AI demand is still real. The macro test is getting harder.

The current market question is not whether artificial-intelligence and cloud spending exist. Recent operating evidence says they do. The harder question is whether that growth can keep compounding fast enough to offset a more hostile valuation backdrop: higher energy costs linked to the Iran conflict, rising government-bond yields, and a Federal Reserve that has resumed tightening.

That is the market tell for the scoped group of DDOG, SNOW, RH, WSM, ETH, LZB, LESL, and TPX: the secular-demand thesis has not broken, but its margin for error has narrowed.

The macro shock is reaching growth assets through rates and energy

A Sept. 20 market review described a renewed combination of Middle East conflict, oil above $100 a barrel, and a 10-year Treasury yield above 5%. It also reported that the Federal Reserve delivered its first rate increase since 2023 as policymakers confronted the inflation risk from the war.^[1]

Those are not merely headline risks for software. Higher energy prices can squeeze households and discretionary purchases, while higher bond yields reduce the present value investors assign to cash flows expected further in the future. That creates a difficult setup for both sides of the scope: growth software must keep proving its earnings trajectory, while furniture, mattress, and home-furnishing names must show that demand can withstand a higher cost of living and financing costs.

The causal chain is still uncertain. The conflict may de-escalate, oil may retreat, or productivity gains may improve the growth outlook. But the market is now asking for more evidence before paying for distant growth.

DDOG: operating momentum remains tangible

Datadog’s second-quarter 2026 release showed revenue of $1.12 billion, up 36% year over year. The company also reported roughly 4,720 customers with at least $100,000 of annual recurring revenue, up from about 3,850 a year earlier, and guided to full-year revenue of $4.45 billion to $4.47 billion.^[2]

The transcript evidence adds an important detail: management said more than 750 AI customers were using Datadog, including all 10 of the leading AI companies it tracks, and that agentic activity measured through MCP tool calls had grown more than 22-fold from Q4 2025.^[3]

That is the bullish side of the hypothesis. AI-related infrastructure is not only a valuation story; it is generating observable usage and enterprise expansion. The counterpoint is that consumption businesses can be volatile when customers optimize workloads or negotiate volume discounts. Earlier management commentary explicitly described that risk, even while calling AI a long-term tailwind.^[3]

SNOW: acceleration is broadening, but consumption cuts both ways

Snowflake’s Q2 fiscal 2027 transcript showed product-revenue growth accelerating to 37% year over year for a third consecutive quarter. Management attributed the result to strength in the core data platform and a meaningful step-up in AI revenue; net-new customer additions rose 32% year over year, and the company had 829 Global 2000 customers.^[4]

That supports the same basic conclusion as DDOG: AI is becoming a demand catalyst inside an established cloud-data workflow, not just a product announcement. But Snowflake’s consumption model makes the evidence two-sided. Management has said product revenue and customer-consumption patterns are more informative than any single quarterly beat, while analysts have raised the possibility of “sticker shock” or optimization as AI agents create more workloads.^[4]

For SNOW, the next proof point is therefore not simply headline growth. It is whether new AI workloads translate into durable customer expansion without a later pullback in consumption.

The consumer names face a different geopolitical transmission channel

RH, WSM, ETH, LZB, LESL, and TPX do not have the same direct AI-demand exposure. Their test is household resilience. Higher fuel and utility costs can compete with discretionary home purchases; higher rates can weigh on housing turnover, credit conditions, and the affordability of large-ticket goods.

The available price snapshot shows that the group is not trading as one block. On the Sept. 18 close, WSM rose 2.41%, TPX rose 1.04%, and LZB rose 0.88%, while LESL fell 9.11% and RH slipped 0.24%. DDOG fell 2.58% and SNOW fell 1.76% that day.^[5] The dispersion argues against treating the macro shock as a complete rejection of the growth thesis or as a uniform consumer collapse.

There is also a data-quality caution: the quote feed returned a stale February 2025 regular-session observation for TPX rather than a current Sept. 2026 snapshot. That name should not be used as a precise read on the latest trading session until a current source is available.^[5]

What the market is actually testing

The hypothesis that earnings growth and resilient demand can support this group over the next year remains plausible, but it now requires two conditions to hold at once:

1. Software demand must convert AI activity into recurring revenue. DDOG and SNOW have supplied credible evidence of usage, customer additions, and AI-related expansion. The risk is that optimization, pricing concessions, or slower enterprise budgets interrupt the conversion.
2. Consumer demand must absorb an energy-and-rates squeeze. The home and furnishings names need to show that customers are still willing and able to make discretionary purchases even as inflation and financing conditions become less comfortable.

The market’s recent behavior fits a “narrower margin of error” interpretation better than a clean break. DDOG and SNOW have both produced strong operating evidence, yet their latest closes were lower. Meanwhile, several consumer names advanced while LESL sold off sharply. Price action is signaling selectivity, not a single verdict.

What to watch next

• Oil and the Strait of Hormuz: whether the conflict produces sustained supply or shipping disruption rather than a short-lived premium. A persistent energy shock would be the clearest route from geopolitics to consumer demand and inflation.
• The 10-year Treasury yield and Fed reaction: whether yields remain above the psychologically important 5% area and whether policymakers describe energy inflation as temporary or as a risk to expectations.^[1]
• DDOG usage quality: expansion among large customers, AI-observability adoption, free-cash-flow conversion, and any renewed discussion of optimization or volume discounts.
• SNOW consumption durability: product-revenue growth, net-new Global 2000 additions, AI workload monetization, and evidence that agent-driven usage is incremental rather than offset by later cost controls.
• Consumer demand indicators: traffic, order backlogs, promotions, cancellations, and management commentary from RH, WSM, ETH, LZB, LESL, and TPX. The key distinction will be resilient demand versus demand purchased through heavier discounting.

The base case is not “geopolitics wins” or “AI wins.” It is that strong company-level execution can still matter, but the macro hurdle has risen. If energy and yields stabilize, the operating evidence in DDOG and SNOW can regain influence. If they keep climbing, even real demand growth may not be enough to protect long-duration valuations or discretionary spending.