Oil Shock Tests AI Demand and Home Spending

The market is separating two claims that are often bundled together: that enterprise AI demand can keep producing software growth, and that consumer demand can stay resilient through a geopolitical energy shock. On October 23, those claims were being priced differently.

Brent crude rose to $116.27 a barrel after Houthi attacks in Saudi Arabia intensified fears of energy-supply disruption, while the Nasdaq recovered modestly after the prior session’s AI-related selloff. The Dow remained under pressure, and the 10-year Treasury yield was just below 5%.^[1] This is not a clean “risk-off” signal. It is a market asking which revenue streams are tied to production workloads and which depend on household confidence, freight costs, financing conditions or discretionary budgets.

The market tell: software rebound, furnishings pressure

At the 16:00 ET close, DDOG was $269.65, up 4.62%, and SNOW was $325.43, up 3.65%. CRWV closed at $54.235, down 2.46%, while AN fell 0.49%, RH 1.94%, WSM 3.16% and LZB 1.38%. ETH was up 0.40%. In after-hours trading at 16:27 ET, DDOG was $269.8312, only 0.07% above its close; SNOW was $326.9125, 0.46% above its close; and CRWV was $54.35, 0.21% above its close.^[2]

The useful signal is the relative split, not any single daily percentage. Software names with direct exposure to cloud, data and AI workloads were able to rebound even while oil and geopolitical risk were rising. The consumer and home-related names in this scope did not show the same resilience. That is consistent with a market that still believes some enterprise workloads are mission-critical, but is less willing to assume that discretionary demand will absorb a new energy and inflation shock.

Why the geopolitical shock matters for this scope

The immediate channel is energy. Reuters reported that Houthi attacks in Saudi Arabia raised fears of further disruption to Middle East supplies, with Brent and U.S. crude both rising sharply; the article also noted that the 10-year Treasury yield remained below 5% after a large prior-day decline.^[1] Higher oil can work through transportation, materials, utilities and household budgets. If the shock persists, it can also complicate the rate path by keeping inflation expectations elevated even as growth concerns increase.

For software, the first-order effect is less direct. DDOG, SNOW and CRWV are still exposed to capital spending, cloud budgets and the valuation effect of higher long-term yields. But their operating drivers can be more specific than “the economy”: data ingestion, observability, AI applications and production workloads. That distinction is why software can rebound on a day when the macro backdrop is deteriorating without proving that the geopolitical risk is harmless.

DDOG and SNOW: the hypothesis has real operating evidence

Datadog’s latest available earnings evidence is unusually relevant to the hypothesis. In its Q2 FY2026 call, management reported $1.12 billion of revenue, up 36% year over year, and said growth among non-AI customers had accelerated to the high 20s. It also described roughly 33,400 customers and about 4,720 customers with at least $100,000 of annual recurring revenue.^[3]

The more important point is the mechanism. Datadog said AI-native and broader enterprise customers were increasing cloud and modern-technology usage, while the company expanded products for observing, securing and acting on AI workloads. In the Q1 FY2026 call, management said more than 6,500 customers were sending data through at least one AI integration and that those customers represented about 80% of ARR.^[3] That supports the idea that AI demand is becoming a measurable workload and platform opportunity, not only a stock-market label.

But the qualification matters. Datadog’s model is usage-sensitive. Customers can optimize cloud and observability consumption, and one large renewal can affect reported growth. The bullish case therefore requires production workloads to broaden faster than optimization and renewal pressure can offset them.

Snowflake offers a similar but not identical read. In its Q2 FY2027 call, product revenue grew 37% year over year for the third consecutive quarter of acceleration. Management said net-new customer additions increased 32% year over year and that the platform supported more than 41% of the Global 2,000.^[4] The company described AI as a multiplier because it can bring new workloads onto the platform, increase adoption of Snowflake’s own AI products and lift broader platform consumption.

That is stronger evidence than a product-launch narrative alone. Still, Snowflake also uses a consumption model, so the durability test is whether AI projects become recurring production workloads rather than short-lived experiments or pulled-forward usage. Even management’s own discussion of the business emphasizes that consumption can normalize after migrations and new workloads arrive.^[4]

Home and discretionary demand: a harder test

The home-related part of the hypothesis is more conditional. RH, WSM and LZB remain exposed to housing turnover, household confidence, imported merchandise and freight. AN adds sensitivity to vehicle affordability and financing conditions. A sustained oil shock can pressure those businesses through both the consumer and cost sides of the income statement.

The case for resilience is not absent, but it is company-specific. Williams-Sonoma’s earlier operating evidence showed share gains even while the broader home-furnishings category was weak, which would make execution and brand strength more important than a broad housing recovery. RH’s expansion strategy similarly depends on inventory, new locations, international execution and demand for higher-ticket discretionary goods. Those are valid growth levers, but they do not amount to proof of sector-wide demand resilience.

The current tape reinforces that distinction: WSM and RH were down on the regular close while DDOG and SNOW rose.^[2] One session cannot establish a trend, and this research pass found less comparable fresh transcript evidence for RH, WSM, ETH, LZB and AN than for DDOG and SNOW. The appropriate conclusion is not that the consumer names cannot grow; it is that the burden of proof is more operational and more exposed to macro friction.

CRWV is its own bridge case. AI infrastructure demand can support revenue growth, but the business also carries capital intensity, financing, customer-concentration and execution risks that make it more sensitive to the market’s changing view of AI spending. Its lower close on a day when DDOG and SNOW rose is a reminder that “AI exposure” is not a single factor.

What the evidence says—and does not say

The base case is therefore split. The hypothesis is best supported in DDOG and SNOW, where management has described accelerating workloads, customer expansion and AI adoption. It is less supported as a blanket claim across RH, WSM, ETH, LZB, AN and CRWV. In those names, the next leg depends on share gains, capital discipline, financing conditions, supply-chain costs or proof that AI infrastructure spending converts into durable economics.

What to watch next

1. Energy and shipping: whether attacks or countermeasures produce an actual reduction in oil flows, insurance availability or shipping capacity, rather than only a risk premium. Brent’s move above $116 was the market’s current warning signal.^[1]
2. Rates and inflation expectations: whether the 10-year yield stays near 5% as oil rises. That combination would be a tougher valuation and demand backdrop for both software multiples and home-related purchases.
3. DDOG: whether non-AI growth remains near the high-20% range, whether AI usage continues moving into production and whether large renewals show optimization pressure.^[3]
4. SNOW: whether the 37% product-revenue trajectory and AI-driven consumption persist after the current acceleration cycle, with particular attention to production workloads rather than announced projects.^[4]
5. CRWV: utilization, customer concentration, financing and cash economics—not just demand headlines.
6. RH, WSM, ETH, LZB and AN: comparable-store or revenue trends, pricing, tariff and freight effects, inventory discipline and evidence that customers are spending despite higher energy and financing costs.

The market is not yet rejecting the earnings-growth hypothesis. It is sorting it. Workload-linked software has the clearest current operating confirmation; home and other discretionary exposure needs more company-specific evidence; and the geopolitical energy shock is the variable that can challenge both through rates, confidence and costs. That is a narrower and more defensible conclusion than treating all eight names as one resilient-demand basket.

This article is research commentary, not financial advice.