The Oil Shock Test: Can AI Demand Outrun Geopolitical Supply Risk?

A fresh geopolitical shock is moving through markets less as a single “risk-off” trade than as a chain of costs: crude, freight, delivery times, household budgets and ultimately corporate margins.

A drone attack has closed Saudi Arabia’s East-West pipeline, a route designed to bypass the Strait of Hormuz. The Los Angeles Times reported that the closure could remove as much as 4% of global oil supply, while renewed Houthi pressure threatens traffic through the Bab el-Mandeb chokepoint. Oil moved above $107 a barrel on Tuesday, and the forward market was paying a premium for early delivery—signals that traders were pricing scarcity rather than simply reacting to headlines.^[1]

The market question for the specified scope—DDOG, SNOW, RH, WSM, ETH, LZB, LESL and TPX—is whether earnings growth and resilient demand can absorb that kind of external pressure over the next year.

The actual market tell: resilience is selective

The latest quote snapshot shows a split. In pre-market trading on September 16, 2026, SNOW was $321, down 0.61% from its 16:00 ET close; DDOG was $229.32, down 0.41%; and LZB was $30.18, down 0.46%. RH and WSM were firmer pre-market after sharp prior-session declines: RH was $125.7858, up 0.77% versus its close, while WSM was $224.79, up 1.50%. LESL was $0.4663, down 0.15%. ETH was $23.05 at the September 15 close, down 4.87%; TPX’s available quote was stale, dated February 26, 2025, so it is not used as a current signal.^[2]

That dispersion matters. The data do not show a clean collapse in software demand or a broad consumer recovery. They show investors distinguishing between businesses with a direct usage or pricing engine and businesses exposed to physical goods, shipping, fuel or discretionary budgets.

SNOW supplies the strongest evidence for the resilience thesis

Snowflake’s latest available earnings-call evidence is unusually direct. Management said product revenue growth reached 37% year over year and raised fiscal 2027 product-revenue guidance to 36% growth, attributing the acceleration to an AI flywheel: AI products increase platform consumption, and platform usage supports further AI adoption.^[3]

That is the bullish case for the scope. Enterprise AI workloads can create demand that is less sensitive to a temporary fuel shock than furniture deliveries or pool maintenance. But it is not a guarantee. The same transcript record shows analysts asking about “sticker shock” as agents proliferate and workloads expand.^[3] Consumption-led models still have to prove that new workloads become durable production spending rather than experimentation later optimized away.

RH and WSM show where geopolitics reaches the income statement

Physical retail provides the cleaner transmission channel. RH reported Q2 2026 revenue of $922.2 million, up 2.6%, but its filing also showed adjusted EBITDA margin of 19.4% inclusive of $55.1 million, or 600 basis points, of tariff benefit.^[4] A benefit tied to tariff refunds can support a quarter without proving that the underlying supply chain has become less exposed.

RH management has separately described a difficult housing backdrop and said tariffs had caused product delays, out-of-stocks and repeated price negotiations.^[5] The geopolitical risk is therefore two-sided: higher freight and energy costs can squeeze margins, while price increases can test demand and delay purchases.

Williams-Sonoma offers a more constructive but still conditional example. Its Q2 call said comparable revenue rose 6.2% and full-year outlook was raised. Yet gross margin fell approximately 160 basis points year over year, with merchandise margin down about 230 basis points as tariffs lifted weighted-average cost of goods sold. Management called Q2 the peak tariff impact and expected pressure to moderate.^[6] That is evidence of operating resilience, not immunity.

What this means for the rest of the scope

DDOG and SNOW are the clearest software tests: the thesis needs enterprise budgets and usage to remain durable even if higher energy prices slow the economy. The available SNOW evidence supports that case more strongly than it proves it for DDOG.

RH and WSM are the best observable tests of physical-goods exposure. Their recent disclosures show that demand can hold up while margins absorb tariffs, but the result depends on pricing, sourcing, freight and the consumer’s willingness to tolerate higher tickets.

LZB, LESL and TPX belong in the same risk framework, but this research pass does not establish a current company-specific catalyst for each. Their inclusion should not be mistaken for evidence that every name has the same earnings trajectory. ETH is also a distinct exposure: its latest available quote was materially lower, but the supplied market data do not identify whether that move was driven by this geopolitical shock, crypto-specific positioning or another catalyst.^[2]

Base case and failure case

The balanced base case is that the shock remains concentrated in energy and logistics. Software demand, especially AI-linked consumption, continues to provide an earnings-growth counterweight, while retailers mitigate part of the cost through sourcing changes, pricing and operating leverage.

The failure case is a second-order loop: prolonged disruption keeps crude and diesel elevated, freight and insurance rise, central banks face renewed inflation pressure, and discretionary consumers defer large-ticket purchases. In that scenario, the thesis fails first through margins and affordability—not necessarily through an immediate collapse in software usage.

There is an important uncertainty in the evidence. The reported pipeline capacity at risk is large, but the duration of the outage and the ability to reroute barrels are not yet established in the sources reviewed. The market’s early-delivery premium is a warning signal, not proof of a lasting supply deficit.^[1]

What to watch next

1. Repair and rerouting: whether Saudi Arabia restores East-West pipeline operations and whether alternative export routes remain available.
2. Shipping insurance and traffic: whether commercial traffic through Hormuz and Bab el-Mandeb normalizes or disruptions broaden.
3. Energy pass-through: diesel, freight and delivery-cost commentary in the next RH and WSM updates.
4. Software consumption quality: whether SNOW and DDOG describe AI usage as production workloads with durable consumption, rather than pilots or temporary bursts.
5. Margin versus demand: whether consumer names preserve full-price selling as input costs rise, or rely increasingly on promotions.
6. Rates and inflation expectations: whether the energy shock changes the policy path enough to pressure long-duration software valuations.

The current evidence supports a qualified version of the resilience thesis: earnings growth can offset geopolitical risk when demand is usage-driven or when a company has pricing and supply-chain tools. It does not support treating the entire scope as equally insulated. The market is already drawing that distinction.