管理層發言
Ladies and gentlemen, thank you for standing by. My name is Krista, and I'll be your conference operator today. At this time, I would like to welcome you to the Better Home & Finance Holding Company Fourth Quarter and Full Year 2025 Results Conference Call. Operator instructions. I would now like to turn the conference over to Tarek Afifi, Corporate Finance and Investor Relations Manager. Please go ahead.
Welcome to Better Home & Finance Holding Company's Fourth Quarter and Full Year 2025 Earnings Conference Call. My name is Tarek Afifi on Better's Corporate Finance team. Joining me on today's call are Vishal Garg, Founder and Chief Executive Officer of Better; and Loveen Advani, Chief Financial Officer of Better. In addition to this conference call, please direct your attention to our fourth quarter and full year earnings release, which is available on our Investor Relations website. Also available on our website is an investor presentation. Certain statements we make today may constitute forward-looking statements within the meaning of federal securities laws that are based on current expectations and assumptions. These expectations and assumptions are subject to risks, uncertainties and other factors as discussed further in our SEC filings that could cause our actual results to differ materially from our historical results. We assume no responsibility to update forward-looking statements other than as required by law. During today's discussion, management will discuss certain non-GAAP financial measures, which we believe are relevant in assessing the company's financial performance. These non-GAAP financial measures should not be considered replacements for and should be read together with our GAAP results. These non-GAAP financial measures are reconciled to GAAP financial measures in today's earnings release and investor presentation, both of which are available on the Investor Relations section of Better's website and when filed in our annual report on Form 10-K filed with the SEC. More information as of and for the period ended December 31, 2025, will be provided upon filing our annual report on Form 10-K with the SEC. I will now turn the call over to Vishal.
Thank you, Tarek. Good morning, everyone, and welcome to our fourth quarter and full year 2025 earnings call. Before I begin, I'd like to give a warm welcome to our new Chief Financial Officer, Loveen Advani. Loveen is a seasoned strategic and operational finance leader with a strong track record of guiding companies through growth and transformation. He has repeatedly demonstrated the ability to align strategy, capital allocation and execution. His experience and leadership style will be instrumental as we execute our strategic and financial priorities in our next chapter of anticipated growth. What's more, I love him because he gets his hands dirty and his hands on keyboard. When I first met him, he sent me over a model, and we started spending time on it one-on-one late at night. That is the kind of CFO that this company needs for the next stage of its Blitzscale growth, and we are so, so happy to have Loveen on board with us. Better is a vertical AI platform fundamentally reshaping and revolutionizing the home finance industry. We are building the AI-native frontier of consumer finance and in doing so, enabling players with massive customer bases to provide mortgages and HELOCs in an AI-first way to their customers, while empowering the established network of local retail mortgage originators. Adoption across the ecosystem confirms this shift is real and accelerating. This is the power of the Tinman AI platform. Over the past decade, we have built a first-of-its-kind AI-driven matching engine that connects consumer credit data, income data, asset data and property data with the preferences of roughly 40 different investors on our platform, allowing us to approve mortgages and home equity loans nearly instantly. The result is a process that is faster, cheaper, easier and just plain better. We are in the middle of a genuine transformation from what was once a direct-to-consumer mortgage business serving consumers who came to Better.com to an AI-native mortgage platform serving the entire mortgage industry. Over the past decade, we built the technology, the infrastructure and the investor relationships to manufacture mortgages faster and cheaper than anyone else. Today, we're taking that foundation and extending it across the entire ecosystem, powering partners with massive customer bases and enabling local retail brokers and originators to scale in ways that simply were not possible before. That shift is now showing in our results and in the momentum we are building with our enterprise partners. These are large complex partnerships with longer sales and setup cycles than anything we manage in our D2C business, and growing them is not something we do alone. It requires deep collaboration with our partners at every step from integration and onboarding to conversion, optimization and product expansion. The pace of ramp is a shared journey, and we are working hand-in-hand with each of our partners to get things scaling. The progress we are seeing is real. The early data is highly encouraging, and we are more excited than ever about what lies ahead. Let me walk you through what we are seeing across each of our key partnerships. As you know, we launched the largest platform partnership in Better's history with Intuit Credit Karma, a leading personal financial services company serving more than 40 million monthly active users. Last year alone, Intuit Credit Karma processed 47 million tax returns and reached over 140 million members. In fact, more than 80% of Americans who took out a mortgage last year are members on the Intuit Credit Karma platform. Through this partnership, we are integrating the breadth and depth of Credit Karma's member data, including credit, income and home attributes such as full credit bureaus, tax returns and detailed home valuations directly into the Tinman AI platform. As you might remember from our public announcement, Credit Karma's goal is to save its members $1 trillion in interest savings on their mortgages. This is no small task, as it implies that our collective partnership, which is saving consumers about $25,000 of lifetime interest on average since we launched in October 2025, needs to fund 40 million mortgages to achieve Credit Karma's goal. In October 2025, after over nine months of working together, we went live on the Credit Karma app and since have rapidly ramped and have only penetrated less than 1% of their monthly user base that we believe is eligible for the product. The opportunity is massive, and our primary focus is deepening integration of the Tinman AI platform across the various Credit Karma consumer touch points to better serve the full needs of its entire member base. Also through our Tinman AI platform, we continue to make great progress extending our platform to power local retail mortgage lenders, providing them with the infrastructure to build and scale their businesses on top of our technology. We continue to scale NEO with their local loan officer teams across the United States experiencing rapid growth. Here, Better enables retail mortgage lenders to build their business on the Tinman platform with near zero customer acquisition cost on this channel. It's been incredible to see the NEO team grow their business from the $1.5 billion run rate they had when they joined to the $2.4 billion run rate they ended 2025 with on the Tinman AI platform. It's proven that the Tinman AI platform eliminates friction, giving originators the opportunity to scale responsibly with 28 new loan officer teams onboarded onto the platform in 2025. Within six months of fully rolling out, NEO increased funded loans per mortgage adviser by 91%, per processor by 17% and per underwriter by nearly 50%. Retail mortgage teams around the country are taking notice of these enhancements and are leaving their existing platforms to join the Better platform and to embark on our shared journey of making retail home lending cheaper, faster, easier and just plain better. Next, our top five U.S. nonbank mortgage loan originator partner went live this February with just 2% of its loan officers on the Tinman AI platform. And in the coming months, we are working towards expanding to all 3,000-plus loan officers. Early reports indicate superior loan officer experience for users of Tinman versus the prior implementation on their legacy software stack. As this rollout scales to their full loan officer base, we expect this partnership to be transformative for both organizations, adding a significant platform volume opportunity for Better while giving one of the largest mortgage originators in the country a competitive advantage in how they serve their customers. In addition, Finance of America, which is an industry-leading reverse mortgage lender with access to millions of customers who are typically home equity-rich but cash flow disadvantaged, is in its early stages of ramping. Together, we are launching the first HELOC and HE loan product offerings to their customers powered by our Tinman AI. We have high hopes of being able to reach a population that Better has traditionally not reached—the senior market—with our partnership with Finance of America and expect to see significant results from that partnership in the coming quarters ahead. And finally, we announced a major milestone, the launch of the first conversational credit decision engine for mortgages and home equity loans integrated directly into ChatGPT through our Tinman AI app. Loan officers, banks and fintechs can now receive decision-ready credit outputs in as little as 47 seconds, reducing origination timelines by an average of 21 days. Better is the only application authorized to display credit decisions within ChatGPT, powered by our proprietary MCP technology built on top of Tinman. Tinman can instantly underwrite approximately 95% of mortgage and home equity loan types, and any institution with a ChatGPT enterprise license can deploy it; no traditional aggregators, no markups. This opens a significant new distribution channel and a clear path to expanding into a direct-to-consumer channel over time. As you might remember, OpenAI and ChatGPT have over 800 million users globally and over 80 million users in the United States with that number growing rapidly. We believe this is the third version of the Internet, and we are first to market with a clear differentiated offering from the other folks that have launched apps on OpenAI and ChatGPT and with the ability to not provide a marketplace or provide a solution, which then requires consumers to leave the platform, but actually to provide a solution that enables consumers to fulfill the entire transaction directly within their ChatGPT interface. Since our OpenAI announcement, we have seen a massive immediate response from across the financial services industry. Within days of releasing a short demonstration video last week, we've received inbound interest from over 40 financial institutions, mortgage companies, banks and fintechs, all reaching out at the most senior levels to request a demo and work with us on deploying our ChatGPT application. As an example, a bank CEO in the South reached out after seeing the announcement. They want to grow their mortgage business, but not the way they tried before through hiring large teams, building out fixed infrastructure and taking on the operational burden that comes with it. What resonated with them was the simplicity of the ChatGPT app and the idea that any loan officer in any branch can instantly qualify a consumer for a mortgage through a conversational interface, minimal setup time, minimal training, maximum reach. This is exactly the problem we set out to solve. The mortgage industry has long been trapped in a cyclical model, scaling up headcount in good markets and cutting in bad ones with fixed costs that punish originators when volumes decline. Tinman fundamentally changes that dynamic. The infrastructure we have built and proven with our current partners can be deployed for any bank, fintech or local originator team. We are giving institutions the flexibility to grow their mortgage business without the operational burden that has historically made that growth so difficult to sustain. We have two strategies when it comes to go-to-market on the Tinman AI platform. The first is to own the future with partnerships like the ones we have done with Credit Karma and OpenAI, where we are developing new ways to reach tens of millions of consumers that are substantially easier and faster for consumers to use and leveraging our technology to create a customer experience and value proposition moat that no one else in the industry can match. The second is to bring the path forward, which is what we have done with NEO and Finance of America and the top five mortgage originator. Better is the mechanism by which these local market experts and large existing mortgage originators with deep relationships can continue to serve both their customers and referral partners. With Better's partnership, NEO is becoming one of the fastest-growing retail lenders in the country. The people didn't change. The relationships didn't change, only the tech platform did. I'll now touch on our financial highlights, and Loveen will provide greater detail shortly. In the fourth quarter of 2025, we generated $1.5 billion in funded loan volume and $44 million in revenue, representing year-over-year increases of 56% in loan volume and 77% in revenue, respectively. This growth spanned all three of our core product categories: refinance, purchase and HELOC. Our Tinman AI platform generated $646 million in volume in the fourth quarter, representing over 40% of total volume and surpassing our prior guidance of $600 million. This outperformance reflects the demand and growing confidence of our partners in our platform. While the fourth quarter is always seasonally softer, our growth year-over-year outperformed that of the industry average, which was relatively stagnant. According to MBA data, in the fourth quarter, total residential funded loan volume increased by 4% year-on-year. Compared to Better's funded loan volume, which grew 56% over the same period. For the full year 2025, we delivered $4.7 billion in funded loan volume and $165 million in revenue, up 32% and 52% year-over-year, respectively. We achieved this growth despite an approximately $1 billion headwind from the conclusion of our Ally partnership, a testament to the resilience of our model. We remain on track to reach $1 billion in monthly volume by May 2026 and to reach adjusted EBITDA breakeven by the end of the third quarter 2026. To win in a commoditized market, you have to win on three things: customer acquisition cost, operational cost and cost of capital, what we call the three pillars of competitive advantage. On customer acquisition, our model inverts the traditional origination dynamic. Rather than paying for customers in an open market, our partnerships are structured so that customers are brought directly to us. Credit Karma's over 140 million members, NEO's 70 local branches and 140 mortgage advisers and our top five nonbank originator partnering with over 3,000 local mortgage advisers represent embedded distribution at scale, a structural CAC advantage that competitors find extraordinarily difficult to replicate and one that is not easy to sustain without a technological moat. On operational costs, Tinman automates up to 80% of the repetitive loan production tasks and our Betsy tool resolves underwriting issues instantly by pulling loan facts, guidelines and drafting communications in seconds. The result is a platform that scales production through AI efficiency and growth without additional overhead. Our cost to process, underwrite and close a loan, and this we're talking about mortgage loans and HELOCs combined together, is about $800 a loan, which is far less than anyone else in the industry. We believe that the initial launch, our Home Token, will allow us to book an extra $500 per funded loan in revenue. And as we scale that, we believe long term, we're going to be able to achieve significant gains in loan revenue as well as funding cost to the consumer and interest rate to the consumer, which we believe will translate into a significant competitive advantage and moat as a result of the efforts that we have put in. On cost of capital, we continue to improve our warehouse terms while working to expand capacity to support partnership volume growth. In parallel, we are working towards a secured tokenized credit facility via the stablecoin ecosystem that we estimate could lower funding costs by up to 100 basis points once implemented, a structural funding advantage that would be difficult for any traditional mortgage originator to match. Over the past three years, we have built the foundation for this moment. I can tell you, honestly, the last time I felt this excited about Better's future was in March 2021. And we have line of sight once again into growing into the largest mortgage company in America. This is truly a turnaround that we have worked for years to bring to life and one that has been able to be built on the implementation of AI across our entire business and leveraging the Tinman platform that we started working on back in 2014. This is why we think that the moat that we have is more sustainable than the traditional incumbent. We built an end-to-end system that takes eight different systems in the mortgage industry and pulls them all together into one system so that it's not just the rules that are captured, but all of the context around the human decisions on the data and the rules. And that learning data across $110 billion of loans is what allows us to continue to push forward and lower our cost to produce, improve our conversion rate and build for our partners that are building the future. And we believe that we can continue to do this because the competitive advantage of richer learning data only compounds over time, the more transactions and the more partners you bring into the ecosystem. We are now firmly in our next phase of growth with momentum, scale and a clear path to adjusted EBITDA breakeven. Partnerships are expanding, adoption is rising, our platform is proven and our AI capabilities are best-in-class, and we are just getting started. With that, I'll turn it over to Loveen to provide a detailed walk-through of our financials.
Thank you, Vishal. I'm pleased to join Better at such a pivotal moment. The company's differentiated platform positions it as a leader in AI-powered home finance. I look forward to partnering with Vishal and the team to drive disciplined execution, enhance financial performance and create value for shareholders. As Vishal outlined, we're in the midst of a meaningful strategic transformation, shifting from a direct-to-consumer originator to an AI-native platform powering the broader mortgage ecosystem. From a financial perspective, this transition is significant. Enterprise partnerships of this scale carry longer ramp timelines, but they also carry a far greater volume potential and far better marginal economics than our legacy D2C model. What gives me confidence is that the financial trajectory is already beginning to reflect this shift. Our platform partnerships are growing rapidly and contributing an increasingly meaningful share of our overall business. To put that evolution in concrete terms, in 2024, our total funded volume was $3.6 billion with 0% contribution from Tinman's AI platform partnerships. In 2025, we grew total funded loan volume to $4.7 billion with 35% coming from our Tinman AI platform. Looking ahead to 2026, we see a clear path to over 60% of our loan volume coming from our Tinman AI platform business. This is a fundamental reshaping of our revenue mix and a reflection of how we're executing on this transition. Let me now review our fourth quarter and full year 2025 financials. Better continues to generate opportunities independent of the broader economic and mortgage market conditions. With a large addressable market and less than 1% share today, we have demonstrated the ability to grow regardless of macro conditions. Starting with the fourth quarter of 2025, compared to Q4 2024, funded loan volume grew 56% to approximately $1.5 billion. Revenue increased 77% to approximately $44 million. This growth was primarily driven by funding more loans through our Tinman AI platform partnerships. Looking at loan volume by product, refinance grew to 8%, purchase increased 22% and home equity rose 18%. By channel, 44% came through Tinman AI platform partners and 56% through direct-to-consumer. By product mix, 49% was purchase, 37% was refinance and 14% was home equity. For full year 2025, compared to full year 2024, funded loan volume grew 32% to approximately $4.7 billion. Revenue increased 52% to approximately $165 million. These results were driven by the launch of our Tinman AI partnerships and continued growth in our direct-to-consumer business. By product, refinance increased 119%, home equity grew 78% and purchase rose 14%. By channel, 36% came through Tinman AI platform partners, 62% through direct-to-consumer and the remaining 2% from our former Ally partnership. By product mix, 61% was purchase, 21% was refinance and 18% was home equity. Turning to cost efficiency. In Q4, total net revenue grew 77% year-over-year, while expenses remained approximately flat. This demonstrates clear operating leverage. We are scaling the revenue at lower marginal costs driven by efficiencies from Tinman AI platform. We continue to streamline overhead while ensuring sufficient resources to support new partnerships. We expect these partnerships to contribute meaningful growth through 2026 and beyond. Unit economics in our direct-to-consumer channel continue to improve. We have integrated AI across every part of our sales and operations workflow. Per loan contribution margin improved 28% quarter-over-quarter from approximately $1,800 to approximately $2,300 per loan. We continue to expect reducing origination costs through higher conversion, lower customer acquisition costs and improved labor efficiency. In the fourth quarter, our adjusted EBITDA loss was approximately $24 million. That compares to a $28 million loss in Q4 of last year and a $25 million loss in the prior sequential quarter. While we aim to reduce losses further on a sequential basis, we're constantly evaluating expense discipline versus investing in growth opportunities. The continued ramp of our business with positive marginal economics is accelerating our path to adjusted EBITDA breakeven. We believe we are at an important transition point, moving from a primarily direct-to-consumer fintech to a true AI platform for the mortgage industry. This gives us confidence in our expectation to achieve adjusted EBITDA breakeven by the end of Q3 2026. Now a brief update on our balance sheet and capital positioning. We ended Q4 2025 with $227 million in cash, restricted cash, short-term investments and assets held for sale. We maintain strong relationships with our financing counterparties with three warehouse facilities totaling $575 million in capacity as of December 31, 2025. We appreciate our warehouse lenders' continued support as we deploy Tinman AI across the mortgage ecosystem. Turning to our outlook. For our total loan volume, we expect $1.4 billion to $1.55 billion in Q1 2026, of which the midpoint is a 70% year-over-year growth from Q1 '25. Based on how our partners are ramping, we continue to believe that we will reach a $1 billion total monthly loan volume by May 2026. We expect to achieve adjusted EBITDA breakeven by the end of Q3 2026. This will be driven by volume growth across both our Tinman AI platform and direct-to-consumer channels, per loan contribution margin improvement, pricing gains and corporate cost reductions. I would note that these growth opportunities have varying expansion timelines, so progress towards breakeven may not be linear. With that, I'll turn it back to the operator for Q&A.
分析師問答
Operator instructions. Your first question comes from the line of Ramsey El-Assal with Cantor Fitzgerald.
I wanted to ask about guidance. Your guide assumes that the Q1 loan volume is roughly flat, I think, at the midpoint versus Q4. Obviously, you have a lot of exciting things going on in the company. Just wondering if you could walk us through the drivers. The partnership volume grew nicely versus Q4 quarter-to-date. Does that mean you're expecting flatter growth on the direct side of things? Or what are the drivers we should consider?
Ramsey, it's Loveen. Thanks for the question. So it's flat because of seasonality. If you go to Page 17 of our investor deck, we made that point and we've shown the last six quarters. So if you look at Q4 '24 to Q1 '25, it was down, right? And this year, from Q4 '25 to our guidance of Q1 '26, it's flat or slightly up. That just shows the kind of growth in the platform.
Got it. Okay. And a quick follow-up for me. I wanted to ask about profitability. Your current target, obviously, is to reach adjusted EBITDA profitability by the end of Q3 this year. How should we think about—how are your thoughts evolving on medium-term and longer-term profitability, especially kind of in the context of this accelerating shift towards the partnership model? How should we think about your profit profile going forward?
Yes, absolutely. Look, I just started a month back. The first task is to get to profitability by Q3 2026, right? After that, we'll evaluate our growth opportunities along with incremental positive contribution margin. When we evaluate new partnerships, we'll be thinking about a contribution margin in the range of 10% to 15% to as high as 25% to 30%. We'll be looking at that range as we think about our growth opportunities.
I think there's three different buckets of the product. The first bucket of the product is what we do on D2C. The second bucket of the product is what we do on Tinman AI platform, where we're closing the loans in our own name. That's what we're doing with NEO. That's what we were doing with Credit Karma. That's what we're doing with others. And then the third is what's the margin on the business where the lender is closing in their name. That's what we're doing with Finance of America. That's what we're doing with the top five mortgage lender, with the top three fintech. All of those—those lenders are closing in their name, and we're giving them the platform to do it. It's their salespeople, processors, underwriters and the closers in our software. And so each of those has a different margin profile and a different revenue per loan profile, depending on the amount of work that we are doing in that. On D2C, of course, we're doing everything from customer acquisition to sales to processing, underwriting, closing and investor marketplace. In some partnerships, we're doing everything; in others, we're doing only parts. Sometimes we're doing capital markets, sometimes we're not. And so it just depends on that—what the revenue per loan is going to be and what the margin is going to be. As the revenue per loan kind of comes down, the margin actually expands because it becomes more and more where they're just using the platform. So the platform-alone business can be 60% margin. The D2C business, as you can see from a contribution margin perspective, is a 20% plus margin business on a contribution margin basis. So we're going to get to know that and define that. We feel very confident in the guidance we're giving and particularly the growth that we're manifesting. But I think at those types of growth rates, you can't exactly know what people are going to buy. And we're in the transformation phase of the business. So we'll know more over the coming couple of quarters.
Your next question comes from the line of Kartik Mehta with Northcoast Research.
Vishal, the partnership metrics suggest some massive top-of-funnel demand. And I'm wondering what kind of metrics you're seeing from preapprovals to a funded loan and how kind of that underpins getting to the $1 billion target?
Yes. So Kartik, if you think about it in the context of our D2C business that we've previously disclosed, that ends up being around 5%. So if the partner volume starts coming in on a cohort basis, let's assume I get $1 billion of pre-approvals, I end up funding about 5% of them. And that funding can take place over three, four, five, six months as it bakes because some people aren't fully ready. They come back, they need to get their spouse to agree. All these different things that happen with this fairly significant life-stake financing transaction for consumers. Remember, on average, 32% of their income is going towards us. So it's a major transaction. There's a bunch of things that go back and forth between when we approve them to when we are able to actually realize the funding event for that. But on a cohort basis, it bakes to 5%. Now in some partners, it ends up being higher because those partners have better brand or deeper matching or deeper integration. In other partners, it ends up being a little lower. We're going to see that play itself out.
And Vishal, where are you in the process from the stablecoin ecosystem use for funding? Obviously, you talked about that, lowering the funding costs and it seems very interesting. So I'm just wondering where you are in that process?
I think we're six months away from when it starts to hit the bottom line.
Your next question comes from the line of Brendan McCarthy with Sidoti.
Welcome, Loveen. I just wanted to start on the Credit Karma partnership. At this point, does that span all of your mortgage products? Or is it strictly geared towards refi?
Right now, we have started with refi, and we believe we will then launch HELOC and then from there, purchase.
Understood. That's helpful. And I think looking at the addressable market there, 140 million obviously is a lot larger than original expectations. Maybe just over the long term, what do you think is a reasonable penetration rate to drive volume?
In the long term, we expect Credit Karma Home Loans, powered by Better, to be the single largest originator of mortgages in this country.
Understood. That's great. Transitioning to the expectation for breakeven adjusted EBITDA at the end of Q3. I assume that will kind of coincide with the $1 billion in monthly funded loan volume. Can you break down your expectations there for volume contribution from D2C, NEO and then Credit Karma as well?
Yes. As I said in my prepared remarks, the Tinman AI platform contribution was 0% in 2024. It was about 35% in 2025, and we're expecting about 60% of total volume from that platform, which includes Credit Karma, NEO and other partnerships.
Understood. And turning to fourth quarter results, just looking at the gain on sale margin, I think it declined sequentially just by a little bit here. I assume was that mostly just given the higher refinance D2C growth?
Yes.
And then last question for me. I saw in the slide deck, it sounds like there's a top three personal lending fintech in the pipeline. I think you mentioned it's currently in the pilot phase. Any detail you can give on that? Is that going to be geared toward more the Tinman mortgage software partnership side? Or do you think it will be similar to NEO or Credit Karma where you'll be doing the originating?
We think in the beginning, it's going to be similar to Credit Karma where we're doing the originating. Then this fintech also has a pretty prominent bank, and so they may choose to onboard to their balance sheet. I've said publicly the bank capital regulation requirements are going to dramatically change the mortgage landscape. The number of calls we have had from banks post the launch of the ChatGPT app—we have over 45 financial institutions in the United States and outside the United States that have interest in utilizing that platform. I thought it was going to be mortgage brokers and retail mortgage lenders; the number of banks that have called, in anticipation of what is happening, has surprised me. I'll double-click into this. If you are a midsized bank today and you've got disintermediation from stablecoins, you can't just go and get internet deposits cheaply anymore. Your deposit cost of capital is creeping up. You've got to go find assets. When you've got to go find assets that generate a higher yield, you can't just sit there and put it in treasuries and short-duration instruments because then your cost structure just doesn't allow you to compete with stablecoin. So what is the thing that you can do with economic growth slowing, with the American consumer a little bit stretched? Are you going to go long credit cards? Are you going to go long personal loans? Or now with mortgage regulatory reform and particularly bank capital levels, are you going to go long credit risk or long duration? Banks are built to go long duration. We're going to see the bank bid for mortgages explode. The bank bid for HELOCs will explode. We are uniquely positioned to accommodate the bank bid vis-à-vis our competitors in HELOC land. Our competitors in HELOC land have built a one-size-fits-all approach proud of it like a box for securitization. We tell any bank: you bring your guidelines, you bring your regional preferences, you bring any of those, and we will accommodate those instantly and to as detailed as you want. I think you'll see a lot of that and a lot more partnerships in that regard. Also, these fintechs are all signing up for bank charters, and they see it, too. You'll see the lines blur between fintech and fintech bank. I thought it might be worthwhile to just share a little bit of context around that, particularly in light of today's announcement around bank capital rules.
Your next question comes from the line of Eric Hagen with BTIG.
Really good conversation here. Really appreciated your thoughts just now on the bank capital. You noted the cost to underwrite is substantially lower than the industry average—around $800, if I heard you correctly. Why don't you think those savings are being passed on to borrowers? What's the gating factor, which is sort of like bottlenecking the ability to pass along those savings in your opinion?
We are passing the savings on our side to the borrowers. We are passing on the savings while trying to continue to improve our contribution margin on our path to profitability because we want to be able to keep passing those savings on to borrowers for many years to come. Our rates are about 30 basis points cheaper on average than the average mortgage rate. Our rates are over 50 basis points cheaper than Rocket and loanDepot. The customer in a purchase market is really guided by the local realtor and the local loan officer. You haven't seen that shift yet. But as refi comes back, you remember from 2016 to 2021, we went from $500 million of volume to $58 billion of volume as refi ramped. As rates come down and refi comes back, there's serious scale possibility because in refi, we have a clear winning proposition. The American consumer may not differentiate between small basis-point differences, but if you tell them, 'Do you want to save $422 a month versus $375 a month?' that math is easy to do. On the B2B side, that savings is direct and visible. The average bank cost to produce is about $14,500. When we go to these banks and say we'll do it for you—for example, flavor A for $4,000 a loan, flavor B for $5,000 a loan, flavor C for $6,000 a loan—that's a very disruptive sale. Now it takes time because that's a CEO/CFO sale. If you're pitching AI implementation to a mid-level person in the bank, it's disruptive. Tinman lowers the amount of rote work that exists in the mortgage industry. Much of the mortgage industry is still 'stare and compare' underwriting. Call up any of my competitors and ask the loan officers how those loans are underwritten and they will tell you. That's just a fact.
Your next question comes from the line of Rohit Kulkarni with ROTH Capital.
A couple of questions on the Tinman AI platform to help us understand what the ramp looks like based on what you know right now—40% of volume already in Q4. Where do you see that share go as the year progresses? And then based on a lot of these recent developments, what are the gating factors for you to scale up that distribution for Tinman? Is there some technical integration, some regulatory compliance approvals, partner training? Walk us through what it would take for you to convert all the leads that you have on Tinman and how that cycles into the overall proportion of funded volume.
Rohit, thanks for the question. I'll take the first and then hand it over to Vishal for the second. The trend is: Tinman AI platform was 0% of our revenue in 2024. Last year, on a full-year basis in 2025, it was about 35% of our revenue. This year, we're saying we expect it to be around 60% of our revenue. We're not giving full-year guidance on loan volumes here, but if you read the tea leaves and look at the trends, our guidance for the first quarter loan volumes is about 77% growth, and if you extrapolate—and this isn't formal guidance—the share of Tinman AI platform increasing from 35% to about 60% shows that subsection is growing really fast.
Our large institutional partnerships—where the companies are 10x to 100x our size—typically have these timelines: from first demo to term sheet signed is usually three months; from term sheet signed to platform launch is usually two months after that; from platform launch to pilot done is about 90 days from that; and then post-pilot to full institutional buy-in and penetration of their customer base takes around nine to 12 months because there are many connections to make. The good thing is after it's connected, it's one system. With what we've done with ChatGPT, we're trying to shorten that sales cycle because it's an interface their internal people already know. That can cut the nine-month timeline down to six months, or even three months if they move fast and don't have legacy constraints, which is why many non-mortgage companies are entering the mortgage business through us. For organizations already entrenched in mortgage with large incumbent infrastructure, it takes longer. The bottleneck previously was our Biz Dev team; last quarter we had two people, and now there are five. We want to ensure revenue aligns with cost. We don't want to hire a large go-to-market team before revenue materializes. Our goal is to monetize our lead while aligning expenses and revenue together.
Okay. Great. Specifically on Credit Karma, perhaps talk about how Credit Karma is helping amplify the benefits and perhaps improve distribution visibility in their member base. What is the dual handshake that once you are deeply embedded in a fintech partner like Credit Karma, how does that change the way they promote or provide higher visibility to your offering?
Credit Karma is a very advanced company. They have a system called Lightbox, and we have integrated ourselves into Lightbox. The system is determining those offers. Right now, we're at less than 1% penetration of their member base as of March 13. We're very excited about the future.
Okay. Great. One last one: earlier you covered contribution margin or marginal margin on D2C versus partnerships. How does that compare right now and over time? Is that an implied assumption within your EBITDA breakeven in the second half?
We're not sharing that level of granular detail for competitive reasons right now. The partner contribution profit per loan varies depending on the extent of services and resources used. Our adjusted EBITDA breakeven is based on achieving the penetration rates on the partners we have signed up.
Your next question comes from the line of Ryan Tomasello with KBW.
Just another question on the Tinman AI platform. There's obviously a range of different models out there in the market that are also providing this broad tech infrastructure to support origination and funding in the mortgage category. That includes some players building that on blockchain rails. Vishal, can you talk about broadly what differentiates Better in this third-party infrastructure category from those peers? And then over time, do you think this platform could be extensible into other categories of consumer credit outside of the mortgage and HELOC market?
Tinman is the only platform in its class that allows loans to be sold to a wide network of investors who can bring their own guidelines and their own pricing into the platform. Some competitors offer integration but keep product criteria tightly controlled, limiting flexibility. We let you use any title company, appraisal company, home insurance company, and accommodate complex scenarios—HOAs, jumbos, non-QM, bank statement loans, DSCR loans—so you can serve the broadest possible customer base. That flexibility matters to partners who want to serve customers on their terms. Many incumbents are still billing by the seat and force customers into multiple disconnected systems. We're billing by the outcome. Digital mortgage has often increased the cost to produce loans because of multiple systems, middleware and consultants. Tinman's agentic AI architecture removes that. You don't need to train people across eight systems; the machine learns from human operations and automates the tasks. There will still be human decisions required for regulatory reasons, but those humans will be able to make far more decisions per day because the machine handles the routine work. That is the future we're driving toward and where we are unique.
Appreciate all that commentary, Vishal. One more on the stablecoin partnership: can you quantify the cost of capital advantage that funding source provides versus traditional facilities? How might that partnership evolve beyond warehouse into permanent financing? And bigger picture, what value do you envision DeFi unlocking for the mortgage market over time?
I'll keep it simple. The initial funding cost advantage is about 100 basis points, which is very meaningful. We think that could translate to roughly $500 extra per loan. Mortgage is underpenetrated among stablecoin issuers. As stablecoins become more pervasive, issuers seeking broader yield will look at DeFi mortgage assets. Mortgages that we make—many guaranteed by GSEs or agencies—are attractive from a Sharpe ratio perspective. A Fannie Mae mortgage combines government guarantee with a tangible asset and borrower, which is compelling. Over time, DeFi can lower spread premiums and funding costs. Currently the premium to hold a fixed-rate GSE mortgage over a 10-year treasury is about 200 basis points; we think DeFi could help get that down to approximately 100 basis points over time.
Your next question comes from the line of Owen Rickert with Northland Capital Markets.
First for me, to go from $1.5 billion in volume to $3 billion in volume per quarter, what needs to happen?
We need to penetrate our existing partners more, and we need to continue to grow NEO and D2C where it makes sense and where we make money on those D2C loans. To go from $1.5 billion to $3 billion is primarily about deeper penetration of the existing partners we already have signed up and implemented with.
Okay. Great. For the four ramping partnerships, can you just rank those in terms of opportunity? We know Credit Karma is obviously #1, but how would you rank Finance of America, the top five nonbank originator and that bank partner?
I think it's Credit Karma Home Loans powered by Better as the largest opportunity, then the top five nonbank originator, then Finance of America, and then the top three leading fintech.
Lastly from me, beyond those four partners, do you have the bandwidth to get potential partners five, six and seven live in 2026? Or is 2026 more just about ramping those four?
No, you should see us launch one marquee partner every quarter, and you should see a bunch of smaller partners launch every quarter.
That concludes our question-and-answer session. I would now like to turn the conference back over to Vishal Garg, Founder and CEO, for closing comments.
Thank you, everyone, for joining. Q4 2025 is a transformational turnaround quarter for the business as we move from being a direct-to-consumer originator on Better.com to being a platform to power every originator in the mortgage industry. We thank you for your interest, and thank you for being a participant and a partner in our journey to making that happen and in doing so, making home finance cheaper, faster and easier and just plain better for all Americans. Thank you.
Ladies and gentlemen, this does conclude today's conference call. Thank you for your participation, and you may now disconnect.