Lets do some $OPEN math. Someone asked what Opendoors net profit looks like at $20B revenue with a 10% contribution margin. What $20B means in homes. → $20B ÷ $375K average home = 53,350 homes sold in the year → ÷ 4 = 13,340 homes a quarter → ÷ 52 = 1,026 homes a week Contribution profit. → $20B × 10% = $2.0B a year, or $500M a quarter Now operating costs. Q1 2026 ran $63M a quarter on $720M of revenue, split across three lines. Scaling each to roughly seven times the revenue: → sales and marketing $33M → $60M, it grows with volume but the agent channel is far cheaper than paid ads → general and admin $22M → $35M, mostly fixed, plus the Doma team coming in → tech and development $8M → $15M, AI native, barely scales with volume → $110M a quarter, so $440M a year That's 2.2% of revenue against roughly 8.75% today. That gap is the operating leverage, and it's the whole reason scale changes everything here. → $2.0B contribution − $440M opex = $1.56B adjusted EBITDA Then the costs below EBITDA. Interest first. Opendoor borrows against nearly every home it owns, so the bill depends on how long each home sits before selling. At a 90 day hold you're carrying roughly a quarter of the year's homes at any moment. → 53,350 homes × (90 ÷ 365) = 13,155 homes held at any time → × $375K = $4.93B of inventory → 90% of that financed = $4.44B of debt → × 7% = $311M a year On that rate, Q1 2026 showed $23M of interest on $1.14B of debt, which annualises to about 8%. But that figure also includes their convertible notes, so 7% on the home loans alone is the fair number. And this is why Kaz's cohort chart matters, the one showing each new batch of homes selling faster than the last. → 120 day hold = 17,540 homes held = $415M interest → 90 day hold = 13,155 homes held = $311M → 60 day hold = 8,770 homes held = $207M Same revenue, same margin, over $200M of profit swing purely from selling faster. Speed isn't a nice slide, it's cash. Then stock comp, which is equity paid to staff rather than cash. It's a real cost because it dilutes existing shareholders. → regular grants run $15M a quarter, so $60M a year → market condition RSUs are $105M a quarter right now, but those only pay out if the share price hits set targets, and they expense over a fixed vesting period rather than forever → blending the two, call it $125M a year once things settle Depreciation and amortisation is small. → $8M a quarter, so roughly $32M a year → small because homes sit on the books as inventory, not as fixed assets Adding up everything below EBITDA: → interest $311M → stock comp $125M → depreciation $32M → $468M total Which gives the answer. → $1.56B EBITDA − $468M = $1.09B pre tax → years of past losses carry forward, so tax is minimal at first → $1.09B net profit, about $1.13 EPS → 30x = $34 | 45x = $51 | 60x = $68 Once those losses are used up and it's taxed at the standard 21% US corporate rate: → $861M net profit, about $0.89 EPS → 30x = $27 | 45x = $40 | 60x = $54 So the answer is roughly $860M to $1.09B of net profit. Timing, and whether 2027 is realistic, in the conclusion below 🧵👇