The energy shock is testing AI demand before it tests the consumer

The market’s latest message is selective, not uniformly defensive. On the October 16 close, the scope’s most rate- and AI-sensitive names fell sharply: Datadog (DDOG) dropped 7.2%, Snowflake (SNOW) 8.4%, and CoreWeave (CRWV) 11.9%. By contrast, AutoNation (AN) gained 1.2%, while Williams-Sonoma (WSM) was little changed. RH fell 4.1% and La-Z-Boy (LZB) 1.3%. Those moves are consistent with investors reducing exposure to long-duration growth and questioning the financing intensity of AI infrastructure, while not yet pricing a generalized collapse in discretionary demand.^[1]

The market tell: duration and funding sensitivity are doing the work

The immediate geopolitical backdrop is an energy shock tied to continuing conflict and attacks on energy infrastructure and shipping. UBS reported that markets were assessing whether traffic through the Strait of Hormuz could recover reliably, while also flagging the risk that renewed attacks could reverse improving flows. Its scenario linked energy availability to inflation, bond yields, consumer spending, and corporate margins.^[2]

That transmission matters for this watchlist because the most vulnerable companies are not necessarily those with the largest direct fuel bills. Software and AI-infrastructure businesses are valued partly on cash flows expected years ahead, and higher yields or a higher risk premium can compress those valuations even when end demand remains healthy. The October 16 price action is therefore a financing-and-duration signal first; it is not, by itself, proof that cloud or AI demand has broken.

What the evidence says about AI demand

The Federal Reserve Bank of San Francisco described AI infrastructure investment as a continuing growth engine. It said real investment in information-processing equipment had grown at an average annualized rate above 20% since the start of 2025, while AI-related imports had risen rapidly. At the same time, the institution said the energy and tariff shock was contributing more to headline PCE inflation than information-processing equipment itself.^[3]

Datadog’s earnings-call commentary supplies a company-level check on that demand. In its Q2 FY2026 call, management said infrastructure-product consumption was accelerating across the customer base and described rising usage of GPU monitoring and agent-monitoring products. It also said customer concerns had shifted from validating AI outcomes toward controlling costs.^[4]

That last point is the hinge. Resilient usage can support DDOG, SNOW, and CRWV over the next year, but only if customers keep expanding workloads faster than they optimize spending and if capital markets continue to fund the infrastructure buildout. A cost-conscious customer is not necessarily a lost customer; it is a customer demanding clearer return on compute, software, and data.

Why the rate channel matters

The San Francisco Fed’s October assessment said PCE inflation was 3.4% year over year in August, with elevated energy and core-goods contributions linked to the Middle East conflict and higher tariff rates. It expected inflation to decline as those shocks fade, but also presented a longer-conflict scenario with a higher inflation path. The same analysis said market participants expected one more 25-basis-point federal-funds-rate increase in 2026 and additional hikes in 2027.^[3]

This is the scenario tension behind the share-price split. If energy flows normalize, the market can look through a temporary inflation pulse and return its attention to secular AI adoption and earnings growth. If disruption persists, the combination of higher energy costs, sticky inflation, and tighter financial conditions can keep pressure on long-duration software and heavily financed infrastructure even while usage remains positive.

Consumer names are a useful cross-check

The relative resilience of AN and WSM versus the AI cohort suggests that the market has not yet adopted a single, broad “consumer recession” view. But that does not clear RH, WSM, or LZB of risk. Furniture and home-related purchases are sensitive to financing costs, confidence, housing turnover, and shipping expenses. A prolonged energy shock could reach these companies through household budgets and freight, with the impact depending on pricing power and the health of higher-income consumers.

The current evidence is therefore mixed: AN’s gain and WSM’s flat session point to selective rather than indiscriminate selling, while RH’s decline shows that premium discretionary exposure remains vulnerable. ETH and LZB add another test of whether consumers delay durable-home purchases when rates and geopolitical uncertainty rise; the October 16 moves were modestly negative for both.^[1]

Testing the one-year thesis

The bullish case for DDOG, SNOW, CRWV, AN, RH, WSM, ETH, and LZB requires three things to be true: enterprise and consumer demand remain resilient, companies preserve enough pricing power to offset energy and logistics pressure, and the geopolitical shock does not force a sustained reacceleration in inflation and rates.

The evidence currently supports the first condition more clearly for AI usage than for consumer durables. It supports the second only unevenly. The third is the largest external uncertainty: both UBS and the San Francisco Fed identified energy-flow normalization as important to the inflation and rates outlook, while warning that a longer conflict would produce a less favorable path.^[2][3]

The base-rate interpretation is not that earnings growth has become irrelevant. It is that earnings growth now has to compete with the discount rate and with investors’ scrutiny of capital intensity. That helps explain why a company can report healthy demand signals and still see its stock reprice lower.

What to watch next

• Shipping and energy infrastructure: evidence that Strait of Hormuz traffic is recovering reliably, or fresh attacks that disrupt flows.^[2]
• Inflation and rates: whether energy and tariff effects begin to fade, or whether the longer-conflict scenario starts to look more plausible.^[3]
• AI customer behavior: growth in inference, agent activity, and observability usage versus explicit evidence of cost optimization and slower expansion.
• Funding intensity: whether AI-infrastructure companies can finance expansion without a sustained rise in credit and equity risk premia.
• Consumer durability: whether AN, WSM, RH, ETH, and LZB continue to diverge, or whether higher rates and freight costs broaden the pressure across home and discretionary categories.

The market is asking a narrower question than “are geopolitics bad for stocks?” It is asking whether resilient demand can outrun the cost of energy, capital, and uncertainty. The answer remains open—and the next evidence is more likely to come from shipping flows, inflation expectations, and customer spending behavior than from a single day’s index move.