Mr. Cooper Closes, Redfin Attach Hits 40%, Margins Expand to 20% — Maintaining Outperform.

Rating Action

We initiated coverage of Rocket Companies at Hold at Q1 2025 on the view that the Redfin acquisition (announced March, closed July) and the announced Mr. Cooper deal were transformational in scope but unproved in execution, and that the equity required tangible integration evidence and a stable origination backdrop before the synergy story could be underwritten with conviction. We upgraded to Outperform at Q2 2025 when the first-three-weeks Redfin signal (200K prequal clicks, 23% lead conversion, 7K agent referrals, "exceeding expectations" CFO commentary) and the $4B Mr. Cooper bond oversubscription cleared the most binding execution gates, layered on top of an operating beat (adj rev $1.34B above guide, +13% rate-lock volume vs. flat MBA industry view, gain-on-sale held at 280 bps, $80M independent cost-action layer separate from M&A synergies) that demonstrated the standalone Rocket franchise was inflecting independently of the integration story.

Today we are maintaining Outperform. Q3 delivered the three datapoints we wanted to see following the Q2 upgrade. First, Mr. Cooper closed on schedule on October 1, removing the largest unmodeled risk in the story; even more importantly, the first 30 days produced concrete evidence of executable integration — co-branded identity live on day one, 40,000 servicing leads flowing into the Rocket pipeline by day nine, first Mr. Cooper client closed in three days, 400 loan officers onboarded, $400 million of expense synergies with line-of-sight identified. The combined ~10 million-client servicing portfolio and $5 billion of stable recurring annual cash flow now form the cycle-resilient cash-flow base the equity has historically lacked. Second, the Redfin attach rate climbed from 27% to 40% in four months against a 50% target, and 13% of September DTC purchase closings already came from Redfin clients — the funnel synergy is a real, measurable operating tailwind, not a slide-deck aspiration. Third, gain-on-sale held at 280 bps while volume reaccelerated 26% sequentially, demonstrating that Rocket's AI-augmented capacity model captures rate-driven demand surges without surrendering unit economics; the September Pipeline Manager Agent case study (10% conversion lift) is the first quantitative confirmation of the operating-leverage thesis. Combined with EBITDA margin expanding to 20% from 13%, a 63% improvement in production capacity per worker over two years, and a Q4 guide that initiates the fully-consolidated era, every operating thesis point we underwrote at the Q2 upgrade has confirmed or improved. Catalysts to reinforce Outperform: continued Redfin attach climb toward and beyond 50%, Mr. Cooper recapture rate improvement on the combined book, FY2026 guide articulating cycle-resilient revenue durability, FHFA servicing-cap dialogue resolving as expected. Downgrade catalysts: integration-induced client churn surfacing in 2026 cohorts, gain-on-sale margin compression from competitive intensity at the next refi wave, regulatory action on the servicing cap that materially constrains scale.

Results vs. Consensus

Q3 was a clean operating beat on every line that matters in this name — revenue above the high end of guidance, EBITDA margin nearly doubling sequentially, gain-on-sale stable while volume reaccelerated, and management gaining share in both purchase and refinance against a still-difficult mortgage backdrop. The standout is not the absolute level of any one number, but the simultaneous combination: revenue acceleration plus margin expansion plus stable unit economics in a quarter when management was also closing the largest acquisition in industry history.

The shape of the beat matters. Adjusted revenue exceeded the high end of guide while gain-on-sale margin held, meaning the upside came from genuine volume capture rather than price reach. Adjusted EBITDA margin expanded by approximately 700 basis points sequentially, a pace that would be implausible if the Q2 run-rate were the structural baseline — the implication is that Q2's softer profitability reflected market conditions, not deteriorating unit economics, and Q3 reverts to a level that prefigures the operating leverage we expect once the consolidated business is fully scaled. The $124 million GAAP loss looks worse than the underlying business; once the $90 million of one-time costs and $50 million of Redfin purchase-price amortization are stripped out, the underlying operating cadence is materially profitable.

Segment Performance

Direct-to-Consumer (DTC)

The DTC segment is where the Redfin synergy is most visible. Net rate lock volume of $36 billion and closed loan volume of $32 billion both came in at the strongest levels in over three years, with growth split roughly evenly between purchase share gains (Redfin-led) and the September refinance wave (AI-led). The Q3 DTC purchase pipeline reached record levels exiting the quarter and entering Q4. Critically, gain-on-sale margins held at 280 basis points despite volume reaccelerating 26% sequentially — in a typical mortgage upcycle, the marginal locks come at progressively thinner spreads as competitors scale capacity. Rocket's ability to add volume without margin sacrifice is a function of the AI-enabled capacity model: the marginal loan is processed at a lower variable cost, freeing pricing flexibility without compressing the contribution.

Partner Network (Rocket Pro / Broker)

Less granular detail was provided on the call, but the Rocket Pro broker underwriting AI agent — which collapses what was a four-hour underwriting workflow into less than 15 minutes — is functionally a competitive moat in the broker channel. Brokers select among wholesale lenders on a combination of price and service speed; Rocket's offer to brokers is now structurally faster, which should support partner-network share over the next several quarters as broker awareness builds. The AI-enabled wholesale offering is a quietly important second leg of the platform that is not yet getting full credit.

Rocket Money

No material standalone disclosure for Rocket Money in Q3. Management's emphasis on the integrated client experience — ~60 million combined client and prospect relationships across Rocket, Redfin, and Mr. Cooper — treats Rocket Money as part of a unified consumer funnel rather than a discrete reportable segment. This is a deliberate framing choice: as the Mr. Cooper servicing book layers in, Rocket Money's role evolves from an independent personal-finance product into the pre-mortgage and post-closing engagement layer of a vertically integrated homeownership platform. We expect more concrete monetization metrics in 2026 as the consolidated reporting structure stabilizes.

Title

Title services were not broken out individually in the Q3 commentary. The relevant operational story is the purchase-agreement-review AI agent, which compressed a previously 80-step manual workflow by 80% and now delivers accuracy that exceeds the legacy review process — this is title-adjacent work and improves the unit economics of the bundled mortgage-plus-title offering on the DTC purchase funnel. The 150,000 team-member hours saved annually translate directly into a cost-to-close advantage at the integrated transaction level.

Real Estate (Redfin, post-close)

Redfin was consolidated for a full quarter for the first time and is now the most consequential operational engine in the story. The headline numbers: nearly 50 million MAUs flowing into the funnel, the "Get Prequalified" button driving over 500,000 user-started applications in September (more than double the July baseline of approximately 250,000), and the mortgage attach rate climbing from 27% to 40% in four months — ahead of the 50% long-term target and ahead of plan. 13% of Rocket's DTC purchase closings in September were Redfin-sourced, a level that for a company of Rocket's size effectively rewires the purchase funnel. Redfin revenue performed in line with internal expectations; the $140 million in expense synergies are "all identified" and land at full run-rate in Q4. The $60 million 2026 revenue synergy estimate looks conservative given that attach rate is already at 40%; we expect upward revision once 2026 guidance is set.

Servicing (Mr. Cooper, post-close)

Mr. Cooper closed October 1, after the Q3 P&L window, so it does not appear in the Q3 reported financials — but it is the most important segment for the equity's go-forward thesis. Pro forma combined servicing book approaches 10 million clients and generates approximately $5 billion of stable recurring annual cash flow. Mr. Cooper's pre-acquisition cost-to-service was roughly one-third lower than the industry average, a structural advantage now layered onto a larger combined book. The first 30 days of integration are the leading indicator that matters: 40,000 leads flowing into the Rocket origination pipeline by day nine, first Mr. Cooper-to-Rocket client close on day twelve at a 3-day turnaround, 400 loan officers cross-trained, zero client disruption reported. Management's $500 million synergy commitment ($400M expense + $100M revenue) has line-of-sight on the expense piece and emerging traction on the revenue piece via measurable conversion lift on the Mr. Cooper servicing book leads.

Key Topics with Management Quotes

1. The "Category of One" Pitch — Vertical Integration as Cycle Insurance

Varun Krishna built the entire call around a single thesis: Rocket post-acquisitions is a fundamentally different business than Rocket pre-acquisitions, and the new business is structurally cycle-resilient because origination, servicing, and real estate offset each other across the rate cycle. The argument is more than rhetorical — it has empirical backing in the pro forma analysis management ran on the combined entity:

"Even in the toughest mortgage market in decades, a combined Rocket and Mr. Cooper would have delivered positive GAAP earnings every quarter from early 2023, through the third quarter of 2025, on a pro forma basis. This shows our resilience through cycles."
— Brian Brown, CFO

This is the through-the-cycle case the equity has historically lacked. Pre-Mr. Cooper, Rocket was a leveraged origination call — great in low-rate windows, painful in high-rate windows. Post-Mr. Cooper, the $5B of recurring servicing cash flow + an originator that captures rate-cut windows + a large real estate top-of-funnel turns it into a more durable franchise. The Q3 result — revenue beat plus margin expansion plus held unit economics — is a single-quarter proof point of the same dynamic.

2. The September Refi Wave as Operating-Leverage Case Study

The single most informative paragraph in the prepared remarks was the case study of how Rocket captured the late-Q3 rate move. The 30-year fixed rate dropped 40 basis points across the quarter to 6.3%, with a 20-basis-point cliff in the two weeks ahead of the September Fed cut. Rocket's response — and the AI tools that enabled it — is the most concrete evidence yet that the technology investment is paying for itself in real time:

"During the September refinance wave, this agent drove a 9-point jump in client follow-ups and a 10% lift in conversion for both daily credit pools and refinance applications directly increasing locked loan volume."
— Varun Krishna, CEO, on the Pipeline Manager AI agent

A 10% conversion lift on a multi-billion-dollar locked-loan-volume base is a direct, measurable contribution to the quarter's revenue beat. The structural significance is larger than the single-quarter dollars: it provides the first quantitative validation of the "infinite capacity at flat fixed costs" thesis that management has been articulating since the Investor Day. The Q3 EBITDA margin expansion to 20% from 13% sequentially is the macro signature of this micro-level efficiency.

3. The Recapture Math — Why Mr. Cooper Matters Quantitatively

Brian Brown laid out a specific scenario calibrated to the rate environment that quantifies the option value of the combined servicing book:

"If the 30-year fixed rate falls to 5.5%, 25% of our servicing portfolio or about $300 billion in unpaid principal would be positioned to refinance, representing significant growth potential and all at a very low cost of acquisition resulting in strong operating margins."
— Brian Brown, CFO

This is the economic core of the deal. A 5.5% 30-year rate is plausible within the 12–18 month window given current Fed-cut expectations — Fannie Mae's industry forecast cited by management calls for the mortgage market to grow 25% YoY in 2026 with rates potentially dipping below 6%. $300B of refinanceable UPB at low customer-acquisition cost (because these are existing relationships) is a multi-billion-dollar revenue opportunity that, in the pre-Mr. Cooper structure, simply did not exist. Rocket's recapture rate is 3x the industry average; even partial harvesting of that pipeline materially shifts the FY2026 numbers.

4. Redfin Attach Rate Is the Single Best Operating KPI

The 27% → 40% attach rate progression in four months is the most underappreciated number in the print. Varun Krishna walked through the funnel mechanics:

"We've added a prequalification experience and a funnel to every listing on Redfin. So that's millions of access points that represent every single home listing that's on Redfin. And because of that integration, because of the optimization, the number of clients that actually have started applications using that access point that get prequalified button has doubled."
— Varun Krishna, CEO

The bundled Rocket Preferred pricing offer plus the integrated brand is doing the work. Hitting 40% attach in four months against a 50% target establishes a steep glide path that, on extrapolation, lands above the 50% target in 2026 with optionality to push it higher as the refinance funnel is added to the Redfin app and the mortgage origination flow gets pulled forward into the Redfin user experience.

5. Servicing AI — The Sierra Partnership and the Next Leg

An underappreciated piece of strategic disclosure was the partnership with Sierra (the AI-first conversational-agent company) on the servicing side, signaling that the AI thesis is moving from origination-only into servicing operations:

"I would go as far as to say the future of servicing is Agentic AI. And when you think about the use cases in servicing, a lot of it has to do with helping clients solve meaningful problems but also handling simple tasks and automation that drive day-to-day efficiency."
— Varun Krishna, CEO

Mr. Cooper's pre-deal cost-to-service was already approximately one-third below industry average. Layering Agentic AI on top of that base — and applying it to a now-larger 10M-client book — should drive a second leg of cost-to-service compression in 2026–2027 that has not yet been articulated as a synergy line item. This is unmodeled upside.

Guidance

Q4 2025 guidance is the first guide that consolidates Mr. Cooper and a full quarter of Redfin into the reported numbers, so direct comparability to prior-quarter guides is limited. The framing matters more than the headline range:

The Rocket-stand-alone +7% YoY at midpoint is the cleanest read on underlying business momentum, and it is set against a Q4 that historically sees softer mortgage demand around Thanksgiving and the December holidays. Management noted that some industry forecasts have Q4 flat-to-up versus Q3, which would be unusual; Rocket is guiding to a tighter view of seasonality and treating any positive surprise as upside. The Rocket-only Q3 result was up 14% YoY — even at the +7% Q4 stand-alone midpoint, market-share gains continue. The $60M of Redfin revenue synergies in 2026 and the $100M of Mr. Cooper revenue synergies are both on track but largely backloaded into 2026.

Analyst Q&A

Five analysts were on the line. Q&A ran cleanly — no defensive exchanges, no walk-backs — and the substance reinforced rather than undermined the prepared-remarks message. Selected paraphrases:

• Jeff Adelson (Morgan Stanley) pressed on the Q4 guide composition and 2026 outlook. Krishna pointed to the record purchase pipeline entering Q4 and the Fannie Mae industry forecast for +25% market growth in 2026 with rates potentially below 6%; Brown unpacked the Q4 components and noted Rocket-stand-alone Q3 was +14% YoY versus the +7% Q4 guide-midpoint, framing Q4 conservatism as seasonality rather than slowdown.
• Mihir Bhatia (Bank of America) asked about Mr. Cooper synergy timing and the OpEx walk. Brown confirmed line-of-sight to the full $400M expense synergy bucket, walked through the $140M of Q4 one-time costs (vs. $90M in Q3), the $120M of Redfin+Cooper purchase-price amortization (vs. $50M Redfin-only in Q3), and the run-rate $140M unsecured-debt expense going forward.
• Doug Harter (UBS) probed what gets the Redfin attach rate from 40% to the 50% target. Krishna cited the integrated brand, competitive Rocket Preferred pricing bundle, and a roadmap to add the refinance funnel to Redfin and pull mortgage origination forward into the Redfin app. Brown added context on funnel dynamics: high intent, extended time-to-purchase, high-quality leads in pipeline rather than "fall-out."
• Bose George (KBW) asked about Investor Day market-share targets and whether Mr. Cooper's share is incremental. Krishna walked through the three building blocks of purchase-share growth (Redfin top-of-funnel, AI-enabled conversion, servicing-book recapture). Brown indicated revised share targets will be communicated in coming quarters once integration is bedded in — current focus is execution, not target-setting.
• Terry Ma (Barclays) raised the FHFA 20% servicing cap. Brown was direct: caps are not unusual in the industry, the GSE/FHFA dialogue has been productive, capital and liquidity levels are well above required standards, and the current agreement leaves sufficient room to grow and even exceed synergy targets. He framed it as a non-issue.
• Ryan McKeveny (Zelman) asked about the AI strategy on the now-larger servicing book. Krishna disclosed the Sierra partnership and described servicing as the next leg of the AI thesis (24/7 conversational agents, automated escalations, predictive client engagement) — a forward-looking commitment that is not yet quantified in synergy estimates.
• Mark DeVries (Deutsche Bank) asked about the September AI/refi case study. Brown emphasized that the AI gains were as much on the loan-officer/banker side as on the back-office processing side — digital chat experiences and automated document collection allow human bankers to focus on revenue conversations, increasing both capacity and efficiency simultaneously.

The most informative read of the Q&A is what was not asked: there were no questions on credit quality of the servicing book, no questions about gain-on-sale sustainability, no challenges on the synergy quantification. The buy-side has effectively underwritten the integration thesis at this point; the open questions are about acceleration and operating leverage, not solvency or execution risk.

What They're NOT Saying

• No 2026 quantitative guide. Management's enthusiasm about 2026 was unmistakable (Krishna: "we feel very excited about 2026"), but specific dollar guidance was deferred. This is normal Q3 cadence — full-year guides typically come at Q4 or year-end — but the absence is worth noting given the level of integration progress already achieved. Expect a step-up in disclosure at Q4.
• No granular Rocket Money disclosure. Once a featured standalone product line, Rocket Money was effectively absent from prepared remarks and Q&A. Management's framing of "60 million combined client relationships" subsumes Rocket Money into the consolidated funnel narrative. This is either a deliberate de-emphasis (Rocket Money's standalone economics are less compelling than the integrated funnel), or a deliberate timing choice (full integration disclosure waits for the FY2026 reporting structure to settle).
• FHFA dialogue addressed indirectly. The 20% servicing-cap question got a confident "not something we're worried about" but the actual cap level, the specific KPIs being monitored, and the room-to-grow magnitude were left general. Investor Day disclosure should resolve this.
• No revised market-share targets. Bose George's question on whether Mr. Cooper is incremental to the prior Investor Day targets was deferred to future quarters. The answer is almost certainly "yes, materially," but management is keeping powder dry for a more impactful disclosure window.
• No quantification of the Sierra-on-servicing synergy. The future-of-servicing-is-Agentic-AI commentary was qualitative. Given Mr. Cooper's already industry-leading cost-to-service, layering AI agents represents real upside that has not been priced into either the synergy plan or the FY2026 sell-side models.
• Capital-return prioritization is unaddressed. Pro forma cash of approximately $4 billion and total liquidity of approximately $11 billion is substantial, and the $4B unsecured-notes issuance was used for debt refinancing rather than capital return. The capital-allocation hierarchy post-integration — debt paydown, MSR purchases, dividend, buyback — was not explicitly walked through.

Market Reaction

The print landed at the close of October 30 with the call held the same afternoon. Headline reaction in the immediate trading window was constructive: the adjusted-revenue beat plus the materially better-than-Q2 EBITDA margin plus the on-schedule Mr. Cooper close gave the bull case an unambiguous endorsement, while the GAAP loss of $124 million was widely (and correctly) discounted as the mechanical residue of one-time integration costs and Redfin purchase-price amortization. The Q4 guide of $2.10B–$2.30B set against a stand-alone Rocket midpoint of +7% YoY landed in line with the buy-side consolidated-quarter framework, removing a key uncertainty around how the new business model would print in its first fully-consolidated period.

The most important market signal is what did not happen: there was no de-rating event on integration risk, no concern over the Mr. Cooper financing cost, no negative reaction to the $90M of Q3 one-time costs or the $140M Q4 one-time guide. The market is awarding Rocket the credibility benefit-of-the-doubt on the integration story, which we read as an accurate assessment of the operational evidence in the print.

Street Perspective

The bull case being made on the Street centers on three operational data points that this print decisively confirmed: (1) the Redfin attach-rate trajectory is ahead of plan and structural, (2) Mr. Cooper integration risk is materially lower than initially feared given the visible day-1-through-day-30 execution evidence, and (3) Rocket's AI capacity model is a real cost-leverage engine, not a slide-deck construct. Bull-case price-target revisions in the wake of this print are likely to focus on FY2026 revenue synergy realization (the conservative case looks too conservative) and 2026–2027 EBITDA margin compounding once the consolidated cost structure is fully harvested.

The bear case being made on the Street has narrowed materially. The remaining concerns center on (1) FHFA servicing-cap uncertainty as a regulatory overhang despite management's confidence, (2) the tail risk of integration-induced client attrition surfacing in 2026 cohorts that we cannot yet observe, (3) gain-on-sale margin compression at the next major refi wave if competitor capacity catches up, and (4) the duration of integration one-time costs — the $140M Q4 one-time guide is acceptable but the cadence into Q1 and beyond is not yet articulated. None of these alone supports a Hold rating in our framework, and they collectively read as risk-management residual rather than thesis-breakers.

Sentiment among generalist investors has clearly shifted from "transformational acquisitions, prove it" (our Q1 and Q2 stance) to "transformational acquisitions, getting proven." Our upgrade is timed to the inflection where the proof is visible in the print rather than to the moment when the synergies are fully realized in the P&L — we view that gap (between operating-evidence emergence and full-P&L realization) as the alpha window.

Model Implications

• Q4 2025 revenue: midpoint of $2.20B as guided; we model the upper half of the range ($2.20B–$2.30B) given record purchase pipeline entering Q4 and Mr. Cooper recapture momentum already visible in October.
• Q4 2025 EBITDA: implied margin compression sequentially given Q4 mortgage seasonality and a consolidated cost base still absorbing one-time costs; we model 14–16% adjusted EBITDA margin in Q4, with sequential expansion resuming in Q1 2026 as one-time costs roll off.
• FY2026 revenue: we now model double-digit YoY growth on a consolidated basis with upside from (a) Redfin attach climbing past 50%, (b) Mr. Cooper revenue synergies hitting run-rate, (c) any rate-environment tailwind from a 5.5%-handle 30-year mortgage scenario.
• FY2026 EBITDA margin: we model 22–24% as the consolidated run-rate, with upside if AI-driven capacity gains continue to outpace fixed-cost growth and Mr. Cooper cost-to-service compresses further with Sierra-AI deployment.
• Synergies: $140M Redfin expense synergy treated as fully baked; $400M Mr. Cooper expense synergy phased in over 6 quarters; $100M Mr. Cooper revenue synergy treated as low-end (we model upside to ~$130M); $60M Redfin revenue synergy in 2026 treated as low-end given current 40% attach trajectory.
• Capital structure: total combined corporate debt unchanged post-refinancing per management; $11B total liquidity gives substantial flexibility; we do not model material near-term buyback given debt-paydown priority but flag this as future capital-return optionality.
• Servicing book MSR mark sensitivity: at a $5B annual cash flow run-rate on a ~10M-client book, MSR-fair-value movements remain a quarterly noise factor; we treat this as below-the-line for thesis purposes but a real GAAP-EPS variability driver.

Thesis Scorecard

Seven of nine thesis pillars confirmed in Q3, with the remaining two pending purely on disclosure cadence rather than execution. The integration risk that gated our Q1 and Q2 Hold rating is materially resolved.