Extend.

Here is a fact about the extended warranty that everyone in retail knows and almost nobody says out loud: it is the single most declined item at any checkout on Earth. The clerk asks if you want to protect your $89 blender for another two years. You say no thanks. You have said no thanks your whole life. So the interesting question about Extend, the San Francisco software company founded in 2019, is not "how do you sell warranties." It is "how do you build a $1.6 billion company on a product people are conditioned to refuse."

The answer, roughly, is that the refusal was never really about the warranty. It was about the pitch - the awkward upsell, the fine print, the vague sense that the whole thing was designed to benefit everyone except you. Extend's founders looked at that friction and did the thing software people do to friction: they turned it into an API. Merchants integrate a few lines of code (or a prebuilt app on Shopify, Magento, BigCommerce and the rest), and the offer to protect your purchase appears at the right moment, priced by a machine-learning model rather than by a bored employee's intuition. The company reports this approach produces attach rates three to four times higher than legacy protection plans, with some partners clearing 12%. That is a large multiple on a product whose baseline is "no thanks."

Extend's stated ambition is to become the "AppleCare" for everything that isn't an actual Apple product. - The company's own framing of the mission

That comparison is doing a lot of work, and it is worth taking seriously. AppleCare is beloved - or at least tolerated - because the experience is seamless: you buy it in one tap, and when something breaks, the fix is fast and the blame has nowhere to hide, because Apple owns the whole chain. Most extended warranties are the opposite. You buy from a retailer, the plan is administered by a third party, the claim is underwritten by yet another party, and when your thing breaks the parties spend a while pointing at each other. Extend's structural bet - completed in 2022 when it finished building what it calls a vertically integrated, proprietary insurance stack - was to collapse that chain. It is the technology, it is the obligor that owns the risk, and it runs the claims desk. When a plan pays out slowly, there is no one else to blame. Extend decided that was a feature.

The founder is Woodrow Levin, who goes by Woody, and whose resume is the kind that either terrifies or reassures investors depending on the day. Before Extend he traded options in Chicago, started a video game company, and built a startup that offered a digital safe-deposit box with permission-sharing. None of those became Extend. Founders love to sand their failures into a smooth narrative arc; the more honest version is that Levin spent years learning what doesn't work before landing on a boring, enormous market that did. He co-founded Extend with Joe Moss and Michael Darmousseh, and the timing was good: the direct-to-consumer boom was minting hundreds of brands that sold expensive, breakable things - Peloton bikes, Sonos speakers, Traeger grills - and none of them wanted to build warranty infrastructure from scratch.

The extended-warranty pitch used to be a person at a register, guessing. Extend replaced the guessing with a model - and quietly tripled how often people say yes.

The money followed the logic. A $40 million Series B in 2020. Then, in May 2021, a $260 million Series C led by SoftBank's Vision Fund 2, with participation from Meritech, PayPal Ventures, GreatPoint Ventures, Nationwide and a long list of others, valuing the company at over $1.6 billion. It is worth pausing on how strange that sentence is. SoftBank, the fund that famously chases category-defining technology bets, wrote a quarter-billion-dollar check for extended warranties - the thing you decline at the register. The unglamorous corner of insurance turned out to be big enough to be genuinely interesting: the extended-warranty market has long been estimated in the tens of billions of dollars, and almost all of it ran on software that felt like it was built in 2004, because it was.

What Extend did next is the part that tells you where the whole industry is heading. Warranties were only ever the wedge. The real territory is everything that happens after you click "buy" - a zone retail has spent twenty years ignoring while obsessing over the checkout button. Extend pushed outward into shipping protection (for the packages stolen off porches, a problem with the excellent industry acronym of nothing, though the related pain of "where is my order" goes by WISMO), into order tracking, into returns and exchanges, and into automated claims handled with AI identity verification and fraud scoring rather than a call-center queue. By 2025 the company reported its platform had driven more than $1 billion in revenue for its merchant partners across the year - a useful number, because it reframes the sales conversation from "buy our software" to "we made your peers a billion dollars."

In March 2026 Extend bundled all of it under a new banner it calls Shopper Operations, and this is where the story gets a little more pointed. The pitch is AI-native: unify delivery, returns, exchanges and warranty claims into one system, then use what Extend calls Shopper Intelligence to read customer behavior in real time and sort shoppers by value and risk. The uncomfortable, honest premise underneath it is that not all customers are equally good. A small slice of high-risk shoppers - serial returners, fraud, abuse - drive a wildly disproportionate share of costs, and most brands apply one blanket policy to everyone, which means they punish their best customers to defend against their worst. Extend's proposal is to stop doing that: give the loyal customer the generous experience, and quietly tighten the policy on the one gaming the system.

"Most retailers can't distinguish their most valuable customers from those quietly eroding profit... brands can give great customers the experience they've earned while protecting margins." - Woodrow Levin, Co-Founder & CEO

You can hear two things in that quote at once, which is what makes it interesting. One is a genuinely reasonable idea - that good customers deserve better treatment than the person defrauding you. The other is the slightly vertiginous realization that your return policy might soon depend on a real-time score you can't see, calculated from behavior you didn't know was being watched. Both things are true. That tension - between the frictionless experience shoppers want and the margin discipline retailers need - is exactly the seam Extend has spent seven years working, first with warranties, now with an AI platform. Whether "Shopper Operations" becomes a real software category or a well-marketed feature bundle is the open question. But the underlying observation - that the money and the loyalty both live after the sale, not at it - has already been vindicated by 1,200 merchants and a billion dollars. The thing you keep declining at the register turned out to be a doorway.