The Market Is Stress-Testing AI Demand, Not Rejecting It

The opening tape is delivering a useful distinction: investors are not treating all growth as one trade. At 12:07 p.m. ET on October 22, SPY was down 1.06%, QQQ was down 2.18%, and DIA was nearly flat. Yet IGV, the software ETF, was up 0.61%—a small but important divergence on a day when the Nasdaq was under pressure.^[1]

That split is the first test of the hypothesis that earnings growth and resilient demand can support DDOG, SNOW, CRWV, AN, RH, WSM, ETH, and LZB over the next year. The evidence is strongest where AI increases the use of an existing software platform. It is less settled where the story depends on enormous infrastructure spending, financial conditions, or discretionary household demand.

The tape says “selective,” not “broken”

Over the past 180 days, IGV returned 28.12% versus 6.97% for SPY. DDOG gained 99.68% and SNOW 117.89%, though both carried substantially higher volatility and experienced drawdowns along the way. By contrast, XLY declined 1.94% over the same period.^[2]

That historical performance is not proof of future earnings. It does, however, establish the market’s starting point: software has already received a powerful re-rating, while consumer discretionary has not. Today’s divergence therefore matters less as a verdict on the next year than as a reminder that expectations and operating results must now keep moving together.

Quotes are delayed 15 minutes and reflect the regular session as of 12:07 p.m. ET.^[1]

Software has the cleanest operating evidence

Datadog’s latest available call offers the clearest support for the demand side of the thesis. Management said second-quarter revenue growth accelerated to 36% year over year, with non-AI customer growth in the high 20s, and described demand as broad across startups and large enterprises. It also pointed to roughly 33,400 customers and about 4,720 customers with at least $100,000 in annual recurring revenue.^[3]

The important detail is that the argument does not rely exclusively on AI-native customers. Datadog described both AI and non-AI cohorts accelerating, with cloud migration and greater use of AI helping drive platform usage. That makes DDOG a useful test of whether AI is expanding observability and security workloads rather than merely shifting investor enthusiasm from one application category to another.

Snowflake’s commentary describes a similar but distinct mechanism. Management said AI was bringing new workloads onto the platform, increasing adoption of its first-party products, and lifting overall consumption. In the latest available call, Snowflake reported second-quarter non-GAAP operating margin of 15%, more than 400 basis points higher year over year, while arguing that AI adoption was creating a multiplier for core platform consumption.^[4]

The counterpoint is embedded in Snowflake’s own investor discussion: as agents and new workloads proliferate, customers may face “sticker shock” or enter periods of optimization.^[4] Consumption businesses can grow quickly, but the path is not always linear. The next proof point is not simply another AI product announcement; it is whether usage, retention, and margins hold together as customers learn what these workloads actually cost.

Infrastructure is the harder version of the thesis

CRWV fell 8.31% in the current session, the sharpest decline among the listed companies in this scope.^[1] Recent market coverage also identified CoreWeave among the AI-linked names pressured as oil rose and investors revisited AI-related risk.^[5]

A down day does not establish that the business is deteriorating. It does show why infrastructure deserves a different analytical standard from software. For a platform company, stronger AI usage can increase recurring consumption. For a compute provider, demand must also clear questions around capital intensity, capacity, customer concentration, funding, and the economic life of the equipment. The thesis can be right about AI demand and still be wrong about how much of that demand becomes attractive equity economics.

AN was up 3.88% today, a reminder that not every technology-adjacent name is moving with the AI infrastructure basket.^[1] Its upcoming report is scheduled for October 30, with the date marked estimated and the session listed as before the open. DDOG is scheduled for November 4, also estimated and before the open; SNOW is scheduled for December 2, estimated and after the close.^[6] Those dates are checkpoints, not forecasts, and the calendar tool does not provide a confirmed reporting date for CRWV.

Consumer demand is a different test

The consumer portion of the scope is mixed. RH rose 7.57%, WSM 1.84%, and LZB 4.81% today, while XLY was down 0.14% in the current session and remains down 1.94% over the six-month comparison window.^[1][2]

That combination cautions against treating one-day strength as broad demand confirmation. Furniture and home-related purchases are more discretionary and often more rate-sensitive than enterprise software subscriptions. They can benefit from company-specific merchandising, easier comparisons, or a recovery in housing activity without proving that the household is broadly unconstrained.

The macro backdrop reinforces that caution. The latest available snapshot shows 4.2% unemployment, 2.1% real GDP growth, and a 3.72% year-over-year CPI reading. But consumer sentiment was 51.7, down 11.17% year over year, while the 10-year Treasury yield was 5.15% and the VIX was 18.44.^[7] This is not a recession signal in the snapshot, but it is a demanding environment for expensive, rate-sensitive purchases and richly valued growth narratives alike.

ETH adds another high-sensitivity observation to the scope: its quoted instrument was down 3.93% today.^[1] That move is not evidence against enterprise demand, but it is evidence that liquidity-sensitive exposures are not receiving the same support as the best software operators.

What the hypothesis gets right—and where it overreaches

Evidence in favor:

• DDOG’s management described accelerating growth across AI and non-AI customers, with low churn and expanding enterprise reach.^[3]
• SNOW’s management linked AI adoption to new workloads, core-platform consumption, and margin expansion.^[4]
• Software has outperformed the broad market over six months, even after meaningful volatility.^[2]

Evidence against a broad version of the thesis:

• CRWV’s sharp one-day decline shows that AI enthusiasm does not automatically translate into stable infrastructure equity performance.^[1]
• XLY’s six-month decline and weak sentiment complicate the case for a synchronized consumer recovery.^[2][7]
• The macro mix—positive growth but a high long-term yield, weak sentiment, and higher volatility—can reward durable execution while compressing weaker or more speculative stories.^[7]

The balanced reading is that the hypothesis is most credible as a company-selection question, not as a blanket statement about every AI, consumer, or crypto-linked asset in the list. DDOG and SNOW currently offer the strongest operating evidence in this research pass. CRWV requires proof that infrastructure demand produces durable returns on capital. RH, WSM, and LZB require evidence that demand is broad and repeatable. ETH remains a liquidity-sensitive cross-check rather than a direct read on enterprise software demand.

What to watch next

1. DDOG and SNOW earnings: Look for continued non-AI customer growth, consumption durability, retention, AI workload contribution, and whether margin expansion survives additional investment. DDOG’s next scheduled date is November 4, estimated, before the open; SNOW’s is December 2, estimated, after the close.^[6]
2. Infrastructure economics: For CRWV and related AI providers, watch customer concentration, capacity utilization, financing needs, and the relationship between revenue growth and capital intensity—not just headline demand.
3. Consumer breadth: Compare RH, WSM, and LZB commentary with the broader XLY signal. A durable thesis needs more than isolated strength in a few names.
4. Rates and sentiment: A 5.15% 10-year yield and 51.7 consumer-sentiment reading create a useful stress test for long-duration growth and discretionary demand.^[7]
5. Market confirmation: The key question is whether software’s operating evidence continues to separate it from capital-intensive AI exposure, rather than assuming that every part of the theme must move together.

The current tape is therefore best described as a stress test. It is challenging the market to identify where AI creates measurable, recurring demand—and where the story still depends more heavily on financing, valuation, or confidence. That is a healthier and more discriminating test than either declaring the theme finished or assuming that strong historical momentum settles the question.