← back to ranking

OPEN · Opendoor Technologies Inc.

Real Estate - Services · mkt cap $4.1B · calls: Q1 FY2026 vs Q4 FY2025
77.0 conviction · conf-adj 77

conf 3/10 partial

enthusiasm:30.0 · trend:8 · quantifies:0 · impact:0 · under_radar:14 · credibility:12 · business_impact:10 · disruption:0 · commitment:0 · confirmation:3

Enthusiasm latest 10 / prev 9 (rising)

Opendoor's AI thesis is that AI and ML are being embedded into underwriting, pricing, repairs, title, mortgage, customer experience, and internal workflows to make the business faster, lower cost, and less dependent on macro forecasting. Credibility improved in Q1 FY2026 because management moved from broad AI-native claims to more operating metrics tied to fall-through, renovation spend, FTE redeployment, intake time, mortgage rates, and fixed OpEx discipline. The risk is that many quantified benefits are pilots or workflow anecdotes rather than fully reconciled financial impact at company scale.

GROUNDED NEXT-FY IMPACT vs CONSENSUS

Grounded on actual base — revenue $4.4B · net income $-1.3B · net margin -29.7% · diluted EPS -1.7

These are next-fiscal-year annual uplift estimates, not next-quarter numbers.

Aggregate next-FY est. rev uplift: % · next-FY EPS uplift: % · vs analysts: inline · priced in: low (model's call-read: medium; verdict above is the hard-data one used for ranking) · confidence: 3/10

ClaimFigureArithmeticNext-FY Rev %Next-FY EPS %
Title intake 5h→15min
productivity · soft
5h to 15min (~95% time cut)(300-15)/300=95.0% faster; title-intake volume and labor cost per intake not disclosed → cannot anchor revenue/EPS dollars.
Replaced $0.5M email system w/ Claude skill
cost
$0.5MDisclosed base $0.5M. After-tax @21%: 0.5M*0.79=$0.395M. Topline 0; $0.5M/$4,371M=0.011%. EPS% null (net income negative).0
SOX deliverables 20h→1-min query
productivity · soft
20h to 1-min query(1,200-1)/1,200=99.92% faster; deliverable count and finance labor cost not disclosed → unsizable.
Voice-bot seller contract 30min→5
productivity · soft
30min to 5min(30-5)/30=83.3% faster; could lift seller throughput but no contract volume or $ base disclosed → unsizable.
72 manual exports/month → 1 pipeline
productivity · soft
72 exports→1 pipelineLabor automation; time per export and FTE/$ base not disclosed → unsizable.
AI repair-negotiation cut buyer fall-through >double digits
engagement · soft
>10% fall-through reduction% fall-through reduction with no disclosed fall-through base, buyer volume, or revenue-per-closing → cannot map to revenue line.
Pre-list reno spend −10–20%/home (pilot)
cost · soft
10–20% per home% saving with no per-home reno $ base disclosed and pilot-only scope → unsizable.
Ticket triage redeployed 3 FTEs
productivity · soft
3 FTEsRedeployed (not eliminated) → no net cost removed; no salary base disclosed → no saving to book.
Big-bank mortgage 340 bps rev/loan
other · soft
340 bps/loanCompetitor benchmark (3.40% rev/loan), not OPEN revenue; no OPEN loan volume/principal disclosed → context only.
OPEN mortgage rates ~100 bps below market
engagement · soft
100 bpsPricing advantage may lift attach/conversion but no attach rate, loan volume, or mortgage-revenue base disclosed → unsizable.
Customer payments 10–15% lower
engagement · soft
10–15%Customer-benefit metric, not a revenue figure; no customer count, elasticity, or revenue base to size topline.
Seller disclosure automation (hours/home)
productivity · soft
hours per homeVague time figure; exact hours, homes processed, and labor cost not disclosed → unsizable.
Self-assessment capacity nearly doubled (Jan vs Sep)
productivity · soft
~2x assessmentsThroughput ~2x vs Sep 2025; baseline homes assessed and conversion-to-revenue not disclosed → unsizable enabler.
~Half of assessments need 0 site visits
cost · soft
~50% visits avoided~50% site visits avoided; per-visit $ and visit-volume base not disclosed → unsizable.
Coverage 1-in-3 → nearly every Lower-48 homeowner
engagement · soft
~3x addressable households~33%→~100% coverage (~3x / +200% addressable), but no conversion rate or revenue-per-home → cannot size topline $; largest qualitative lever.
>1,000 real-time data pipelines
other · soft
>1,000 pipelinesInfrastructure scale, no cost/revenue/throughput $ attached → unsizable.
Valuation model runtime 12h→5.5h (−50%)
productivity
−50% runtime (12h→5.5h)(12-5.5)/12=54.2% cut; $ outcome captured in the ≥$1M/yr model-pipeline saving → not additive, no separate $.0
Feature-building DAGs 90% cheaper
cost
−90% DAG costComponent of the ≥$1M/yr model-pipeline saving — rolled into that figure to avoid double-count.0
Model-pipeline changes save ≥$1M/yr
cost
≥$1M/yrDisclosed base $1M. After-tax @21%: 1M*0.79=$0.79M. Topline 0; $1M/$4,371M=0.023%. EPS% null (net income negative).0
Vision model: 100k listings 34h→4h, in-house
productivity
34h→4h (~88% faster)Speed/in-house move; cost outcome captured in the >$1M SaaS-replacement figure → not additive.0
Replaced SaaS tools, >$1M cost cut
cost
>$1MDisclosed base $1M. After-tax @21%: 1M*0.79=$0.79M. Topline 0; $1M/$4,371M=0.023%. EPS% null (net income negative).0

Assumptions: Tax rate 21% on cost savings (company is loss-making, so the tax shield is notional — savings shown gross and after-tax). Cost savings treated as annual run-rate fully in next FY (FY2026). NO incremental net margin applied because ZERO claims anchor an incremental-revenue dollar figure — all topline claims (coverage 3x, capacity 2x, 100 bps rate edge) are unanchored TAM/throughput/pricing levers with no conversion rate or revenue-per-unit disclosed, so they cannot be sized. To avoid double-counting, the runtime/DAG/vision-model efficiency claims are folded into the two disclosed $1M savings figures rather than added separately. EPS-uplift % is null throughout: net income is −$1.3B, so any % off a negative base is meaningless. Pure adopter (Opendoor applies AI to its own iBuying/mortgage operations; it sells no AI compute/infrastructure), so supplier-side = nil.

Top line: No claim anchors an incremental-revenue dollar amount, so est_rev_uplift_pct is null. The genuinely material topline lever — AI-enabled coverage expanding from ~1-in-3 to nearly every Lower-48 homeowner (~3x serviceable households), plus ~2x self-assessment throughput and a ~100 bps mortgage-rate edge — is real but unsized: management gave no conversion rate or revenue-per-home, so it can't be converted to a credible $ uplift. Notably consensus revenue is FLAT-TO-DOWN (FY25 est $4,229M vs $4,371M actual), implying analysts are NOT yet crediting an AI-driven volume ramp.

Bottom line: The only HARD, anchored impact is three disclosed cost savings: $0.5M email-system replacement + ≥$1M/yr model-pipeline + >$1M SaaS replacement = ≥$2.5M/yr gross (~$1.98M after-tax @21%). That is just 0.057% of $4,371M revenue and is immaterial against a −$1.3B net loss — it cannot meaningfully move EPS. The richer efficiency story (title intake 5h→15min, SOX 20h→1-query, ~50% of assessments needing no site visit, 3 FTEs redeployed) is directionally consistent with management's 'declining opex as a % of revenue' guidance but is entirely unsized (no FTE/$ bases disclosed).

[EPS uplift n/m (loss-making base)] Quantifiable AI impact = ~$2.5M/yr gross savings = 0.057% of revenue and ~0.15% of the $1.3B loss — far too small to beat consensus, which already models a narrowing loss (EPS −0.40, net income −$285M for FY25 vs −$1.70 / −$1.3B actual). So the hard math is firmly inline/immaterial, not ahead. The one place the math could point ABOVE consensus is the topline: consensus revenue is flat-to-declining ($4,229M) while AI tripled addressable households and doubled assessment capacity — if that converts, it exceeds what analysts assume. But with no anchored revenue figure it stays soft, and the efficiency narrative is already captured in management's 'declining opex %' guidance (i.e. partially priced). Net: efficiency is priced in, the unsized TAM expansion is the only un-priced optionality → medium.

MODEL CONSENSUS (impact)

partial

X and Y agree on every figure: all productivity/engagement claims unsizable (soft); only disclosed $-cost savings sized, all rev=0 and EPS null because net income is negative.

Conflicts reconciled
FieldOpus 4.8GPT-5.5
Rev uplift %0
EPS uplift %
Priced inmediummedium
vs analystsinlineunclear
Confidence33
Top lineNo claim anchors an incremental-revenue dollar amount, so est_rev_uplift_pct is null. The genuinely material topline lever — AI-enabled coverage expanding from ~1-in-3 to nearly every Lower-48 homeowner (~3x serviceable households), plus ~2x self-assessment throughput and a ~100 bps mortgage-rate edge — is real but unsized: management gave no conversion rate or revenue-per-home, so it can't be converted to a credible $ uplift. Notably consensus revenue is FLAT-TO-DOWN (FY25 est $4,229M vs $4,371M actual), implying analysts are NOT yet crediting an AI-driven volume ramp.No hard next-FY AI revenue uplift is disclosed. The largest top-line indicators are soft: coverage expands from about one-third of homeowners to nearly all Lower 48 homeowners, mortgage pricing is about 100 bps below market, and fall-through improves by a double-digit amount, but none provide volume or conversion bases.
Bottom lineThe only HARD, anchored impact is three disclosed cost savings: $0.5M email-system replacement + ≥$1M/yr model-pipeline + >$1M SaaS replacement = ≥$2.5M/yr gross (~$1.98M after-tax @21%). That is just 0.057% of $4,371M revenue and is immaterial against a −$1.3B net loss — it cannot meaningfully move EPS. The richer efficiency story (title intake 5h→15min, SOX 20h→1-query, ~50% of assessments needing no site visit, 3 FTEs redeployed) is directionally consistent with management's 'declining opex as a % of revenue' guidance but is entirely unsized (no FTE/$ bases disclosed).Hard disclosed savings total $2.5M pre-tax and $1.975M after tax: $0.5M email system + $1.0M model pipeline + at least $1.0M SaaS replacement. That is only 0.0572% of current revenue and 0.9158% of the $273M operating loss, before considering the many unanchored productivity claims.
ReasoningQuantifiable AI impact = ~$2.5M/yr gross savings = 0.057% of revenue and ~0.15% of the $1.3B loss — far too small to beat consensus, which already models a narrowing loss (EPS −0.40, net income −$285M for FY25 vs −$1.70 / −$1.3B actual). So the hard math is firmly inline/immaterial, not ahead. The one place the math could point ABOVE consensus is the topline: consensus revenue is flat-to-declining ($4,229M) while AI tripled addressable households and doubled assessment capacity — if that converts, it exceeds what analysts assume. But with no anchored revenue figure it stays soft, and the efficiency narrative is already captured in management's 'declining opex %' guidance (i.e. partially priced). Net: efficiency is priced in, the unsized TAM expansion is the only un-priced optionality → medium.Consensus 2025 revenue of $4.229B is 3.2499% below the supplied $4.371B current revenue base, while the hard AI revenue uplift calculated here is 0.0%. Consensus net income of -$284.8M is already $1.015B better than current -$1.300B net income; the hard after-tax AI savings of $1.975M are only 0.6936% of that consensus loss and about $0.0026 per diluted share, so the disclosed hard math does not clearly move estimates above consensus.

Rows highlighted where the two models disagreed.

QUANTIFICATIONS
title intake cycle time: up to 5 hours to 15 minutes (Q1 FY2026, bottomline)
“At title intake, it used to take us up to 5 hours. It now takes 15 minutes.”
legacy email system replacement: $0.5 million (Q1 FY2026, bottomline)
“One of our marketing managers replaced our $0.5 million life cycle legacy e-mail system with one Claude skill.”
SOX deliverables productivity: 20 hours to 1-minute query (Q1 FY2026, bottomline)
“Afinance team turned 20 hours of SOX deliverables into 1-minute query.”
seller contract time: 30 minutes to 5 (Q1 FY2026, both)
“Our voice bots dropped seller contract time from 30 minutes to 5.”
manual exports automation: 72 manual exports a month with 1 pipeline (Q1 FY2026, bottomline)
“We replaced 72 manual exports a month with 1 pipeline.”
buyer fall-through rate: over double digits (Q1 FY2026, both)
“an AI-powered repair negotiation tool cut our buyer fall-through rate by over double digits”
pre-list renovation spend: up to 10% to 20% per home (Q1 FY2026 pilot markets, bottomline)
“field managers are using AI scoping feedback, helping to reduce pre-list renovation spend by up to 10% to 20% per home in pilot markets”
ticket triage labor redeployment: 3 full-time employees (Q1 FY2026, bottomline)
“a ticket triage automation, redeployed 3 full-time employees from classification to resolution”
mortgage revenue benchmark: 340 basis points in revenue per loan (Q1 FY2026, both)
“Big bank lenders take about 340 basis points in revenue per loan.”
mortgage rate advantage: about 100 basis points below the market average (Q1 FY2026, topline)
“Today, our rates are running about 100 basis points below the market average.”
customer mortgage payment advantage: about 10% to 15% lower mortgage rates per month (Q1 FY2026, topline)
“And that translates to about 10% to 15% lower mortgage rates per month.”
seller disclosure automation: hours per home (Q4 FY2025, bottomline)
“It took hours per home, every single time.”
self-assessment capacity: nearly doubled (January 2026 compared to September 2025, both)
“And because there are no humans in the loop, in January, we nearly doubled the number of homes that we assessed compared to September.”
self-assessment human avoidance: about half (Q4 FY2025 call month, bottomline)
“And so far this month, about half the homes we've assessed have needed 0 people to show up at your doorstep.”
AI-enabled coverage expansion: about 1 to every 3 homeowners to nearly every homeowner in the Lower 48 (Q4 FY2025, topline)
“Thanks to AI, our product went from being available to about 1 to every 3 homeowners to being available to nearly every homeowner in the Lower 48.”
data pipeline scale: over 1,000 pipelines (Q4 FY2025, both)
“We now have a nearly real-time data ingestion with over 1,000 pipelines.”
valuation model runtime: reduced by about 50% from 12 hours to 5.5 hours (Q4 FY2025, bottomline)
“This means that the run time in our model was reduced by about 50% from 12 hours to 5.5 hours.”
feature building DAG cost: 90% cheaper (Q4 FY2025, bottomline)
“The entire pipeline can now run in a script that's about 50 lines of code and the feature building Dags are now 90% cheaper.”
annual savings from model pipeline changes: at least $1 million a year (Q4 FY2025, bottomline)
“This saves us at least $1 million a year.”
vision model processing time: 100,000 listings from 34 hours to just 4 (Q4 FY2025, bottomline)
“it also brought a core AI capability fully in-house and cut the processing time for 100,000 listings from 34 hours to just 4, cheaper, better, faster.”
SaaS replacement cost savings: over $1 million (Q4 FY2025, bottomline)
“On top of this, we also cut over $1 million in cost by replacing SaaS tools with more cost-effective and genuinely better AI alternative.”
home level days in possession model dispersion: about 2x (Q4 FY2025, both)
“We also built a new home level days in possession model using real demand signals. I think this is going to increase our dispersion by about 2x.”
PAST (realized)
CURRENT (now)
FORWARD (guidance)
TRACK RECORD — PROMISE vs DELIVERY

70/100 track record   delivers  6 calls reviewed

Opendoor made few quantified AI promises under prior management, but the post-Q3-FY2025 'Opendoor 2.0' team explicitly self-graded against stated AI/automation-driven targets and delivered most judgeable ones (acquisition velocity, AI assessments, Cash Plus, repeatable cohort margins). The flagship promise—AI-enabled ANI breakeven by end-2026—remains too-early but is tracking ahead, so the short track record skews positive.

Significantly increase acquisition velocity from ~128 homes/week via AI-automated underwriting/assessment — promised Q3 FY2025
delivered Velocity rose ~300% to ~537 homes/week by Q4 FY2025 and 5,000+ contracts in Q1 FY2026, best contract quarter since 2022
Scale AI home assessments to ~750/week and cut assessment time from ~a day to ~10 minutes with humans out of the loop — promised Q3 FY2025
delivered By Q1 FY2026 self-assessment app nearly doubled assessed homes vs September, ~half needed zero people on-site, 6,000 seller-led assessments in March
Roll out Cash Plus hybrid (AI/platform-routed capital-light product) to all markets next quarter — promised Q2 FY2025
delivered Reached all markets; grew 600%+ to ~35% of contracts by Q4 FY2025 and ~1/3 of acquisitions by Q1 FY2026
October 2025 cohort's model-driven margin stability (~90bps vs prior ~260bps degradation, ~3x improvement) is repeatable, not a fluke — promised Q4 FY2025
delivered Q1 FY2026 showed four consecutive months (Oct–Jan) of flat cohort margin curves validated at >80% sell-through
AI/velocity-driven path to adjusted-net-income breakeven (12-mo forward) by end of 2026, never raising equity again — promised Q3 FY2025
too-early Reported on track; already adjusted-EBITDA positive on forward-12mo basis as of April 1, 2026, but the ANI breakeven date has not yet arrived
AI-powered pricing models (school-district, competition features) to improve conversion at like-for-like spreads — promised Q1 FY2025
partial Conversion gains reported in subsequent quarters but never quantified against a specific target; superseded by Opendoor 2.0 velocity model
PRICED-IN (REFINED)
LOW (room left)

Est. revisions flat  ·  Fwd P/E -9.2  ·  EV/Sales 0.8x

Estimate signals are not rising: ratings have been mostly stable, with a Hold-heavy mix, and the last-month price target average is below the last-quarter average. Forward estimates show losses continuing and revenue expectations are mixed rather than baking in rapid AI-led growth. Valuation is not stretched on EV/Sales at about 0.8x, while forward P/E is negative because earnings remain loss-making. With flat revisions and a non-rich sales multiple, AI upside does not appear meaningfully priced in.
COVERAGE — ENTHUSIASM TRAJECTORY + CATALYSTS
4Q4 FY20245Q1 FY20258Q2 FY20259Q3 FY20257Q4 FY20258Q1 FY2026

AI enthusiasm across 6 calls — trend ↗ rising

AI evolved from pricing-model tweaks into a central operating system for offers, inspections, automation, velocity, and capital-light products.

RECENT AI CATALYSTS & NEWS
BUSINESS IMPACT - QUALITATIVE MATERIALITY

9/10 qualitative impact   transformational  medium-term · mixed evidence

Where AI matters: pricing, underwriting, renovation, transaction workflow

AI is being applied to Opendoor's core iBuying engine: pricing, inspections, repair negotiation, seller intake, title, mortgage, and internal workflow automation. The hard dollar savings disclosed are small, but the operating evidence points to potentially large effects on conversion, coverage, velocity, fall-through, and per-home renovation economics.

Caveats: Many benefits are pilot metrics or workflow anecdotes rather than reconciled company-scale financial impact; Housing macro, spreads, inventory risk, and capital availability can overwhelm AI efficiency gains; Valuation AI may become less differentiated if competitors access similar models and data; Management enthusiasm is high while consensus revenue still does not reflect a proven AI-led volume ramp

AI DISRUPTION / CANNIBALIZATION RISK  tailwind · 2/10

AI can make home valuation tools more widely available, but it does not automate away Opendoor's core role as a liquidity provider taking housing inventory, capital, renovation, and execution risk. The bigger economic effect is likely improving its own underwriting and transaction throughput rather than deflating the revenue model.

OPTIONS / MARKET STRUCTURE

option liquidity: fair

ATM IV
TYPICAL BID-ASK
OPEN INTEREST

proxy inputs — dollar-ADV $200M · beta 3.55741 · px $5.41

source: proxy (no options chain on FMP)
FMP /stable/ exposes no options-chain endpoint on this key, so ATM IV, bid-ask spread and open interest are unavailable. Liquidity below is a PROXY from dollar-ADV, beta and price level (a stand-in for option depth), not measured option-market data.

CONFIRMATION — INSIDERS · 13F · LANGUAGE
Mixed — insiders selling, institutions adding, management language 6/10 measured.
INSIDERS selling 3 open-market sell(s) vs 1 buy(s) — net distribution
INSTITUTIONS (13F) adding as of 2026-03-31: 72 new / 82 closed positions; 179 increased / 97 reduced; institutional ownership +5.79pp; -13 net 13F holders
MGMT LANGUAGE 6/10 measured Firm ownership around models and data loops, but little explicit AI framing or quantified AI-specific business impact.
commit “Our signal intelligence officers, hedge fund quants and they're all maniacally focused on data loops.”
commit “They have a mandate, ship a change every single week, optimized for both margin and velocity.”
commit “as our models get better and they're getting better every single week, the whole machine moves faster.”
VERBATIM AI QUOTES
“We built an AI audit tool that automatically reconciles inspection scopes with actual repair decisions, giving our field teams real actionable feedback to improve operating compliance and cost discipline.”
— Kasra Nejatian, Q1 FY2026
“At title intake, it used to take us up to 5 hours. It now takes 15 minutes.”
— Kasra Nejatian, Q1 FY2026
“One of our marketing managers replaced our $0.5 million life cycle legacy e-mail system with one Claude skill.”
— Kasra Nejatian, Q1 FY2026
“We're going all in on AI, and we're doing it responsibly.”
— Christy Schwartz, Q1 FY2026
“Some highlights in addition to what Kaz shared earlier: an AI-powered repair negotiation tool cut our buyer fall-through rate by over double digits; field managers are using AI scoping feedback, helping to reduce pre-list renovation spend by up to 10% to 20% per home in pilot markets; and a ticket triage automation, redeployed 3 full-time employees from classification to resolution.”
— Christy Schwartz, Q1 FY2026
“AI extends that by quite a bit because you're now encoding judgment on top of rules and the leverage becomes really high.”
— Kasra Nejatian, Q1 FY2026
“AI doesn't eliminate this complexity. It just makes navigating it a lot easier.”
— Kasra Nejatian, Q1 FY2026
“AI totally dissolves this asymmetry, right?”
— Kasra Nejatian, Q1 FY2026
“We can build AI concierge that feel to the customer like the expert is sitting at the kitchen table, right?”
— Kasra Nejatian, Q1 FY2026
“AI just totally removes this constraint.”
— Kasra Nejatian, Q1 FY2026
“Opendoor 2.0 is the single most AI-pilled company in the public market that I know of.”
— Kasra Nejatian, Q4 FY2025
“Last week, someone at Opendoor shipped an AI workflow that does all of this kind of automatically with no humans in the loop.”
— Kasra Nejatian, Q4 FY2025
“When I say we default to AI, I don't mean engineers use Copilot. That's not what I mean.”
— Kasra Nejatian, Q4 FY2025
“With our self-assessment app, the seller takes pictures of their home and AI does the assessment work. No humans needed.”
— Kasra Nejatian, Q4 FY2025
“Thanks to AI, our product went from being available to about 1 to every 3 homeowners to being available to nearly every homeowner in the Lower 48.”
— Kasra Nejatian, Q4 FY2025
“We also invested in advanced ML models and data-driven pricing strategies. They're improving our margins per home while maintaining or actually improving our conversion rate.”
— Kasra Nejatian, Q4 FY2025
“Our home sale pricing is now powered by our new ML model that avoids Opendoor's previous policy of just blanketed price drops.”
— Kasra Nejatian, Q4 FY2025
“AI is basically like manna from heaven. It's the thing we need to make the business work.”
— Kasra Nejatian, Q4 FY2025
“And so our machine learning pricing model, it has meaningfully improved the precision of pricing decisions, better demand signals, better timing and better calibration at the individual home level rather than blanketed spread reduction or price reductions.”
— Christy Schwartz, Q4 FY2025
ANALYST QUESTIONS ON AI
Q (Q1 FY2026, Mike Alfred): My question is about the longer-term implications of AI. Do you believe when you look at the strategic direction of the company that we are well prepared for all the things that AI is likely to change about the way the real estate market operates in the coming years?
A: AI extends that by quite a bit because you're now encoding judgment on top of rules and the leverage becomes really high. Like that's real and it's important. And we're capturing a lot of this.
Q (Q4 FY2025, Unknown Analyst): Obviously, this has become really pronounced in the market. But what are you guys using it for internally? And how should that change the customer experience in the Opendoor product moving forward?
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, right?