Square Sells the Reader at a Loss. You Are the Product It's Buying.

The thing is the size of a matchbook, costs about as much as dinner for two, and Block loses money on every single one it ships. That little white square plugs into a phone and turns a farmers-market stall into a business that takes cards. It is the most famous product the company makes - and on the company's own books, on Block's own filings, hardware costs have consistently exceeded hardware revenue.⁹ That is not a manufacturing failure. It is the entire point.

The official story is that Square is a hardware company that struck gold with a clever little reader. The real story is that the reader was never meant to make money - it was meant to make customers. Square sells you the dongle at a loss to get you onto the rails, and then it never stops selling you the rest.

It started with a lost sale. The founding anecdote is that Jim McKelvey, Jack Dorsey's co-founder, couldn't close a deal because he had no way to accept a credit card - and the name 'Square' comes from the shape of the reader they built to fix that.⁷ The company was founded in 2009, and by 2024 more than four million sellers were running on the Square ecosystem, pushing 5.2 billion transactions and $228 billion in volume through it.¹ Every one of those sellers walked in through a piece of hardware sold for less than it cost to make.

The loss is in the filing, not the legend

Plenty of companies sell cheap gear to win customers. What makes Block unusual is how openly it admits the gear loses money. This isn't analyst speculation or a low-margin shrug - it's visible in the primary filings. Block's FY2023 10-K shows hardware costs of roughly $268M against hardware revenue of just $157M - costs running nearly 70% above what the hardware actually earned.⁹ Its Q1 2024 shareholder letter goes further, noting that hardware gross profit is not presented in the gross profit line management guides Wall Street to,⁵ because including those losses would muddy a number that is supposed to measure the actual business. The reader is treated, internally, as a marketing expense wearing a product's clothes.

And the scale tells you everything. In fiscal 2023, hardware revenue was just $0.157B — under one percent of Block's total revenue — while transactions and subscriptions ran into the billions.⁹ The thing everyone pictures when they think of Square is a rounding error on the income statement. Years earlier the pattern was identical: hardware revenue was $84.5M in FY2019, a sliver next to transaction-based revenue in the billions.⁴ The hardware line has always been small on purpose. It's the bait, not the meal.

Why the cheap reader is the most expensive part

Here is the mechanism, worked down. A loss leader only pays off if the cheap thing reliably leads to the expensive thing. For a grocery store, the discounted milk pulls you past the full-price everything else. For Square, the entry-level reader — priced around $59 as of 2026 — pulls a small seller across a threshold most of them would never cross on their own: the threshold of accepting cards at all.⁶ Once they're across it, they're not buying a reader anymore. They're paying a percentage-plus-flat-cent fee on every in-person sale, forever, with no further sales effort from Square. Square's published rates for its Free plan currently stand at 2.6% + $0.15 per in-person transaction.⁶ The hardware purchase happens once. The take rate happens on every transaction for as long as the business exists.

But the transaction cut is only the second-cheapest reason to give the hardware away. The real prize is everything that gets sold on top of the payment rail: the subscriptions and services line - software, lending, banking - that brought in over a billion dollars in 2023.⁸ A seller who started with a dongle gets nudged into payroll, into inventory software, into a loan underwritten against the transaction history Square can already see flowing through its own pipes. None of that is possible without the first cheap swipe. The reader isn't a product Block sells; it's the cheapest customer-acquisition channel in fintech, paid for with hardware margin instead of an ad budget.

Isn't this just a normal razor-and-blades trick?

The fair objection is that this is nothing new - it's razor-and-blades, the oldest trick in retail, dressed up in fintech language. And that's partly true: the structure is ancient. But two things make Square's version sharper than the cliché. First, the 'blade' isn't a consumable the seller has to keep buying from you - it's a recurring percentage of their own revenue, which scales as they grow, with zero marginal selling cost to Block. A razor company has to keep manufacturing blades; Block just keeps clipping a slice of GPV. Second, the lock-in is data, not just compatibility. Once a seller's payments, payroll, and loan history all live inside one ecosystem, leaving means rebuilding their whole operational nervous system somewhere else. The honest counter is that this isn't bulletproof - a seller can switch processors, and competitors give hardware away too. The reason it holds is that switching costs rise with every additional Square service a seller adopts, and the reader is engineered to be the first of many. The loss leader works because what it leads to is sticky, not because the reader is special.

Block sells the most recognizable thing it makes at a loss, and it tells you so in plain federal-filing English. The mistake is to read that as a flaw. The reader was never the business - it was the doorway, priced low enough that four million small sellers walked through it without thinking twice. The genius wasn't building a cheaper card reader. It was understanding that the cheapest customer you'll ever buy is the one who thinks they're buying from you.