管理層發言
Good afternoon, ladies and gentlemen, and thank you for standing by. My name is Kelvin, and I will be your conference operator today. At this time, I would like to welcome everyone to the Neptune Insurance Holdings Fourth Quarter Earnings Call. I would now like to turn the call over to Jon Carlon, Director of Corporate Development. Please go ahead.
Thank you, and good afternoon. With me here today is Trevor Burgess, Chairman and CEO; Matt Duffy, President and Chief Risk Officer; and Jim Steiner, CFO and COO. Before we begin, I'd like to remind everyone that today's discussion will include forward-looking statements, including, among others, statements about our expectations for our future financial performance, growth opportunities, business strategy, market trends and capital allocation plans. These statements are based on our current views and assumptions and are subject to risks and uncertainties that could cause actual results to differ materially. We direct you to our recent SEC filings for a full description of these risks. We undertake no obligation to update any forward-looking statements, whether as a result of new information, future events or otherwise, except as required by law. We will also reference certain non-GAAP financial measures. These measures should be considered only as supplements to their comparable GAAP measures. Additional information, including reconciliations of the non-GAAP measures to their most comparable GAAP measures can be found in our earnings release at investors.neptuneflood.com and in our current report on Form 8-K that was publicly filed with the SEC on February 18, 2026. And now I'd like to turn the call over to Trevor.
Good evening, and thank you for joining us. Let me start with what's on everyone's mind. Last week, our stock sold off seemingly because investors decided AI is coming for insurance. They're right. They just have the direction of the trade completely wrong. Neptune isn't threatened by AI disruption. We are AI disruption. We have been since the day we wrote our first policy. We have no legacy system to rip out. We have no army of manual underwriters to replace. We built this company as an AI-powered API-first platform from scratch, proprietary machine learning models running pricing, underwriting, portfolio management and distribution across the largest private flood data set in the United States. Let me be specific about what that means, the data moat. Over years of underwriting and servicing hundreds of thousands of policies, we have assembled proprietary flood risk data, behavioral retention data, claims performance insights and real-time transaction signals that no competitor can replicate by simply plugging in a large language model.
In an AI-driven world, the model is a commodity. The data is the moat. We have the data. The results. Our underwriting performance hasn't just been good. It has been so good that capacity providers are fighting to get on our platform. We grew from 23 to 40 capacity partners this year, a 74% increase, now supporting 8 programs. And because those partners have seen hundreds of millions of dollars in underwriting profit flow through Neptune, our average commission rate increased 35 basis points year-over-year. That's not us negotiating harder. That's the market telling us our AI works. Where we sit? Before I explain why AI is a tailwind for us, it's worth clarifying how Neptune actually reaches the end customer because I think some of the confusion starts here. We are a managing general agent. We sit between the capacity providers who take the risk and the distribution channel that reach the policyholder.
We do not own large captive agency forces. We reach customers 2 ways, through a network of independent agents and wholesale partners and through direct-to-consumer digital channels. Both flow through the same platform. That distinction matters because when AI reduces the cost of distribution, that savings flows directly to Neptune's bottom line. We are not disrupting ourselves. The distribution upside. Here's the part that I think is most misunderstood. Agent commissions are our single largest expense line. If AI-driven workflows reduce friction in how consumers buy flood insurance and they will, that is not a threat to Neptune. That is margin expansion for Neptune. We already support both agent-led and fully digital flows on the same platform. If every customer in America decides tomorrow that they want to buy flood insurance through an AI agent instead of a human one, we are ready. And our adjusted EBITDA margin goes from an already exceptional 60% to something significantly higher.
The team. I'd like everyone to look at who actually works here because it tells you what kind of company this is. Over 40% of Neptune's employees are engineers or data scientists. This is not an insurance company experimenting with AI. This is an AI company that happens to be in insurance. And that composition shows up in the numbers. In 2025, trailing 12-month revenue per employee increased 15% to $2.7 million and adjusted EBITDA per employee increased 14% to $1.6 million. For context, our revenue per employee puts us between Apple and NVIDIA. These are not insurance company metrics. Those are elite technology platform metrics made possible because of AI. I want to be honest about something. Nobody knows exactly how AI will reshape this industry. We could be wrong about the pace, the path or the specifics. But if the question is who is best positioned to adapt, we like our answer. We are already building with the latest tools.
We are already operating at the cutting edge and our entire infrastructure was designed to evolve. If the future belongs to AI, and we believe it does, we would rather be the company that was built for it than the one trying to catch up. Technology, capacity, distribution. Those are the 3 moats around this company, and we expect every one of them to get deeper as AI adoption accelerates. Now let me talk you through the quarter. The fourth quarter was an outstanding finish to a record year for Neptune. It showcased the stability of our platform, the strength of our execution and the durability of our business model. During the quarter, we successfully launched new capacity programs, delivered record new business sales and scaled our technology seamlessly to meet elevated demand, all while maintaining strong margins and operational discipline. Our first full quarter as a public company built directly on the momentum that brought us here and capped an exceptional year.
A few highlights from the quarter include revenue of $43.8 million, a 39% increase year-over-year, net income of $4.3 million at a 10% margin, down 63% from the fourth quarter of 2024, with adjusted net income of $15.3 million, up 25% from the fourth quarter of 2024. Adjusted EBITDA of $25.9 million, up 34% year-over-year at a 59% margin, written premium of $100.3 million, driving 33% year-over-year premium in force growth and record new business sales posted during the quarter. The strength of the fourth quarter capped an exciting and record-setting year. For the full year 2025, Neptune delivered revenue of $159.6 million, up 34% from 2024. Net income of $37.4 million at a 23% margin, up 8% from 2024 and adjusted net income of $56.9 million, up 38%. Full year adjusted EBITDA of $95 million, up 32% year-over-year at a 60% margin, giving us a Rule of 40 of 93. Year-end premium in force was approximately $370 million, reflecting over $90 million of net growth during the year.
And as a reminder, because this is central to understanding the business, Neptune takes no balance sheet insurance risk, 0. We are an asset-light platform that earns commissions and fees on every policy written and renewed. That model is what allows us to scale at 60% EBITDA margins without taking catastrophe exposure onto our books. Now profitability and technology get a lot of the attention, and they should. But I wanted to spend a moment on something that doesn't always show up in the model, how we performed when it mattered most. During the fourth quarter, the federal government shut down and the National Flood Insurance Program or NFIP went dark. That is the primary source of flood insurance for most Americans, and it was unavailable. While the NFIP was shut down, our platform kept quoting, kept binding and kept onboarding new agents who needed somewhere to send their customers. We made targeted disciplined decisions to support our agent partners during the disruption, including incentives that contributed to record new business sales with minimal impact to margins.
That is the kind of moment that cements relationships. Agents remember who showed up. I also want to highlight a structural advantage of our market that I think is underappreciated. Flood insurance pricing is not subject to the reinsurance cycle volatility that whipsaws other property and casualty lines. The NFIP sets its rate statutorily. And as the dominant market alternative, that creates a stable pricing environment around which we can underwrite with confidence. In 2025, we retained 98% of premium. That retention rate tells you 2 things. Our pricing is competitive and our policyholders are staying. The results we delivered this year reflect a model that is working, technology, data, capacity, distribution and execution all compounding together. We enter 2026 with real momentum and confidence in our ability to continue building value long term. I'll now turn things over to Matt Duffy, our President and Chief Risk Officer, to discuss business updates.
Let me walk through how we scaled our 3 core pillars during the fourth quarter. Our technology platform handled elevated volume without strain during the federal government shutdown. Increased activity drove strong growth in sign-ups, quotes and bound policies. Performance, reliability and service levels remain constant throughout, highlighting the best-in-class fully digital nature of our platform. We also launched the beta version of our Indemnity Earthquake product. Because our underwriting and distribution infrastructure is entirely digital and API-driven, adding adjacent perils requires minimal incremental cost. We can expand into new lines without adding complexity or sacrificing margin. The company was built this way by design. At the same time, ongoing enhancements to our machine learning models continue to improve performance across both new business and renewals, demonstrating the compounding nature of our data advantage.
We are seeing real productivity gains across our engineering organization from AI-assisted development tools. Engineers are shipping code faster, iterating models more quickly and accelerating testing and documentation cycles while maintaining strict quality and security standards. Nearly half of our workforce is our engineering and data science team. When development velocity increases, innovation accelerates and importantly, operating leverage improves. That advantage compounds over time. Our platform was built from inception for real-time digital insurance transactions. If conversational or AI-driven interfaces increasingly become the entry point for insurance purchases, we are ready and integrated. We do not need to retrofit legacy systems or rebuild workflows. If more volume ultimately shifts towards direct-to-consumer channels, our economics benefit. At the same time, we continue investing in tools that make agents more productive.
We are channel flexible by design. On to capacity. Our capacity relationships are a direct validation of 8 years of underwriting performance. It's important to clarify something that is often misunderstood. Neptune does not file rates like admitted insurers. Our pricing algorithms and risk selection models are proprietary. We also do not retain insurance risk on our balance sheet. That means we scale without catastrophe loss volatility or capital strength. Our revenue is driven by recurring premium retention and operating leverage, not by loss ratio swings. That structure allows us to grow capacity efficiently, as shown during the fourth quarter when we launched our seventh capacity program with Palomar, which contributed immediately to financial results. We also executed our eighth program with Somers Re, which went live on January 1. That program was designed to deploy additional capacity quickly as opportunities arise.
Q4 also saw us expand an existing capacity relationship to support the beta launch of our indemnity earthquake product. We now partner with 40 risk-taking providers, up 74% year-over-year. Capacity expansion is earned. It reflects underwriting performance, disciplined portfolio management and transparent reporting. Our partners continue to expand alongside us because the results support it. Subsequent to quarter end, we completed the renewal of another significant capacity agreement. The continued renewal and expansion of these capacity relationships reinforce our ability to scale efficiently across varying market conditions while maintaining underwriting discipline. And distribution. The strength of our distribution platform was particularly evident in the fourth quarter. We delivered record-breaking new business sales despite the absence of storm-driven demand, underscoring the point that our growth is driven by execution and market expansion rather than event-driven activity.
Production associated with the Palomar relationship contributed incremental volume on top of an already record-setting quarter. We launched a new user-based log-in process during the quarter and more than 30,000 insurance agents registered within the first 30 days. This significantly improves how agents interact with our platform and is already deepening the behavioral data set that delivers better tools, insights and feedback through our data science and AI models. For the full year, premium retention reached 98%, up 90 basis points year-over-year, supporting record written premium of $367 million. Those figures reflect the durability of our agent relationships and the consistency of our value proposition. Across technology, capacity and distribution, the progress achieved in the fourth quarter strengthens our operating leverage, reinforces our competitive moat and compounds our data advantage. We enter 2026 with significant momentum. With that, I'll pass it on to Jim.
For the fourth quarter of 2025, Neptune delivered another strong financial performance, capping off a very successful year. Quarterly revenue increased 39% year-over-year to $43.8 million, driven by record new business sales and strong retention across the portfolio. For the full year, revenue increased 34% to $159.6 million, reflecting continued growth in premium in force and consistent execution across the business. We've continued to focus on renewal acceptance as a key driver of revenue retention. For the full year ended December 31, 2025, we retained 98% of premium and 86% of eligible policies, up 0.9 and 1.8 percentage points, respectively, compared to 2024. Our asset-light technology-first model continues to deliver operating efficiency and strong margins even as we incur incremental costs associated with being a public company. For the 3 months ended December 31, 2025, net income was $4.3 million, down 63% over the prior year quarter, while adjusted EBITDA rose 34% to $25.9 million.
Net income included $4.6 million of IPO-related expenses incurred during the quarter. This resulted in a 9.9% net income margin and a 59% adjusted EBITDA margin for the quarter. For the full year of 2025, net income increased 8.2% year-over-year to $37.4 million, representing a 23% net income margin, while adjusted EBITDA increased 32% to $95 million, representing a 60% adjusted EBITDA margin. 2025 adjusted EBITDA primarily excludes IPO expenses and share-based compensation. On a per employee basis, we generated $2.7 million of revenue per employee and $1.6 million of adjusted EBITDA per employee for the full year, increases of 15% and 14%, respectively, compared to the prior year. The continued improvement in these metrics underscores the scalability of our model as we grow. Turning to the balance sheet. Our growth and strong operating cash flow allowed us to further strengthen our capital structure.
During the quarter, we refinanced our existing term debt into a $260 million revolving credit facility, which lowered our interest rate, eliminated required amortization and provides greater flexibility to manage capital efficiently. We ended the quarter with $240 million of total debt outstanding or approximately 2.5x net leverage on a trailing 12-month adjusted EBITDA basis. Subsequent to quarter end, we paid down an additional $9 million of debt during January and February, bringing our current outstanding balance to $231 million. Our priority remains continued debt reduction while maintaining flexibility to invest in growth. IPO-related expenses totaled approximately $8.9 million for the full year, including $0.5 million incurred during the fourth quarter. These costs were reimbursed following the completion of the IPO on October 2, and that reimbursement was reflected as an equity contribution in our Q4 financials.
The proceeds from reimbursement were used to further reduce outstanding debt. Share-based compensation totaled $11.4 million for the year with $11.1 million recognized during the fourth quarter. This increase was driven by 2 IPO-related items. First, upon completion of the IPO, outstanding unvested employee stock options became subject to accelerated vesting, resulting in a onetime noncash compensation expense of $4.1 million. Second, as part of the IPO, Neptune issued restricted stock units to all employees. These RSUs vest annually in equal installments over 3 years. On a go-forward basis, we expect approximately $6.9 million of RSU-related expense quarterly. Based on Neptune's year-end stock price and calculated using the treasury stock method, we had 149,401,852 diluted shares outstanding. The net share settlement to satisfy tax withholding obligations will reduce that total share count by approximately 500,000 shares or 35 basis points. Upon vesting, the net sale of RSUs will have a 64 basis point dilutive effect on the account of basic shares outstanding. With that, I'll turn things back to Trevor for closing remarks.
Neptune remains focused on long-term shareholder value creation. The inherent variability of government policy and weather-related activity make short-term forecasting challenging. And as a result, we generally do not update guidance unless there has been a meaningful change in the underlying trajectory of the business. The strength of our performance in the fourth quarter and across full year 2025 has increased our confidence in the outlook for 2026. Based on that performance, we are increasing our full year expectations. For full year 2026, we now expect revenue of approximately $193 million and adjusted EBITDA margin between 60% and 61%. These targets reflect our continued commitment to profitable growth, operational efficiency and disciplined capital allocation. Where appropriate, we intend to deploy capital to grow the business while returning excess capital to shareholders. To date, that has included a strong emphasis on debt reduction as a straightforward and efficient way to enhance equity value.
Before we open the line for questions, I want to bring it back to what matters. This quarter capped a record year. Revenue grew 39%, adjusted EBITDA grew 34%, and we did it at a 59% adjusted EBITDA margin while investing for growth. Those are the numbers of a company that is AI. I want to be direct about what we're building because I think the market narrative has drifted from the reality. Neptune is not just a successful flood MGA. We are building the technology and data infrastructure for flood insurance in an AI-driven world. That is a much larger opportunity, and we have a meaningful head start. Our data sets get stronger every day. Every policy quoted, every renewal decision, every claim resolved feeds back into our models. That flywheel has been compounding for years. You cannot shortcut it with a foundation model into press release. Data at our scale in our domain is a structural barrier to entry, and we believe it widens with time.
Our underwriting results have made that case to the people who matter most, the capacity providers, putting real capital behind our models. We went from 23 to 40 partners this year. Commission economics improved. That is not momentum you can hand wave away. That is sophisticated risk capital voting with its balance sheet that Neptune works. And our distribution infrastructure is built for wherever the market goes next, agents, embedded digital, AI-native workflows. We don't need to guess which channel wins. We capture all of them on the same platform. But I want to be clear about something. We are not rooting against human agents. There are roughly 25 million buildings in the United States at meaningful risk of flooding and only about 4 million flood insurance policies in force. That is a massive protection gap. AI doesn't close the gap alone. Agents do, armed with better tools, better data and a faster quoting experience.
Our job is to make agents more productive, not to replace them. When we do that well, the addressable market gets dramatically larger for everyone. And if parts of the market do shift towards fully automated buying, our margins expand because our largest cost line compresses. Either way, Neptune wins. So here's how I frame it simply. We have the technology, we have the data, we have the capacity relationships. We have the distribution flexibility, and we just posted a record year, proving all 4 work together. The question is not whether AI disrupts insurance. The question is, who is best positioned when it does. We believe, and our results demonstrate that answer is Neptune. Thank you to our employees, our over 30,000 totally awesome human insurance agents, our capacity partners and our shareholders. We'll now open the line for questions.
分析師問答
Your first question comes from John Barnidge of Piper Sandler. We have the technology, we have the data, we have the capacity relationships, and we have distribution flexibility, and we just posted a record year, which proves all four work together. The question is not whether AI disrupts insurance. The question is who is best positioned when it does. We believe, and our results demonstrate, that answer is Neptune. Thank you to our employees, our over 30,000 totally awesome human insurance agents, our capacity partners and our shareholders. We'll now open the line for questions.
Congrats on the results and appreciate the opportunity. I believe you had talked about earthquake launching. How large of a market do you think this is for the company in '26?
Thank you for asking that question. We are in beta testing right now. And what that means for us is getting the technology right, making sure that we have the right product that meets the marketplace needs. And we know that in California, which will be our first state, 90% of homes are not protected by earthquake coverage. So we think much like flood, it's a hugely underpenetrated market. Too early to say how big that will be for us in 2026.
My other question, if I may. It looks like there was a letter Elizabeth Warren put out shortly before the close today, attempting to connect flood insured with Project 2025. Can you maybe talk about your response to that and maybe build, Trevor, on how you are the AI disruption as opposed to being disrupted?
If this were Polymarket and you had asked, I would have given odds of the likelihood that the Trump administration would invite me to the White House and that I would get a letter from Elizabeth Warren at something like 10,000 to 1. But I think it's a tremendous opportunity for us to continue to talk with important stakeholders about the role of private flood insurance. We're entering our 10th year in business. We are the largest private flood insurer in the United States. We want to make sure that all stakeholders understand the role that private flood insurance can play to close the huge flood insurance gap that exists in the country. We have talked with everyone who is interested in talking with us, from senator staffs to the White House to the FEMA Advisory Council to our local congressperson. This is a bipartisan concern to make sure that people can be properly protected. We look forward to talking to Senator Warren's staff to make sure she understands the role that private flood insurance can play. Listen, when it comes to that, I'm sorry.
Sorry, I didn't realize you're continuing.
Yes. No, I was just going to answer your second question about sort of where we fit in the AI ecosystem. Really, I think the most important thing to understand is that Neptune's platform was built from the beginning using old-fashioned AI. So non-large language model developed AI, really the automation of human tasks. So we've been built from the beginning able to work in digital workflows. We've been able to work in agent-led workflows, direct-to-consumer. If all Americans decide tomorrow, they're going to use ChatGPT to buy flood insurance, we're ready for that, and that would be good for us because it would expand our margins. But I don't think that's actually going to happen. Agents play an incredibly important role here at expanding the marketplace. So there are about 25 million buildings in America that are at moderate to severe risk of flooding, and there are only about 4 million flood insurance policies that exist today. We need agents to be beating the drums, educating their customers that they need to add flood insurance to their coverage universe.
Your next question comes from the line of Andrew Kligerman of TD Cowen.
I was curious about the comment that your priority or emphasis would be on debt reduction, and could you give me a better sense of why that would be the priority given that in the last two weeks what I think was simply, and as you explained Trevor, a real overreaction to AI on your stock caused it to fall too far. So, given where the stock is, is there any consideration of share repurchase? I believe it was something you were exploring during the IPO period. Bottom line, what are your priorities on capital management and why?
Yes. Thank you for that question. We've been paying down debt because it's the easiest thing to do. As a revolver, that gives us availability to use in the future if need be. We find great ways to grow the business or invest, or ultimately, if the Board decides that stock repurchases are the right thing to do, we'll have the dry powder to be able to do that. We also produce significant free cash flow, which will grow that amount over time. As I mentioned on our last call, we now have three options available to us: we can pay down debt, we can buy back shares, and we can potentially someday pay dividends. I think our Board will be interested in the first two for this year because it gives us the most flexibility. We will have to see what the ultimate reaction is to our earnings and how we help educate the market about our stock. But we too were surprised that year to date we're down more than the traditional brokers, which is confusing to us.
It seemed like a real overreaction. Kind of following up on John's question about Senator Warren's letter. Related to that, she seemed to favor NFIP over the private market. And I appreciate that you want to educate the Senator staff. But I was kind of hopeful just given that it just seems like the private markets are so much more efficient. I've been kind of hopeful that the NFIP could be resolved in the form in which it currently exists. Do you, Trevor, think that there's a good probability of that at this point, given the support that we're seeing in the Senate?
I think there's a lot of education that still needs to happen in Congress. And in particular, helping everyone understand that private flood insurance can save the majority of Americans' money on their flood insurance, right? That's a really important point if we're trying to get more people insured, if we're trying to close this massive flood insurance gap, which I've talked about a couple of times on this call. We've got some education work to do. Now the decision around what's going to happen with FEMA and the NFIP seems to lie with the administration. And I don't have any insight into what the decision-making process is there. And we, like everyone, look forward to hopeful report from the FEMA Advisory Council, which was extended and may still produce a report, we'll see. But our job is just to do the blocking and tackling every day to try to sell more policies. The NFIP has been in long-term decline, and Neptune is growing rapidly. And our job is to keep Neptune growing, and there's not much I can do to control what happens at the federal government level.
Your next question comes from the line of Bob Wong of Morgan Stanley.
So my first question is really about how to think about the future evolution of AI, right? So I think you gave some tremendous comments on why AI is an advantage that is built for Neptune and Neptune is built for AI. But as we go forward as technology evolves, as you gather more data, things of that nature, can you maybe help us think about, is there a scenario where potentially other competitors that can perhaps replicate your capabilities or perhaps comes in from a different angle that can, for lack of a better word, give you a run for your money. But just really curious as to how you think about the evolution of the technology going forward.
Yes. I think there are a couple of things here. First, there's a massive total addressable market: 100 million buildings in America. Eighty percent of the sales Neptune has ever done are non-mandatory, so there's a very large market. Second, we have a big data moat. The data we have collected and continue to collect and analyze gives us a tremendous advantage. Even if future large language models know everything on the internet, that data doesn't exist on the internet, so our dataset is unique. I'm open to the possibility that advancements could be made, but there are two other important moats. One is the capacity relationship. Someone needs to take that risk, and we've made hundreds of millions of dollars for our risk-taking partners over the last eight or nine years. This is a relationship business; we've worked very hard to cultivate those relationships and continue to grow them profitably for our partners.
We work with 11 of the 12 largest reinsurers in the world; they would have to decide to leave us and the money we've made them over time to go to a new startup. I'm not saying that's impossible, just unlikely. The last area is distribution. We've worked for years to build the distribution network we have with amazing agents across the country, and it took years to build all of the API connections and integrations. So there are not one or two, but three meaningful moats here that will be hard to replicate.
Okay. Really helpful. Maybe on the second point is the path of growth. You've addressed some of this on the prepared remarks. But if we think about the current quarter, right, the revenue was stellar. One thing I would like to maybe get a little bit more detail on is how you think about the path for that growth going forward. Is there any reason to think that you cannot repeat the same level of growth you just had in the last, call it, 2 quarters or 3 quarters going forward? It feels like the momentum is on your side, so to speak.
Yes. We're certainly excited about the momentum in the business, and that's why we've increased our revenue guidance for 2026. We have a couple of things going on at the same time here. One is increased momentum. Two is the Palomar rollover, which will run through October 1 of this year, so we will not have that effect in the fourth quarter of 2026. And we're larger, which means that even if 86% or 87% of people renew their policies, you're still losing 13%, and you need to make those up just to stay the same. So it gets harder and harder the larger you get. Put all of that together, and we are confident that things look better than they did three months ago, and that's why we increased our guidance.
Your next question comes from the line of Mike Zaremski of BMO.
Just a question around any updates on the task force that was created by the President. And if you all have seen any plans, time line kind of et cetera, around that initiative.
Thanks, Mike. The only thing that we heard is the same thing that everyone in America heard, which was that it was extended by 60 days, and that puts it to sometime end of March, and we'll stand by and see what happens. But we have no other information than that the extension took place.
Got it. So there's nothing you've seen that would cause you to change the way you operate so you can react if something meaningful happens when that date arrives.
We're ready in case something happens, but I don't know what, if anything, will occur. We added our eighth program on January 1 with Somers Re. Somers is backed by a very large global reinsurer. We put that kind of program in place so we can react incredibly quickly if needed. There could also be another government shutdown. The NFIP has now been extended through September. There were multiple shutdowns in the first Trump administration, and we've already had two so far in this administration that affected the NFIP. We need to be ready to serve America if that happens again this fall.
Great. That's helpful. My final follow-up is on the stronger-than-expected revenues, you gave us some good color on the puts and takes. Did you all tease out kind of how much of a bump you got from the shutdown? And separately, are the promotions you called out, are those kind of normal course of business or those were more onetime due to the shutdown?
I think the promotions were more one-time. Our December expense for agents returned to normal levels, as they have in January and so far in February, but we do promotions from time to time. We thought the shutdown was a good opportunity to lean in and take a very minor hit to margin to introduce a lot of new agents to Neptune, who had previously largely only used the NFIP. We were excited about onboarding them. When we last spoke in the middle of December or in November, which I think was the day the major shutdown ended, I likened the impact to a medium-sized storm. With the aftereffects of agents discovering Neptune for the first time and sticking with us because they liked it, maybe it was a medium to large-sized storm. Still, it was not as large as what we would see from a major category 4 or 5 hurricane hitting a major metropolitan area. But the shutdown did have the very positive effect of introducing a large number of agents to us for the first time, and that's part of why the added enthusiasm and trend led us to increase our guidance.
The next question comes from the line of Tommy McJoynt of KBW.
If we looked at the last few years and also just thinking about the next few years ahead, is most of Neptune's growth taking share from the NFIP? Or is it writing new policies that previously had no flood insurance?
It's something like 75% or 80% net new. very few people are actually switching from the NFIP to Neptune. And that's because there's still about half of all policies are receiving subsidies and their risk rating 2.0 price has not yet been achieved. As you remember, the NFIP is increasing prices on consumers' homes by 18% a year until they reach their actuarially accurate rate at the NFIP. So with the ongoing subsidization, there really hasn't even been the opportunity for people to think about a private alternative. So that, I think, will change over time as those subsidizations burn off in the coming 3, 4, 5 years. But for now, the vast majority has been net new, new home purchases, people discovering from their agents that their homeowners policy doesn't cover the risk of flooding and the need to protect themselves from this most dangerous peril. The housing market has not done us any favors, right? The housing market continues to be quite slow. When people buy houses is when they normally shop for insurance. So a return to a healthy housing market would definitely help our business as well because a lot of those people may have NFIP policies, they sell their home. Now that's a new opportunity for someone to shop.
And actually along the same line there, sort of mechanically, when a home does sell that's currently under, we'll call it, risk rating 1.0 pricing with the NFIP, does the new buyer of that home get the full risk-adjusted risk rating 2.0 price even if it's significantly higher? Or are they also grandfathered into the 18% increased cap?
They are able to be grandfathered into the glide path through a manual process of taking over the old policy. Not everyone does that because it is a bureaucratic process, but it is possible to do. Otherwise, they do have to pay the new Risk Rating 2.0 price. But a lot of home sales also are taking place with people who have paid off their mortgage and so are no longer required to have flood insurance. And then when they sell, they sell to someone who does have a mortgage and then that's going to trigger the requirement for a new policy to be bought. And that's a great opportunity for Neptune. So that's part of the reason why we're at near record levels in the United States of cash ownership of homes. And so that churn creates new mortgages, creates new mandatory buying.
Your next question comes from the line of Gregory Peters of Raymond James.
I was going to ask you, I was looking at Page 13 of your investor deck. And one of the comments you made was actually the last bullet point. And I think, Matt, you talked a little bit about it during your comments where you said you launched a new user-based agent log-in system. And you said within 30 days of the launch, more than 30,000 agents created individual accounts. And my human brain is having some problems processing how you had such a broad acceptance of this. Was that bots registering individual accounts? Talk to me about what actually happened that triggered such a strong response rate.
No, it wasn't bots. Sorry, Greg, and thank you for that question. The answer is no, it wasn't bots. This is all human-based and uses multi-factor authentication, so users are required to properly authenticate into the system. This reflects how many users we actually have using Triton, the quoting platform, and Poseidon, the policy management platform on a monthly basis. These are agents who are coming in to quote policies, to buy policies, and also agents who are servicing existing policies. We saw a great uptick, and this gives us a big new data set going into 2026 that provides much more granular behavioral data at the agent level. That will allow us to better help agents as they go through this process, show them ways to sell policies and improve penetration of flood insurance in their books, and also reward agents for the quoting and binding activity they do through the Neptune system.
Okay. I guess a follow-up question relates back to comments where you said you now have 8 programs and 40 capacity providers. Can you remind me what the difference is between program 1 and program 8, or between program 4 and program 7? I'm just trying to visualize the differences between the programs.
The programs are all actually the same. They've all signed up to Neptune and to Triton to allocate individual policies to them. We were awarded a patent last year on our disaggregation system that does that allocation, assigning individual policies to the carrier where it will have the least impact on their probable maximum loss. It's a very sophisticated program. Put simply, we don't want one carrier to have all the policies in Miami; we want to spread that out among the eight different programs. We want to make sure that if a big storm happens, no one carrier is disproportionately damaged because of it. Doing that in a highly sophisticated way to manage their risk capital is a big part of the value add of Neptune to these carrier partners. But they're all exactly the same rule set. They are not setting any pricing and there are no different guidelines. They're signing up to Neptune and basically giving us a premium they're looking for during the year.
Said another way, they submit a premium request and we let them know what we think we'll be able to allocate to them in terms of premium during the year. To be super clear, we call a capacity provider either an insurer or a reinsurer. The eight programs are our eight insurers on whose paper we place the individual policies, and then we have 32 reinsurers who sit behind those eight insurance companies for a total of 40 capacity providers.
Your next question comes from the line of David Motemaden, Evercore.
I had just a question on the agent sales promotions that you guys were running and it feels like they had a pretty big impact and minimal margin impact on the growth, but minimum margin impact. It didn't sound like you guys were going to continue those in a bigger way. I guess margins are really good, but when you're producing the volume growth that you did this quarter, that's at least, I think, a pretty good trade-off. So I guess, why not do more of these promotions to really drive the volume and the top line growth here?
We certainly have tried promotions in a variety of different settings and continue to do so. So new agents, storm season, post-storm recoveries, government shutdowns. And the one we ran for the government shutdown was pretty specific around getting new agents to do business with us. And so it's sort of an introductory to Neptune kind of promotion. But we're not afraid to do promotions. We're very excited about this new data set that Matt was talking about of having the individual traceable information around actual humans, 30,000 human agents now and being able to more precisely tailor incentives to those agents. And so as we collect data from this new individual log-in system, which started on December 1, over this year, we think we will be able to more precisely be able to deploy those kinds of incentives. It is something we are absolutely focused on. It's not that we're taking them away, but it's around using the dollars in the most effective times and places possible.
Got it. Maybe just following up on that last point about the individual user log-in. Are there any early signs? Any encouraging results or testing you've done, or are you still in the discovery phase?
We're still early. And remember that we have a very seasonal business. So January is normally the slowest month of the year, followed by February, right? So this is the slow time. As we approach hurricane season, things tend to heat up and then volume is very dependent on storm activity and last year government shutdowns. So this is really a data gathering time of year for us, and we would look to deploy and take action upon that when it is actually raining.
Our next question comes from the line of Josh Shanker of Bank of America.
Congratulations on the quarter and the year. Trevor, in your opening remarks, you mentioned that you felt that AI would reduce friction, and it seems to be that commission costs would come down for the industry as one of those frictional costs. In this whole week, regardless of what happened to Neptune, and I just want to think broader, do you believe that the AI sort of concern is real for the industry and that commissions will be under pressure for the insurance industry? I say that knowing that you're paying people about 13%, 14% for about 1 minute of work. And I'd love to hear how you think about the industry's impact.
So Neptune has had a direct-to-consumer channel from the very beginning. And it's always been about 2% of our business every year. Even as we grow, it remains about 2% of our business. Consumers like talking with human agents about the risk that they're taking on in their life and getting good advice. Can I foresee a dystopian future where AI agents are better than humans at convincing us about our risk and better at selling to us? Maybe. I'm not sure that's a great world, but I can futurize and think of that kind of outcome. And that's really what I think the market was jumping to: that kind of outcome where AI agents are just better than humans at doing the job of convincing people what to do and providing them advice. Our point is, if that happens, Neptune's margins improve because when we sell direct-to-consumer, we don't have to pay an agent, right? That is, it's just our technology directly linked in.
We may have had to pay Google AdWords. We may have to pay OpenAI. We may have to pay Anthropic. We may have to pay our new AI overlords, unclear. But those costs generally would need to be lower than what we would have to pay an agent for it to make sense for us to dedicate that money. So I think it is possible to envision that kind of dystopian future, but I don't think that it is a near-term concern. I spent last weekend upgrading my parents' phones, and I installed ChatGPT for them, and they were playing around with it. And I asked my mother this week, have they been using it much? And they said, no, it scared them too much. I'm 53, and I can use it and it makes sense. But I think this is going to be a long road to convince large swaths of Americans to stop taking advice from their friend at the Qantas Club and to start listening to an AI.
There are no further questions at this time. And with that, I will now turn the call over to Trevor Burgess, Chairman and CEO, for closing remarks. Please go ahead.
So thanks, everyone, for joining us this hour to listen to our full year and fourth quarter update. I'm incredibly proud of the work that the team is doing here. We look forward to sharing with you updates throughout this year on the progress that we're making. And I want to give a special shout out to the 30,000 human agents who are working with Neptune every single day to help close the flood insurance gap that exists in America. Thank you all so much.
Ladies and gentlemen, this concludes today's call. We thank you for participating. You may now disconnect your lines.