Earnings growth splits the deck: software accelerates, furniture stalls, Leslie's breaks

The idea behind this earnings deck was simple: that earnings growth and resilient demand could carry eight names — the observability and data-cloud pair DDOG and SNOW, and the home-furnishings group RH, WSM, ETH, LZB, LESL and TPX — through the next year. One full earnings cycle later, the evidence has split into three very different stories. The software names are delivering on the thesis. The furniture names are mostly holding margins while waiting for a housing-market turn that has not arrived. And the most distressed retailer in the group has broken it outright.

The cloud row is the part of the thesis that is working

The strongest case comes from the software names. Snowflake reported its fiscal Q2 on September 2 with product revenue of $1.49 billion, up 37% year over year, and total revenue of $1.55 billion, up 35%, with net revenue retention of 126% and 828 customers generating over $1 million in trailing-twelve-month product revenue, up 27%. The stock spiked about 22% as management pointed to momentum in its AI coding agent, CoCo. ^[1]

Datadog’s fiscal Q2, reported August 6, showed revenue up 36% to $1.12 billion, with roughly 4,720 customers at or above $100,000 in ARR versus about 3,850 a year earlier. Management framed FY2026 revenue growth around 30%, with an explicit note of conservatism tied to a usage reduction from its largest customer. ^[2] On the call, CEO Olivier Pomel described AI as a tailwind: more than 750 AI-native customers across the base, all 10 of the largest AI leaders as customers, and agentic MCP tool calls quadrupling quarter over quarter and growing more than 22x versus Q4 2025. ^[3]

What would have to be true for this side of the thesis to keep working is a continuation of what the quarter already showed: demand broadening beyond a handful of mega-cap AI spenders. Datadog reported acceleration in non-AI customers too — the strongest sequential usage growth in twelve quarters — which is precisely the broadening that makes the growth look durable rather than concentrated. ^[3] As of midday Friday, DDOG traded at about $273 and SNOW near $341. ^[4]

The furniture row: margins defended, demand not recovered

The home-furnishings side is a different, more fragile story. Williams-Sonoma is the standout: fiscal Q2 comparable revenue rose 6.2% with a 22.9% GAAP operating margin, and the company raised its full-year outlook — enough that its shares have risen even with housing activity weak. ^{[5] [6]} RH reported fiscal Q2 net revenue up 2.6% to $922.2 million, above the high end of its guidance, with an adjusted EBITDA margin of 19.4% that included $55.1 million, or 600 basis points, of tariff benefit — meaning the normalized margin sits meaningfully lower. Hitting growth, in other words, has partly been a function of tariff windfall, not underlying unit demand. ^{[7] [8]}

The rest of the group reads as stabilization under pressure. La-Z-Boy’s fiscal Q1 showed retail written sales up 16% and written same-store sales up 3%, but the stock drew scrutiny after sharply reduced guidance. ^[7] Ethan Allen — the scope’s “ETH,” listed on NYSE as ETD — reported fiscal Q4 net sales down 8.5% on lower contract and delivery demand, cushioned by strong margins and a special-plus-regular dividend. ^[7] The bedding name formerly known as TPX, now Somnigroup International (SGI) after its Tempur Sealy–Mattress Firm combination, saw Q2 net sales fall 3.0% to $1.82 billion even as EPS rose 11%, the margin gain coming against record operating cash flow and a softened industry backdrop. ^{[5] [6]}

Across the broader furniture complex, the language is strikingly consistent. One peer called the home-furnishings industry “abysmal from an actual unit-sold perspective,” with the housing market working through high levels of unsold homes and acute affordability pressure; others described demand as remaining below historical levels until housing turnover normalizes. ^{[9] [7]} That is the crux: for RH, WSM, LZB and the rest to grow earnings durably, what has to be true is a housing-and-affordability recovery — and the quarters so far do not provide that evidence. What they do show is operators defending margins through mix, pricing and cost discipline, which can support earnings even without a top-line boom.

The one name that breaks the thesis

Leslie’s, the pool-care retailer, has moved from merely weak to distressed. Its stock trades near $0.34, and on Friday — September 25 — reports emerged that the company is preparing to file for Chapter 11 bankruptcy protection as soon as next week, with Nasdaq delisting pressure looming after a period of repeated declines. ^{[10] [4]} For the “resilient demand” hypothesis, this is the decisive negative case: a consumer-discretionary name where demand and the balance sheet broke together, and where no amount of industry-wide stabilization is likely to arrive in time. It is the counterexample that shows the thesis was never going to apply uniformly to every name in scope.

What to watch next

• The next software prints (Nov–Dec): Datadog reports November 5 before the open and Snowflake December 2 after the close. The question is whether non-AI broadening and AI-native expansion hold up, and whether the largest-customer usage reduction Datadog flagged actually lands as guided. ^[11]
• Housing-led names: Williams-Sonoma reports November 18, La-Z-Boy November 17, RH December 10, and Leslie’s December 1. Watch whether comp trends can hold without tariff windfalls — RH’s 600-basis-point tariff benefit is the signal to track most closely. ^[11]
• The housing macro: Every furniture operator has tied its turn to affordability and housing turnover. Short of falling mortgage rates and rising existing-home sales, expect the “margins defended, units not recovered” pattern to persist.
• Leslie’s and the restructured bedding complex: Chapter 11, if filed, will reset the base case for LESL entirely. Separately, watch how Somnigroup’s Mattress Firm integration economics evolve into 2027.

The takeaway from this pass is not that the thesis is simply right or wrong — it is that it holds only selectively. It is compelling for the cloud-software names, partial for the furniture operators that can defend margins against a stalled housing market, and bankrupt for the one name that could not hold on long enough. Treating the group as a single trade would be the mistake; the evidence now argues these are three separate bets.