Energy Shock Tests AI Demand Resilience

The market tell: demand is holding, but the discount rate is moving

The current test for the software-and-demand thesis is not simply whether enterprises still want AI. It is whether resilient usage can offset a higher cost of energy, capital and risk.

On October 16, the latest available quote snapshot showed a sharp split inside the monitored scope: DDOG was down 6.9%, SNOW 7.6% and CRWV 10.6% from the prior close, while AN was up 1.7% and WSM was little changed. RH, ETH and LZB were modestly lower. The prices are delayed 15 minutes and timestamped 12:27 ET, so they are a market read—not a closing verdict.^[1]

That cross-sectional move matters. Pressure was concentrated in long-duration, AI-linked software and infrastructure names rather than spread evenly across every consumer-facing company in the scope. It is consistent with investors demanding more compensation for financing risk and future cash-flow uncertainty, even while the underlying demand question remains open.

Geopolitics is arriving through energy and freight

CNBC reported that attacks around the waterway had surged, that a Singapore-registered vessel sank after being struck in the Strait of Hormuz, and that oil markets were weighing renewed regional attacks and possible U.S. action involving the so-called shadow fleet.^[2]

The supply response is cushioning—but not removing—the shock. The International Energy Agency agreed to release more than 20 million barrels of diesel, in addition to oil remaining from an earlier emergency-reserve release. Yet the same report described Brent near $105.51 and WTI near $97.43 in the cited trading snapshot, with freight and geopolitical uncertainty still central to pricing.^[2]

For this watchlist, the transmission mechanism is more important than oil alone: higher energy and freight costs can keep inflation expectations elevated; higher inflation risk can keep long-term yields high; expensive capital raises the hurdle for leveraged data-center expansion; and weak consumer confidence can pressure discretionary purchases and big-ticket home goods.

The macro backdrop is mixed, not recessionary

The latest macro snapshot showed 4.2% unemployment, 3.72% year-over-year CPI inflation, a 3.75% federal-funds rate, a 5.19% 10-year Treasury yield, and a 15.27 VIX. Real GDP growth was 2.1%, while consumer sentiment was 51.7 and down 11.17% year over year.^[3]

That combination explains why the market can sell growth multiples without yet pricing a broad collapse in demand. Growth remains positive and labor conditions are not recessionary in the snapshot, but the long end of the curve and weak sentiment make the market less forgiving of high expectations.

Evidence for demand resilience

Datadog’s latest available earnings-call discussion supplied the clearest primary evidence in this scope. CEO Olivier Pomel said Q2 revenue grew 36% year over year to $1.12 billion, growth accelerated across the customer base, and non-AI customers were growing in the high 20s. He also said customers were adopting AI alongside cloud and modern technologies, increasing usage of Datadog to observe, secure and act on those workloads.^[4]

The same call described roughly 33,400 customers and about 4,720 customers with at least $100,000 of annual recurring revenue. Those figures support the resilient-demand side, although one company’s commentary cannot validate the entire group.^[4]

Snowflake is another read-through: a September Reuters report described an AI-powered results update that lifted the company’s outlook and buoyed software stocks. The accessible result supports the direction of the report, but not enough detail to responsibly reproduce the full financial table here.^[5]

Evidence against a simple AI-demand-wins narrative

CoreWeave is the direct financing-risk signal in the group. Current reporting cited investor anxiety around its debt load even as other coverage pointed to large customer backlogs and rapid revenue growth.^[5] That combination is why demand alone is not enough: contracted growth must become capacity, utilization and cash flow without an unacceptable financing burden.

The price response reinforces that distinction. CRWV’s 10.6% decline was larger than the moves in DDOG and SNOW in the October 16 snapshot.^[1] The move cannot prove fundamental deterioration; it does show that investors are differentiating between AI demand and the capital structure used to serve it.

The consumer names provide a counterweight. AN and WSM were positive or nearly flat, while RH and LZB were modestly lower.^[1] This is too short a window to call a durable rotation, but it argues against describing the session as a uniform liquidation of every growth or discretionary exposure.

What this means for the monitored scope

The thesis remains conditional. Datadog supports demand resilience; the 5.19% 10-year yield and weak sentiment argue that valuation sensitivity remains high; and shipping attacks, emergency fuel releases and shadow-fleet policy discussions show that geopolitics is an economic channel.^[4][3][2]

DDOG, SNOW and CRWV are the cleanest test of whether AI usage can outrun the cost of capital. AN, RH, WSM and LZB are more useful for observing whether household confidence and financing conditions are broadening into discretionary demand. ETH adds separate digital-asset sensitivity, but this pass did not identify a defensible catalyst for its modest move; it should not be forced into the same causal story.

What to watch next

1. Shipping and sanctions: whether attacks around the Strait of Hormuz ease or intensify, and whether shadow-fleet sanctions are tightened or relaxed.^[2]
2. The long end of the Treasury curve: whether the 10-year yield responds to improving supply conditions or remains elevated because inflation risk persists.
3. AI utilization versus AI capex: whether future commentary distinguishes customer usage and revenue from the cost and timing of capacity. CRWV is the clearest watchpoint because backlog and leverage must move together.
4. Breadth beyond software: whether AN and WSM’s relative firmness persists and whether RH and LZB stabilize. A broader consumer rebound would support demand; renewed weakness alongside high yields would point to affordability pressure.
5. Management language on budgets: whether non-AI customer growth remains broad and whether AI workloads translate into durable expansion rather than a concentrated cohort effect.

The disciplined conclusion is neither “geopolitics breaks the AI cycle” nor “AI demand makes macro irrelevant.” The current evidence supports a narrower claim: demand is still showing resilience, but the market is charging more for the capital, energy and execution risk required to monetize it. That is why the next fundamental disclosures matter more than a single day’s drawdown.