Castle

Here is a fact about money that should bother you more than it does. The average American will spend hours comparing blenders, read forty reviews before buying headphones, and agonize over a $9 monthly subscription. Then that same person will buy a house - the single largest purchase of their life - on the strength of a spreadsheet, a weekend of open houses, and a strong feeling in the gut. And once they own it, the data situation somehow gets worse.

This is the observation at the center of Castle, a fintech company headquartered not in San Francisco but in Hudson, New York. The pitch is deceptively small: an app that consolidates your home bills. The ambition underneath it is not small at all. Castle wants to build the payments infrastructure for what it estimates is the roughly $3 trillion Americans spend on their homes every year - the utilities, the property taxes, the school taxes, the insurance premiums, the heating bills, the contractor invoices, the maintenance that arrives unannounced and always in the wrong month.

The interesting thing about home spending, from a fintech point of view, is that it never got unbundled and then rebundled the way everything else did. Your investments got a dashboard. Your checking account got an app. Your credit card got real-time fraud alerts. But the money you spend on the roof over your head still moves, in 2026, substantially by cash and check, across three or more different payment methods a month, monitored for fraud by essentially no one. It is the largest consumer category in the country and it is running on the technology of 1995.

Castle's founding insight - and the reason the company exists at all - is that this is not a small annoyance. It is thousands of dollars a year. The company says the average Castle customer saves over $10,000. Not by budgeting harder or living smaller, but by letting software do the unglamorous thing no human wants to do: read every home-related transaction, cross-reference it against a database of tax credits, rebates, exemptions and discounts, and flag the money you are owed or the money you are being overcharged. The savings were always there. Nobody had the patience to find them.

Madeline Hung did not set out to build a fintech company. She set out to build a house. During the COVID-19 pandemic, with a Manhattan lease running out, she and her husband went looking upstate, fell for the Hudson Valley, and bought a historic photography studio - a building with no kitchen, no bedrooms, and no plumbing. Over the following twelve months they designed and renovated it themselves, sourcing materials, finishing cabinets, painting trim.

Hung grew up in a family of real estate enthusiasts, where a well-solved home-improvement problem was discussed like a genuine accomplishment. She was, in other words, exactly the kind of detail-obsessed person who keeps elaborate spreadsheets. And yet, despite the spreadsheets and the research, she found tracking the actual spend shockingly hard. Renovation costs blurred. Post-move-in expenses - maintenance, taxes, insurance, heating - accumulated in ways she couldn't quite reconstruct. She kept finding what she now calls "$5,000 rounding errors": money that had clearly gone somewhere, avoidable with better information, invisible without it.

If a person that organized couldn't keep track of her own home, the reasoning went, then the problem wasn't the person. It was that no tool existed. That is the moment a renovation became a company.

There is a certain kind of startup that goes after the flashiest possible market - the next neobank, the next trading app, the next thing with a slick card design. Castle went the other way and picked the least sexy, most universal expense there is: the roof over your head. And that, on reflection, is the whole point. The biggest markets are frequently the ones no one bothered to build good software for, precisely because they looked too boring, too fragmented, too hard.

The lesson buried in Castle's model is one every consumer fintech eventually learns. The money is rarely in the spending itself - it's in the paperwork around the spending. The rebate you qualified for and never claimed. The property tax exemption you didn't know existed. The utility overcharge that recurs quietly every month. Homeowners don't need another app that scolds them for buying coffee. They need software that quietly claws back money they are actually owed. Castle's bet is that automation, not advice, is what people will pay for.

Whether Castle becomes the payments layer for a $3 trillion category or a well-loved app for a devoted niche is, at seed stage and fifteen people, genuinely unknown. What's clear is the shape of the wager: that the last unbundled corner of consumer finance is the address you already live at, and that the company willing to read all that boring paper first will own the interesting part.