One Good Cohort Is Not Yet a Business, But It Is the First Real Evidence

Results vs. Consensus

Quality of Beat/Miss

• Revenue: Higher quality than last quarter's beat, and for the opposite reason. In Q3 the revenue beat came from dumping aged inventory at a margin cost. Here, management guided revenue down approximately 35% sequentially and delivered down 20%, and the CFO attributed the difference explicitly to selling through more aged inventory than forecast because resale velocity improved. Same mechanism, better execution: the company cleared the old book faster than it expected to, which is exactly what it said it would try to do.
• Margins: The worst reported contribution margin in the recent record, at 1.0%, alongside the best forward evidence the company has produced. Both are true because of a single disclosure: 94% of homes sold in Q4 were acquired before October, so the reported margin is almost entirely a readout on the prior regime's book. Gross margin of 7.7% actually rose 50bps sequentially, and adjusted gross margin fell to 6.1% from 7.0%, the difference being inventory valuation adjustments on homes still held.
• EPS: The $0.02 adjusted beat is real but small. The GAAP figure requires care: the -$1.26 loss per share includes a $933M non-cash loss on extinguishment from settling the majority of the 2030 convertible notes in equity, an event announced at the Q3 call and therefore not new information. Excluding it, the adjusted net loss of $62M is essentially flat against Q3's $61M. Note also that Q4 stock-based compensation was roughly $105M, derived from full-year SBC of $159M against $54M through nine months, and that gap between GAAP and adjusted results is now structural rather than one-off.

Key Operating Metrics

Opendoor reports a single operating segment. The table above is the segmentation that matters, and this quarter it splits cleanly into two opposing stories. The stock and flow metrics that describe the new regime, homes purchased, homes under contract, aged inventory percentage, fixed operating expenses, all improved. The metrics that describe the outcome of past decisions, contribution profit and margin, all deteriorated. That divergence is the quarter.

Acquisitions and the funnel

1,706 homes purchased is a 46% sequential increase against a guide of at least 35%, and the weekly contract cadence disclosed on the call is running well ahead of that: 537 signed in the most recent week against 236 at the time of the Q3 call and 128 in the final week of Q3. Management also disclosed that Opendoor availability went from roughly a third of US homes to nearly every homeowner in the Lower 48 in about ten weeks, built across close to 200 MLS datasets and more than 100 brokerage regions.

"It took Opendoor from 2015 to 2025 to become available as an option for about 1/3 of the homes in the U.S., 10 years. Opendoor 2.0 almost tripled that in about 10 weeks." — Kaz Nejatian, CEO

Assessment: The coverage expansion is the most underappreciated disclosure in the release, because it changes the shape of the top of the funnel permanently rather than for one quarter. Management says it is generating thousands of incremental qualified leads per week with no incremental marketing spend. That claim is unverifiable from outside and is exactly the kind of thing that should show up in the Q1 acquisition line if true.

Resale velocity and inventory health

Aged inventory fell from 51% of listings to 33% in one quarter, an 18-point move, and days in possession fell roughly 25% comparing the last week of Q4 against the last week of Q3. Total inventory fell to $925M across 2,867 homes.

"The percentage of Opendoor homes on the market for greater than 120 days decreased from 51% at the end of Q3 to 33% at the end of Q4, an 18 percentage point improvement in a single quarter." — Christy Schwartz, CFO

Assessment: This is the single most checkable improvement in the print and the one we flagged last quarter as the metric to watch. It is also the mechanism by which Q1 margins should inflect, since the remaining legacy book is now a third of listings rather than half. One caution: a falling aged-inventory percentage during a quarter of rising acquisitions partly reflects denominator growth, not only faster clearance of old homes.

Cash Plus and the capital-light mix shift

Cash Plus, the product where a seller takes less cash upfront in exchange for lower fees and retains more of the price risk, moved from roughly 19% of contracts in the last week of Q3 to 35% of a substantially larger base, which management sized as more than 600% growth in absolute terms.

"Cash plus. This is an underappreciated as a risk management tool. When a seller opts in, they're retaining more of the price risk, and we earn fees with less capital at risk. That structurally shifts our exposure." — Christy Schwartz, CFO

Assessment: This is the most important structural development of the quarter and it is arriving two steps ahead of the stated four-step plan. If a third of volume carries materially less balance-sheet risk, the terminal capital intensity of the business is lower than the Q3 framework implied, which is precisely the change that would eventually justify a different multiple.

Key Topics & Management Commentary

Overall Management Tone: Management opened by grading itself against the four-step plan laid out in November, replaying its own prior remarks and marking each item green, which is a level of self-imposed accountability that is rare and, on this evidence, earned. The posture was noticeably more confident than the Q3 call and considerably more data-led, with the October cohort doing most of the argumentative work in place of the rhetoric that carried the prior quarter. Where the tone remained loose was on anything requiring a number rather than a direction: the mortgage attach rate, the acquisition ramp shape, and the dilution arithmetic were all handled qualitatively.

1. The October cohort as the load-bearing evidence

Almost the entire bull case in this print rests on one cohort. October 2025 was the first full month of acquisitions under the new pricing and selection regime, and by the time of the call it was more than 50% sold or under resale contract, clearing at roughly twice the velocity of the October 2024 cohort and around 50% faster than October 2023.

"What makes this cohort significant isn't just the margin level. It's the shape of the cohort. From 10% to 50% sold through, our margin degradation relative to home price appreciation has been the lowest of any cohort in our history, not any October cohort, any cohort period." — Kaz Nejatian, CEO

The CFO put a figure on the shape claim: from 10% to 50% sell-through, cohort margins were up about 18 basis points rather than degrading, and this occurred over a period in which national home prices fell roughly 300 basis points.

Assessment: The shape argument is stronger evidence than the level argument. A cohort that holds margin as it sells through has been priced with genuine dispersion rather than optimism at the front end, which is exactly the deficiency management diagnosed in the legacy book. The caveat is sample: this is one month of acquisitions, roughly double the depressed prior run rate, and the second half of any cohort is where the difficult homes concentrate.

2. Self-grading against the four-step plan

The call opened with management replaying its own November remarks and marking each of the four steps. On the stated evidence: step one on track, acquisition velocity up 300% since September with 537 homes contracted in the most recent week; step two delivered, with unit economics improving alongside volume; step three delivered, direct-to-consumer acquisition contracts up nearly 700% week-over-comparable-week with days in possession down almost 25%; step four beginning, via Cash Plus at 35% of contracts.

"We laid out a plan for you. We're going to turn this company around, and we laid that plan out during Opendoor 2.0's first open house. That plan is working. We're green all across." — Kaz Nejatian, CEO

Assessment: We agree with the grading on the evidence presented, and the willingness to be graded is itself worth something in a story that has previously run on assertion. The obvious risk in a self-scored framework is that the scorer chooses the metrics, and the metric management did not grade itself on, contribution margin, is the one that printed at 1.0%.

3. The choice to run acquisitions below the top of the range

The most substantive exchange on the call concerned why acquisition contracts are tracking at or below the low end of the company's own published dashboard targets. Management's answer was that this is a deliberate allocation of engineering capacity toward product rather than growth.

"In a product company, you always basically have 2 choices. You can invest in product or you can invest in growth. ... When you invest in product, you're putting fuel in the tank. When you invest in growth, you're spending that fuel. ... Right now, I want to invest in having the best product we can before we go and invest in growth again." — Kaz Nejatian, CEO

The concrete example given was pulling four engineers onto the mortgage product in early January instead of improving the consumer app, producing a beta launch within ten weeks. The CFO separately reframed the 2026 acquisition trajectory as more weighted to the back half of the year, with Q1 and Q2 spent refining conversion and pricing.

Assessment: The reasoning is defensible on its own terms and the mortgage build is genuine evidence of throughput. It is also, functionally, a guidance reduction for the first half of 2026 delivered as a strategy statement. The exit-rate target of roughly 6,000 quarterly acquisitions is unchanged, which means the required ramp in the second half is now steeper. That is a risk transfer from H1 to H2, and it deserves to be named as such.

4. Adjusted EBITDA profitability pulled forward to Q2

Asked where the company stands against its profitability plan, the CEO volunteered a milestone that was not in the release: annual-basis adjusted EBITDA profitability beginning in Q2 2026, a step that precedes the year-end adjusted net income target.

"Before we become ANI profitable, we're going to become adjusted EBITDA profitable. Now I think adjusted EBITDA is not the best metric for measuring our business, and I'm not about to give you guidance. ... But to be transparent about it, the plan I have on this laptop right now has us being adjusted EBITDA profitable on an annual basis starting in Q2." — Kaz Nejatian, CEO

Assessment: This is a new and considerably nearer checkpoint than anything offered in November, and it is the most useful thing said on the call for modelling purposes. The hedge attached to it, that it is a plan rather than guidance, is becoming a familiar pattern: this management team consistently offers specificity and simultaneously disclaims it. We treat it as a soft commitment and will grade it in August.

5. Product-led cost reduction

Management explicitly rejected cost-cutting as a strategy and framed the expense reductions as a byproduct of engineering work. The specifics were unusually granular: annualised hosting costs down from $12M entering 2025 to under $5M exiting it, valuation model runtime cut from twelve hours to five and a half with a roughly 90% reduction in feature-building pipeline cost, and third-party vision tools replaced in-house, cutting processing time for 100,000 listings from 34 hours to four.

"We didn't focus on cutting costs to improve margins. We focused on improving the product and taking pride in our code and the costs started disappearing kind of on their own." — Kaz Nejatian, CEO

Assessment: Fixed operating expenses of $35M against $37M in Q3 and $43M a year ago corroborate the direction, and trailing twelve-month operations expense held at 1.3% of revenue as committed. The individual savings cited, however, total a few million dollars annually against a cost base an order of magnitude larger. The engineering detail is credible and the aggregate impact is smaller than its prominence on the call implies.

6. Product surface: tailored offers and the self-assessment app

The pricing model changed from a single take-it-or-leave-it offer to a structure where the seller chooses how much cash to take upfront and the fee adjusts accordingly. Management framed this as a product fix rather than an underwriting fix, on the logic that a seller with 80% equity has different needs from one with 20% and the old product forced both to pay for risk neither party wanted.

"Every dollar that a seller does not need upfront is a dollar that we don't need to put at risk and a dollar that they don't need to pay for. So they pay less and we make more." — Kaz Nejatian, CEO

Alongside it, a self-assessment app now lets sellers photograph their own homes for AI assessment. Management said assessments in January nearly doubled against September, and that about half of the homes assessed so far in February required nobody to visit the property.

Assessment: The tailored-offer structure is the cleanest articulation yet of how tighter spreads and lower risk can coexist, and it is the mechanism behind the Cash Plus mix shift. The self-assessment app attacks the variable cost per acquisition directly. Both are the right shape; neither has a disclosed conversion or margin figure attached.

7. Mortgage, built in ten weeks

The mortgage product moved from a January start to a beta launch in the week of the call. Management declined to size the opportunity, saying it would underpromise on attach rates, while hinting at a partner-facing angle with homebuilders.

"We built a mortgage product in less than 10 weeks. People told me it would take at least a year. We did it in 10 weeks." — Kaz Nejatian, CEO

Assessment: Ancillary services attached to a captive transaction remain the most credible route to a structurally higher contribution margin, and a ten-week build is a real datapoint on organisational velocity. We continue to model no ancillary contribution until attach rates are disclosed. A mortgage business also brings regulatory and warehouse-funding surface area that has not been discussed at all.

8. The executive team was replaced wholesale

Management disclosed that of eighteen direct reports, ten did not work at Opendoor a year ago, and that no member of the executive team listed in the current 10-K appeared in the prior filing. The company now has a new chief operating officer, chief financial officer, president, chief growth officer, chief people officer and chief business officer.

Assessment: Complete executive turnover in five months is consistent with a genuine refounding and also removes essentially all institutional memory of how the prior book was priced. Investors should read the October cohort's performance in that light: it is the work of a team with no attachment to the old model, which is both the point and the risk.

9. Downside protection against a housing decline

The most useful risk disclosure came in response to a shareholder question about a further 5% to 10% decline in national home prices. The CFO gave a structured three-part answer covering selection, velocity and Cash Plus, and declined to claim immunity.

"Are we immune from a 5% to 10% decline? No, nobody is. But the question isn't whether it hurts, it's whether Opendoor is structurally positioned to navigate and withstand it. ... Over the past 5 months, national home prices came down roughly 300 basis points. During that exact period, our contribution margins on that cohort held steady or slightly improved." — Christy Schwartz, CFO

Assessment: This is the best answer given on either of the two calls we have now covered, because it is falsifiable and it was tested against an actual, if modest, price decline. A 300 basis point move is not a 5% to 10% move, and the cohort has not yet met a genuine downturn. But the framing that velocity is the hedge, not spread, is analytically correct and represents a real change in how this balance sheet is being run.

10. The stock price and management incentives

Asked how investors should assess progress with the shares down significantly during his tenure, the CEO disclosed a $1 salary with compensation contingent on share price levels not seen in years, then declined to manage to the share price.

"My salary is $1 and unless stock levels see levels they haven't seen in years, I don't get paid. So I don't want you to take what I'm about to say the wrong way, but ... I don't manage the stock price. I manage the business and look at what we're building." — Kaz Nejatian, CEO

Assessment: The alignment is real and unusually strong. It also means the CEO's personal payoff is convex in the share price, which is the incentive structure that historically produces aggressive growth and aggressive equity issuance. Investors should hold both facts at once: this is excellent alignment on direction and imperfect alignment on dilution.

11. Full-year 2025 in context

The full-year figures put the transition year in perspective. Revenue of $4,371M was down 15% from $5,153M, gross profit fell to $350M from $433M at an 8.0% margin against 8.4%, and net loss widened to $1,300M from $392M, of which $924M was loss on debt extinguishment. Contribution profit fell to $150M from $242M with margin at 3.4% against 4.7%, and adjusted EBITDA improved to -$83M from -$142M. Adjusted net loss narrowed to $195M from $258M. Homes purchased fell to 8,241 from 14,684.

Assessment: Read across the full year, 2025 was a controlled shrinkage: revenue and volume down sharply, cost base down more sharply, and the adjusted loss lines genuinely narrower than 2024 despite a smaller business. That is the correct shape for a company deleveraging its operational risk. The equity holder paid for it in share count.

Guidance & Outlook

The guidance architecture is unchanged from November: directional guideposts, a weekly public dashboard, and two metrics committed to quarterly disclosure. What changed is that management now has a track record against it, and on this quarter's evidence the guideposts were set conservatively. Three of the four quantified Q4 items came in better than guided, and the fourth, contribution margin, landed where management said it would.

Two forward statements matter more than the rest. The first is that Q1 contribution margin exits the quarter at the highest level since Q2 2024, which given the Q2 2024 comparison implies a monthly exit run rate somewhere near 5%. The second is the unscripted comment that the internal plan has adjusted EBITDA turning positive on an annual basis from Q2 2026.

"October's margins thus far have come in above our long-term target range. That's a good problem to have, but you should not expect every quarter to look like October on a margin basis. You should expect us to reinvest that spread advantage into growth, faster turns, broader coverage and more competitive offers, while we aim to maintain contribution margins within our target range." — Christy Schwartz, CFO

Implied Q1 shape: Revenue down approximately 10% sequentially puts Q1 near $660M, which remains inventory-constrained by the low Q3 and Q4 purchase volumes. With adjusted EBITDA guided to a loss of low-to-mid $30Ms against Q4's -$43M, and contribution margin exiting materially higher, Q1 should be the first quarter where the reported margin line begins to reflect the new book rather than the old.

Street at: Consensus had modelled roughly $595M for Q4 against the $736M print, so the Street was too pessimistic on sell-through by a wide margin. Expect Q1 estimates to be marked up on revenue and for the debate to move to the margin exit rate.

Guidance style: Conservative, on one quarter of evidence. Management set guideposts it beat on three of four quantified lines. That is a useful pattern to note and too small a sample to rely on.

Analyst Q&A Highlights

The format remains the shareholder-question livestream introduced last quarter, with sell-side questions submitted in writing and read by investor relations. There are still no live follow-ups, which continues to protect management from the second question.

Acquisition volume tracking below the published target range

The sharpest question of the call came from the sell side and pointed directly at the company's own accountability dashboard: contract volumes were running at or below the low end of the published targets. Management acknowledged the substance and reframed it as a deliberate allocation choice rather than a shortfall.

Q: "According to your accountability dashboard, acquisition contract volumes are running at or below the low end of your targets. Aside from seasonal factors, what are the primary macro or pricing drivers preventing a faster ramp?"
— Ryan Tomasello, KBW

A: "I reserve the right to wake up smarter every single morning than I was yesterday. ... So if we wanted to hit the top end of the range, we could do it very easily. But that would come at a cost. ... I'm choosing to work on the funnel rather than the number that just happens to come out at the end of the funnel."
— Kaz Nejatian, CEO

Assessment: A dashboard published as an accountability mechanism has, within one quarter, produced a miss that was answered by explaining why the metric is the wrong metric. The underlying argument about product investment is reasonable, and the mortgage beta is evidence it was spent well. But the pattern is worth watching closely: the value of self-imposed accountability depends entirely on not renegotiating it when it binds.

What the ramp in acquisitions revealed about home quality

A question on surprises encountered while scaling purchases produced the most quantitatively useful answer of the call, on the shape of cohort margin degradation.

Q: "What has stood out to you so far about the quality and profitability of the homes you're buying since ramping up acquisitions? Any surprises as you pick up the pace and have those led to any changes in your approach?"
— Dae Lee, JPMorgan

A: "What surprised us was how quickly some of the changes we put in place showed up in our results. ... In this last October cohort, we saw that from 10% to 50%, our margins basically held flat. They were actually up about 18 bps, and that was during a tough market. And so seeing that flat degradation was very promising."
— Christy Schwartz, CFO

Assessment: Flat-to-positive margin degradation through the first half of a cohort's sell-through is the single most persuasive number produced this quarter, because it speaks to pricing precision rather than to luck on the easy homes. It also establishes a clean benchmark: if the same statistic holds for the November through January cohorts, the thesis moves from one proof point to a pattern.

Resilience to a further decline in home prices

A shareholder question on a 5% to 10% national price decline drew a structured answer on how the risk toolkit has changed, and an explicit refusal to claim immunity.

Q: "If home prices drop another 5% to 10% nationally, what happens to your margins and inventory risk? How prepared is Opendoor for that scenario?"
— Victoria B., shareholder question

A: "Opendoor 2.0 is built to move homes, not hold them. The longer you hold when the market turns, the more it hurts. So the whole game is turn faster. ... Previously, the primary lever was spread, price conservatively enough to absorb a market move, and that works, but it's blunt and it has terrible adverse selection."
— Christy Schwartz, CFO

Assessment: Velocity as the primary hedge is the correct answer for this business model and is a genuine change from the prior regime's spread-based defence. The unresolved question is what happens to velocity itself in a real downturn, since the buyer side slows at exactly the moment the seller side accelerates. The toolkit is better; the correlation risk is unaddressed.

Timing and shape of the ancillary-services opportunity

A sell-side question on extending into adjacent products produced the mortgage beta disclosure and an unusually candid admission about the central unknown.

Q: "There's some testing of a mortgage product with Lennar. Where is Opendoor in extending the services opportunity? What should our expectation for 2026 be for Opendoor to be able to attach more adjacent products to transactions?"
— Andrew Boone, Citizens

A: "So I think the primary unknown question is this. What will be our attach for mortgage and adjacent products, right? How many people who buy or sell a home from Opendoor take a mortgage from us? I have a hypothesis, but not a good answer."
— Kaz Nejatian, CEO

Assessment: The honesty is welcome and the answer is the correct one at this stage. It also confirms that nothing in the 2026 profitability framework depends on ancillary attach, which means the mortgage product is genuine optionality rather than a load-bearing assumption. We continue to value it at zero until an attach rate is disclosed.

Becoming the default channel for home sellers

A long-form shareholder question about consumer behaviour produced the only market-share datapoint offered on the call.

Q: "How do we get to the point where sellers think, I'm going to try Opendoor first before even considering going via a realtor?"
— Catherine, shareholder question

A: "In markets that we've been operating for a while in our mature markets, 20% of folks who go on to list their home before they call a person, they try opendoor.com. That's the beginning of looking like a default."
— Kaz Nejatian, CEO

Assessment: A 20% pre-listing consideration rate in mature markets is a real and previously undisclosed figure, and it is the best available proxy for whether the brand has consumer pull independent of marketing spend. The obvious follow-up, what the equivalent figure is in the newly covered markets that make up most of the expanded footprint, was not asked and not volunteered.

How AI is actually deployed in underwriting

A video-submitted question on internal AI use drew a specific description of what changed in the underwriting workflow rather than a generic capability claim.

Q: "What are you guys using it for internally? And how should that change the customer experience in the Opendoor product moving forward?"
— Shareholder video submission

A: "Our analysts are no longer doing valuations today. They're auditing what AI has prepared. They are working on evaluating the output rather than doing paperwork to prepare the input."
— Kaz Nejatian, CEO

Assessment: Consistent with last quarter's eleven-touchpoints-to-one claim and now corroborated by the fixed operating expense line falling while volume rose 46%. The framing that humans handle the last mile, the atypical asset the model prices badly, is the right division of labour and implicitly concedes that the model has known blind spots.

What They're NOT Saying

1. Dilution arithmetic, again: Weighted-average shares rose to 869.8 million from 741.9 million, up 17% sequentially and 21% year over year, as the convertible settlement landed. Neither the call nor the release quantified the share count consequences or the pro-forma fully diluted count including the $9, $13 and $17 warrants. This is the second consecutive quarter the question has gone unaddressed.
2. Stock-based compensation: Full-year SBC of $159M against $114M in 2024 implies roughly $105M in Q4 alone, which is the dominant driver of general and administrative expense rising to $129M from $48M. The call did not mention SBC once, and the entire discussion of cost discipline was conducted on measures that exclude it.
3. Contribution margin for the quarter: The self-grading exercise covered acquisitions, velocity, direct-to-consumer mix and operating leverage. It did not grade the metric management itself calls the key one, which printed at 1.0%. That omission is a choice.
4. The spread: Still undisclosed in basis points or dollars, in any period, despite spread compression being the central mechanism of the entire strategy. Management describes dispersion qualitatively and asks to be judged on outputs.
5. Cash Plus unit economics: Cash Plus is now 35% of contracts and is presented as both a growth and a risk-management win, but no fee, take rate or contribution figure was given for it. As mix shifts toward a product with different economics, reported contribution margin becomes harder to interpret without that disclosure.
6. What happens to the H2 ramp if H1 investment does not pay off: The 6,000-per-quarter exit target is unchanged while the first-half trajectory has been lowered, which mathematically steepens the second half. No sensitivity was offered around what the exit rate looks like if the funnel work delivers less than expected.
7. Mortgage funding structure: A mortgage product launched in beta requires warehouse lines, licensing and compliance infrastructure. None of it was discussed, and none of the associated cost appears in the guidance framework.

Market Reaction

• Pre-print setup: The shares closed at $4.65 on February 19, down 20.2% year to date against the S&P 500's +0.2%, and down 27.2% over the preceding thirty days from $6.39 on January 20. Over twelve months the stock remained up 201.9%. The 52-week closing range entering the print was $0.51 to $10.52.
• Next-day session: The shares opened at $5.50 on February 20, an 18.3% gap up, traded between $4.74 and $5.59, and closed at $5.00, up 7.5% or $0.35. Volume was 132.3 million shares against a thirty-day average of 56.4 million, or 2.3 times normal. The S&P 500 closed up 0.7%.

The initial reaction was considerably stronger than the close. Reporting in the hours after the print had the stock up roughly 16.5%, consistent with the 18.3% opening gap, and half of that was given back during the session. The gap up is the market pricing the guidepost sweep and the October cohort disclosure. The fade is the market reading the actual income statement, where contribution margin printed 1.0% and per-home contribution profit fell to $4K.

Note the symmetry with last quarter. In November the stock gapped down 18.7% on a bad print and closed flat as the call reframed it. In February it gapped up 18.3% on a good narrative and closed up 7.5% as the financials reasserted themselves. In both cases the session compressed roughly half the initial move. That is what a market does when it believes the story and cannot yet find it in the numbers.

The pre-print setup matters here too. Entering the print the stock had fallen 27% in thirty days and 20% year to date, so this was a beaten-down name going into a quarter where management had a scorecard to defend. A 7.5% advance on 2.3 times normal volume recovers less than a third of the thirty-day decline. The market's verdict is that the evidence is real and not yet sufficient, which is close to our own.

Street Perspective

Debate: Is the October cohort a proof point or a data point?

Bull view: The bull case on the Street is that the cohort's flat margin degradation from 10% to 50% sell-through, achieved while national home prices fell roughly 300 basis points, is evidence of genuine pricing dispersion rather than a favourable draw, and that every subsequent cohort inherits the same machinery plus incremental learning.

Bear view: The bear camp contends that a single month's acquisitions at roughly double a depressed run rate is a small and self-selected sample, that the difficult half of any cohort sells last, and that management has strong incentives to lead with its best month.

Our take: The bulls have the stronger argument on mechanism and the bears the stronger argument on sample. What tips it slightly toward the bulls is that the shape claim, not the level claim, is what management chose to emphasise, and shape is the harder thing to get lucky on. We upgrade this pillar from neutral toward constructive without yet crediting it fully, and we will grade it again on the November through January cohorts.

Debate: Does the H1 slowdown jeopardise the 2026 exit rate?

Bull view: Some desks argue that front-loading engineering investment into funnel conversion and the mortgage stack is exactly what a product-led operator should do while inventory is being rebuilt anyway, and that the levers described are more numerous than they were three months ago.

Bear view: Skeptics observe that contract volumes are already tracking at or below the low end of the company's own published range, that the ramp has now been explicitly rebalanced toward the second half, and that an unchanged exit target with a lowered first half is simply a steeper required trajectory.

Our take: The bears are right about the arithmetic and the bulls are right that the H1 spend is defensible. The tell will be Q1: if acquisitions grow another 40%-plus sequentially while contribution margin exits near 5%, the reallocation was costless. If acquisitions grow 20% and margin exits at 4%, the 6,000 exit rate becomes a stretch and the year-end profitability commitment goes with it.

Debate: Is adjusted net income the right measure of this business?

Bull view: The bull framing is that GAAP results are dominated by non-cash items, the $933M extinguishment charge this quarter and roughly $105M of stock-based compensation, and that adjusted net loss narrowing to $195M for the year from $258M is the honest read on operating trajectory.

Bear view: The bear camp responds that stock-based compensation at this scale is not a non-cash irrelevance but the mechanism by which the equity claim is diluted, that adding it back while share count rises 21% year over year double-counts the benefit, and that a profitability target struck on a measure which excludes it is a target that can be met without per-share value creation.

Our take: The bears have this one, and it is the most important unresolved issue in the story. The extinguishment charge is genuinely a one-time item and should be excluded. Stock-based compensation running near $105M a quarter against a company targeting adjusted net income breakeven is not an adjustment, it is the cost of the labour. We will assess the 2026 target on both bases and expect the gap between them to be the central debate by the second half.

Model Update Needed

Valuation impact: At $5.00 on 869.8 million weighted-average shares the equity is capitalised at roughly $4.3B, against trailing twelve-month gross profit of $350M and trailing contribution profit of $150M. That is roughly 12 times trailing gross profit and 29 times trailing contribution profit, on a business whose contribution profit fell 38% during the year in question. The multiple is only defensible against a 2026 in which contribution profit roughly triples, which is what the company's own framework implies. Our fair-value range moves to $4.50 to $6.50, modestly wider than last quarter to reflect both the improved evidence and the higher share count. The current price sits near the middle.

Thesis Scorecard Post-Earnings

Overall: Thesis strengthened. Three pillars improved this quarter, including the upgrade of the central contribution-margin pillar from neutral to confirmed and the first genuine challenge to the housing-risk bear point. The offset is that the dilution bear point escalated from emerging to confirmed, and it is now the binding constraint on how much of the operational improvement reaches a per-share number.

Action: Hold, with the balance of evidence improving. We move to Outperform on a Q1 print that shows contribution margin for the quarter at or above 4% with acquisitions above 2,300 homes, since that combination would demonstrate the cohort thesis surviving contact with scale. We move to Underperform on a quarter where margin improvement requires acquisitions to stall, or on any indication that the year-end profitability target is being met through further equity issuance.