NVR — NVR, Inc.

Sentiment as of 2026-09-13 is soft but not capitulatory. The Q2 2026 print (released 2026-07-23) showed revenue of $2.33B, down ~10.5% YoY, net income of $236.5M (down ~29%), and homebuilding gross margin of 19.2% versus 21.5% a year earlier, plus $21.7M of contract land deposit impairments — a clear miss on the revenue line. Price action confirms the malaise: as of 2026-09-13 the stock sits 2.24% below its 20-day SMA, 3.22% below its 50-day, and 9.23% below its 200-day, with returns of -2.19% (1wk), -2.12% (1mo), -3.36% (3mo) and -4.73% (6mo). RSI(14) at 43.45 is neutral-to-weak rather than oversold. The 25-analyst rating distribution is genuinely mixed (11 buy, 10 hold, 4 sell), and the consensus price targets (~$7,149 consensus, ~$7,423 recent average, only 2 contributing analysts) look stale relative to the falling EPS estimates. The offset keeping sentiment from turning outright bearish: new orders rose 9% YoY to 5,885 units, backlog rose 9% to 10,998 units ($4.99B), and the cancellation rate held stable at 15%.

This is deceleration and margin compression, not a one-quarter blip. Revenue peaked at $2,851M in Q4 2024 and has since gone $2,505M (Q3'25), $2,713M (Q4'25), $1,881M (Q1'26, -20.0% YoY) and $2,326M (Q2'26, -10.6% YoY). Operating margin has stepped down in three distinct phases: ~18.9-19.7% through 2023 and 2024 (Q4'23 18.91%, Q1'24 19.18%, Q2'24 18.92%), ~16.0-16.9% through 2025 (Q1'25 16.02%, Q2'25 16.43%, Q3'25 16.88%, Q4'25 16.49%), then 12.04% in Q1 2026 and 13.25% in Q2 2026. Consolidated gross margin fell from 25.58% in Q2 2024 to 20.85% in Q2 2026 — roughly 470bp. EPS tells the same story: quarterly peak of $139.93 (Q4 2024) has fallen to $108.37 (Q2'25), $112.13 (Q3'25), $121.54 (Q4'25), $67.76 (Q1'26) and $83.62 (Q2'26). Against that, the leading indicators are stabilizating rather than deteriorating: new orders +9% YoY, backlog +9% to 10,998 units worth $4.99B, cancellations flat at 15%. Cash generation is lumpy and seasonal — TTM operating cash flow of ~$1,067M (Q3'25 $438.7M, Q4'25 $439.7M, Q1'26 $339.7M, Q2'26 -$151.5M) versus TTM net income of ~$1,141M. The honest read: a cyclical-volume and margin trough with early demand-side stabilization, but no evidence yet of a margin inflection.

Fair-to-slightly-expensive on trough earnings, cheap only if you underwrite the recovery. As of 2026-09-13, the stock trades at 15.16x the data provider's TTM EPS of $416.69 (summing the four most recently reported quarters gives $385.05, i.e. ~16.0x), versus a 5-year average P/E of 14.21x — so it is roughly 7% above its own trailing-average earnings multiple while earnings are actively falling. The more telling number: on the next-FY consensus EPS of $361.10, the forward P/E is ~17.1x, higher than trailing, because estimates assume the trough deepens further. PEG (-0.78) is meaningless here since growth is negative. P/S of 1.76x sits below the 5-year average of 2.13x, and EV/EBITDA of ~11.6x (my calc: EV ~$16.3B on ~$1.4-1.5B TTM EBITDA) is a modest premium to typical homebuilder range. P/B of 4.98x looks extreme but is an artifact of buybacks shrinking equity to $3.39B while ROIC stays at 25.8% — book value is not the right lens for this name. Verdict: you are not being paid a valuation cushion for the uncertainty; you are paying a full multiple on depressed and still-declining earnings.

The stock at ~17x forward is pricing the trough as temporary, not pricing a recession. That is the crux: if Q2 2026's 19.2% homebuilding gross margin marks the bottom and the +9% order growth / +9% backlog converts into 2027 revenue at recovered margins, today's multiple looks entirely reasonable and normalized EPS of $400+ supports a mid-$6,000s price. If instead the margin compression is structural — lot cost inflation from third-party developers plus permanent affordability-driven incentives — then 17x on a $361 EPS stream that is still sliding is too rich, and a de-rating to 15x on that estimate implies ~$5,400 (close to the 52-week low of $5,501.01). The four sell ratings against eleven buys among 25 analysts capture exactly this binary. Practical read: the market is neither underpricing a turnaround nor fully pricing in a prolonged downturn — it is waiting for the Q3 2026 print (est. late October 2026) to show whether Q2 gross margin was the floor. Until that confirmation, the stock is a 'show me' story with roughly symmetric skew.

Bearish-to-neutral trend, no timing edge either way. As of 2026-09-13, RSI(14) stood at 43.45 — neither oversold nor overbought, so no mean-reversion signal. Price was below all three key moving averages: -2.24% vs the 20-day SMA, -3.22% vs the 50-day, and -9.23% vs the 200-day — a persistent downtrend, with the wide 200-day gap indicating months of distribution rather than a recent pullback. Returns of -2.19% (1wk), -2.12% (1mo), -3.36% (3mo) and -4.73% (6mo) describe a slow grind lower, not a crash. The stock sits ~12% above its 52-week low of $5,501.01 and ~27% below its 52-week high of $8,489.85 — mid-range in a downtrend, which is the worst place to add: no capitulation discount and no momentum. Best entry setups would be either a retest of the 52-week low (where the dividend-adjusted value case strengthens) or a reclaim of the 50-day SMA on a margin-stabilization earnings print.

• Mortgage-rate path and Fed policy: persistently elevated rates are the single biggest swing factor for both buyer traffic and the incentive/price-concession burden; NVR's Q2 2026 gross margin compression is largely a rate-affordability symptom.
• Housing affordability and pricing concessions: builders are funding rate buydowns and price cuts to move inventory, which is exactly what pushed NVR's homebuilding gross margin to 19.2% in Q2 2026 from 21.5% a year prior.
• Land and developed-lot cost inflation: NVR's asset-light model buys finished lots from third-party developers, so rising lot costs feed straight into cost of sales with no raw-land hedge.
• Normalizing resale/existing-home supply: as the mortgage 'lock-in effect' fades and resale inventory returns, builders lose scarcity pricing power — a structural headwind to both volume mix and margin.
• Sector positioning: homebuilders trade as a leveraged rate/sentiment beta, so multiple expansion or contraction can dominate a single quarter's fundamental result.

• Q3 2026 earnings release, estimated late October 2026 (Oct 27-28) — the key question is whether homebuilding gross margin bottoms near Q2's 19.2%, plus cancellation trends heading into the winter season and backlog conversion pace.
• Mortgage-rate trajectory — a sustained move lower would improve affordability and reduce the incentive spend currently compressing margins; a move higher would delay any earnings trough.
• Continued share repurchases — with only ~2.70M shares outstanding, buybacks at $6,161 mechanically raise EPS per share and are a genuine per-share compounding engine (~4-5% share count reduction per year historically).
• Conversion of the +9% order growth and $4.99B backlog into revenue and margin — this is the evidence the market needs that Q2 2026 was the trough.
• Further land deposit impairments — Q2 2026 already carried $21.7M; a larger number would signal stress in the third-party developer lot market NVR depends on.

• Margin compression proving structural rather than cyclical: operating margin has fallen from 18.92% (Q2 2024) to 13.25% (Q2 2026), and lot costs plus affordability incentives do not resolve quickly.
• Earnings still falling — four consecutive quarters of YoY declines, and the next-FY consensus EPS of $361.10 sits below TTM EPS, meaning the trough may not yet be in.
• Balance-sheet cushion is thinning: cash fell from $2.61B (12/31/24) to $1.20B (6/30/26) and stockholders' equity from $4.21B to $3.39B, largely through buybacks that exceeded earnings — less dry powder if the downturn extends.
• Earnings-quality watch: Q2 2026 showed $236.5M net income against -$151.5M operating cash flow, an accruals ratio of 7.09%, and it has worsened each year (Q2 2024: 2.63% → Q2 2025: 5.03% → Q2 2026: 7.09%). Mostly seasonal, but worth monitoring.
• Valuation risk: ~17x forward earnings for a declining EPS stream offers no cushion if the 2027 recovery does not materialize; a de-rating to 15x on $361 EPS implies roughly $5,400.
• Affordability shock: another leg up in mortgage rates would hit volumes and price simultaneously, and NVR's pre-sold model protects against spec inventory but not against demand disappearance.

NVR is a genuinely high-quality business — 25.8% ROIC, 34.7% ROE, $1.2B of cash against $908M of long-term debt, an asset-light lot-option model that avoids raw-land risk, and an aggressive buyback shrinking a ~2.70M share count ~4-5% a year. But as of 2026-09-13 you are paying ~17x forward earnings for a company whose EPS is still falling, with the market already having priced in the recovery. The demand-side inflection is real (orders +9%, backlog +9% to $4.99B, cancellations stable at 15%) and is the reason not to be bearish, but it has not touched the income statement yet — operating margin went from 18.92% in Q2 2024 to 13.25% in Q2 2026 with no turn visible. This is a wait-for-confirmation name, not a buy at $6,161 mid-downtrend: own it on a margin-stabilization print or a pullback toward the $5,501 52-week low, where normalized earnings power of $400+ per share starts to look genuinely cheap. Not a sell, not a buy — a fair-priced quality cyclical in a trough.

Normalized mid-cycle EPS of ~$410 — set between FY2025 actual EPS of $436.87 and the next-FY consensus of $361.10, reflecting a trough where orders and backlog are up 9% but margins have not recovered — multiplied by 15.5x P/E, a modest premium to NVR's own 5-year average of 14.21x justified by 25.8% ROIC, the asset-light lot-option model, and roughly 4-5% annual share-count reduction, but below peak-cycle multiples given residential-construction cyclicality. Cross-checks bracket the result: 10.5x EV/EBITDA on normalized EBITDA of ~$1.5B implies roughly $5,950/share and 15x on $410 implies ~$6,150, so $6,355 sits at the upper end of a $5,950-$6,400 range — consistent with a stock that is roughly fairly valued with a slight edge, well below the stale sell-side consensus of ~$7,149.