infrastructure financing signals. The macro snapshot showed the 10-year yield at 5.24% and high-yield spread at 3.15%.^[2]
• Demand quality: Separate usage, retention and profitable expansion from headline bookings or capacity commitments.
• Consumer dispersion: Compare traffic, inventory and promotions across AN, RH, WSM and LZB rather than treating one day’s sector move as a verdict.
• Price confirmation: A recovery that broadens beyond the largest growth names would make the resilient-demand case more credible; continued concentration in defensives would argue that the market still distrusts duration.

The conclusion is deliberately conditional. The supplied growth-and-demand thesis remains plausible for parts of the scope, but October 14 showed that resilience must be measured against the cost of capital, not just against revenue ambition. That is the test the next round of company updates will have to pass.