Tools chosen: kalshi_direct kalshi_related earnings_data claude_news gdelt_news code_execution polymarket_related

Sub-questions (Fermi decomposition)

Planner reasoning

Carvana's retail unit sales have grown from ~313k (2023) to ~417k (2024) to roughly ~580-600k expected in 2025; reaching >900k in 2026 would require ~50%+ YoY growth, so the key empirical work is nailing down the 2025 actual base and 2026 guidance/analyst consensus. The Kalshi market price for this and neighboring strike levels (the KXCVNAA series is a threshold ladder) is the primary anchor and lets me back out an implied distribution.

# Current state Carvana resolves this Kalshi market based on FY2026 total retail units sold (reported via 10-K/8-K). As of the latest confirmed data (Q2 2026 earnings, reported 2026-07-29), H1 2026 actual retail units = 384,718 (Q1: 187,393 +40% YoY; Q2: 197,325 +38% YoY), following FY2025 actual of 596,641 units (+43% YoY). YoY growth is decelerating each quarter; management's FY2026 guidance disappointed investors enough to trigger a stock selloff despite record results. # Timeline of key events - 2023: FY2023 retail units = 312,847 (confirmed, 10-K). - 2024: FY2024 retail units = 416,348, +33.1% YoY (confirmed, 10-K). - 2025-Q1 to Q4: Quarterly units 133,898 / 143,280 / 155,941 / 163,522; FY2025 total = 596,641, +43% YoY (confirmed, 8-Ks/shareholder letters). - 2026-01 (implicit): Management guides FY2026 "significant growth" in units and Adj. EBITDA, sequential Q1 increase (confirmed, Q4 2025 letter). - 2026-Q1: 187,393 units, +40% YoY (confirmed, 8-K). - 2026-05-25: Stock split; analysts update price targets (reported). - 2026-06-23/25: June sales growth "appears to slow"; Jefferies stays bullish (reported). - 2026-07-29: Q2 2026 results — 197,325 units, +38% YoY, record revenue $7.376B; guides Q3 sequential increase and FY2026 Adj. EBITDA $2.7-3.0B (confirmed, 8-K/earnings call). - 2026-07-30: Stock tumbles as full-year guidance "disappoints" despite record quarter (confirmed via multiple outlets — Yahoo, ZeroHedge, Benzinga). - 2026-08-14/19: Bulls/bears split on valuation and business model post-guidance (reported/opinion). # Event Will Carvana report above 900,000 total FY2026 retail vehicle units sold? # Outcomes to forecast Yes / No (single threshold: >900,000) # Kalshi market anchor **KXCVNAA-28JANUNITS-900000 "Above 900 thousand": current YES price = 3%**. 7-day change +2pp, 30-day change +2pp (off a very low base), avg daily volume ~1,007 contracts. Thin but rising slightly — market treats this as a low-probability tail outcome. # Sub-question answers 1. **FY2023/2024/2025 quarterly units** — 2023: 312,847; 2024: 416,348 (+33.1%); 2025 quarters: Q1 133,898, Q2 143,280, Q3 155,941, Q4 163,522, FY2025 total 596,641 (+43%). [claude_news/SEC filings] 2. **Implied 2025 total & required 2026 growth** — Actual FY2025 = 596,641 (not estimated). To exceed 900k in 2026 requires ~50.9% YoY growth. With H1 2026 actual already at 384,718 (+39% avg), H2 2026 would need to grow ~61% YoY over H2 2025 (319,463) to cross 900k — far above the current ~38-40% trend. [code_execution model + claude_news actuals] 3. **Management guidance/consensus** — Management guided "significant growth" in units/EBITDA and sequential quarterly gains, but the FY2026 guidance issued with Q2 results (Adj. EBITDA $2.7-3.0B) disappointed the Street enough to crash the stock despite record Q2 — implying units guidance is below bullish expectations, consistent with continued deceleration rather than acceleration. [gdelt_news, claude_news] 4. **Other ladder strikes** — Only the 900k strike was returned by the related-markets scan; no other KXCVNAA strikes surfaced, so the market's implied full distribution can't be directly inferred. [kalshi_related — gap] 5. **Capacity constraints** — Carvana has real estate for 3M units/year long-term and integrated 16 ADESA sites by end-2025 (~1.5M annual capacity built), so physical capacity is not the binding constraint for 2026; growth deceleration appears demand/guidance-driven, not capacity-driven. [claude_news] 6. **Historical 50%+ growth base rate** — Recent data shows -3% (2022), -24% (2023), +33% (2024), +43% (2025); no year in the available data shows 50%+ growth. Earlier 2017-2021 hypergrowth figures were not returned by research (gap), but visible modern-era pattern shows deceleration after 2025's +43%, not further acceleration. [claude_news; gap on 2017-2021] # Key facts (high-confidence, factual) 1. [SEC 8-K/claude_news] FY2025 retail units = 596,641 (+43% YoY, confirmed). 2. [SEC 8-K] Q1 2026 = 187,393 (+40% YoY); Q2 2026 = 197,325 (+38% YoY) — both confirmed actuals. 3. [gdelt_news] Q2 2026 earnings beat but FY2026 guidance disappointed; stock fell. 4. [claude_news] Carvana ~1.5% of US auto market at ~800k run-rate; long-term target 3M units by 2030-2035. # Cross-market signals - Kalshi related: Ford 2026 US sales "Above 2.15M" market trading 31% (much higher confidence, different asset class) — not directly comparable but shows market willing to price >30% on adjacent auto-volume bets when plausible; CVNA 900k bucket at just 3% signals strong skepticism. - Polymarket: No matching markets found (0/100 scanned). - Sportsbook implied: N/A. # Analyst opinions and speculation - Jefferies remains bullish despite June sales deceleration signal (reported, 2026-06-23/25). - Post-Q2 2026, "bulls and bears split" on business model/valuation (Yahoo, 2026-08-19); some analysts frame CVNA as bargain at 13x EBITDA (InsiderMonkey, 2026-08-14). - Media framing: "guidance disappoints despite record quarter" — suggests Street/consensus units expectations for FY2026 sit meaningfully below bull case, aligning with sub-900k outcomes. # Directional lean per outcome - **Yes (>900k)**: Requires H2 2026 growth to re-accelerate to ~61% YoY vs. current ~38-40% trend and prior guidance disappointment — no supporting evidence found; capacity is sufficient but demand/guidance signals argue against. Weak case. - **No (≤900k)**: H1 2026 actuals (384,718) plus decelerating trend imply FY2026 total ~800k-850k at current growth rates; guidance letdown in July 2026 reinforces this; Kalshi price (3%) and quantitative model (3-15% base case) both support No. Strong case. # Gaps / unknowns - No visibility into other KXCVNAA ladder strikes (e.g., 800k, 850k) to triangulate implied market distribution. - No explicit 2017-2021 CVNA growth-rate data to fully test the "50%+ growth base rate" question. - H2 2026 (Q3/Q4) actuals not yet available — outcome still depends on realized H2 growth acceleration/deceleration. - Exact sell-side consensus unit estimate for FY2026 not directly quoted in sources. # Calibration anchors - Kalshi current YES price: 3% (anchor). - Quantitative model (code_execution, updated with actual 2025/H1-2026 data) implies ~3-15% probability under realistic growth-deceleration scenarios, tightening toward Kalshi's price once actual H1 2026 figures are used instead of estimated bases. - Precedent: Carvana has never posted two consecutive years of ~40-50%+ growth historically (2022-2023 were negative); current deceleration pattern (46%→38% quarter-over-quarter YoY) argues against the acceleration needed for Yes.

1. Both forecasts converge tightly around the 3% Kalshi anchor (3% and 4%) without independently stress-testing tail scenarios like M&A, a surprise demand surge from a competitor's exit, or accounting/reporting changes (e.g., wholesale-to-retail reclassification) that could inflate the unit count beyond organic growth trends—these are not discussed at all. 2. Neither forecaster addresses the "Other ladder strikes" gap explicitly in their probability construction; since the brief notes no 800k/850k strikes were found, both should acknowledge more explicitly that their point estimate (~800-830k) sits close enough to 900k that model uncertainty (e.g., ±10% variance in H2 execution) could matter more than the tight 3-4% they land on. 3. Both rely on linear/sequential extrapolation from Q1/Q2 to project H2, but neither considers that Carvana's growth has historically been lumpy with occasional acceleration (e.g., 2025's +43% following 2024's +33%)—a reacceleration in H2 (holiday inventory builds, ADESA capacity ramp-up effects) is dismissed too quickly given "capacity is not the binding constraint" per the brief. 4. Neither critiques the Kalshi price itself for potential mispricing due to thinness (~1,007 avg daily volume) — a 3% price on a thin market could be stale or reflect only a few large traders, yet both forecasters treat it as a strong independent confirmation rather than a weakly-informative signal that happens to agree with their own models. 5. Forecast 2's slight upward adjustment to 4% "to allow for execution upside" is arbitrary and not tied to any specific evidence (e.g., Jefferies' bullish stance despite deceleration signals, mentioned in the brief but not incorporated into either rationale's numerical reasoning). 6. Neither forecast discusses resolution-date risk: the market closes 2028-03-31, well after FY2026 results would normally be confirmed (10-K), so there's no meaningful ambiguity about reporting timing—but neither explicitly rules out restatement/revision risk that could shift the reported figure post-initial 8-K release.

2% (82% confidence)

Yes: 2% No: 98%

With H1 2026 actuals confirmed at 384,718 units, crossing 900k requires H2 of 515,282 — roughly +61% YoY vs H2 2025's 319,463, versus an observed and decelerating trend of +40%/+38%. Sequential extrapolation (Q3 ~205-215k, Q4 ~210-225k) lands FY2026 near 800-830k, a gap of ~10% that would require a sharp reacceleration rather than the continued deceleration that management's disappointing post-Q2 guidance signals. The critique's tail scenarios (M&A, reclassification, competitor exit) are real but each would need to add ~80-100k units in a single half-year, which is an extreme low-probability event; even a generous allowance for such tails plus thin-market noise keeps the YES probability in the low single digits. I settle marginally below the 3% Kalshi anchor because the confirmed H1 data is more informative than the thin market price and points to a shortfall that is large relative to plausible execution variance.

Total pipeline time: 222.1s