Ramp

It is a Tuesday night and somewhere a finance team is not staring at a spreadsheet. The receipts have already matched themselves. The month-end close that used to eat the weekend finished on schedule. This quiet is the product. Ramp does not sell software so much as it sells back the hours other companies quietly take. Seventy thousand businesses, from family farms to the Fortune 100, now run their money through it. The bill for the busywork has come due, and Ramp has been collecting the refund.

For a company built on corporate cards, that is a strange place to begin. Card companies are supposed to want you to swipe more. Ramp built its entire pitch on the opposite. The result, six years in, is a $44 billion valuation and a finance platform that does not look much like the industry it disrupted.

Finance was built to reward waste

Here is the awkward truth the industry preferred not to mention. The traditional corporate card business runs on interchange, a small cut of every dollar spent. The more a company spends, the more the card issuer earns. Rewards programs then dangle points to encourage exactly that. The incentives, politely, were upside down. The card in your wallet was quietly rooting against your budget.

Then add the human cost. Expense reports filed at midnight. Receipts photographed in parking lots. A monthly close that turned accountants into archaeologists, digging through statements to figure out what a charge from three weeks ago actually was. None of this work created value. It just had to be done, by someone, forever.

Three friends, one contrarian wager

Eric Glyman and Karim Atiyeh met in a computer science class at Harvard. Their first company, Paribus, tracked prices and clawed back refunds when things you bought went on sale. Capital One acquired it in 2016. The lesson stuck: people will happily let software fight for their money if the software actually wins. In 2019 the pair reunited with Gene Lee, an engineer from the Paribus days, to build Ramp.

The wager was simple and slightly heretical. Build a card company that profits when customers spend less, then surround it with software that automates the finance work nobody wanted to do. Competitors were busy out-pointing each other on rewards. Ramp went after the one luxury the rewards race ignored entirely: time.

It is the kind of bet that sounds obvious in hindsight and reckless in the moment. A startup voluntarily capping the metric that the entire industry optimizes for is not a normal pitch to walk into an investor meeting with. But the founders had a track record of being early to an unloved idea, and a habit of shipping faster than the incumbents could schedule a meeting about it. Ramp launched its first card in 2019 and never really stopped adding to it.

One system instead of seven tabs

What started as a card is now a single place where corporate money moves. Cards come with spend controls that block out-of-policy purchases before they happen, rather than apologizing for them afterward. Bill Pay reads invoices, routes approvals, and sends free same-day ACH and free wires, domestic and international. Procurement turns the chaos of vendor requests into an intake-to-pay workflow. Travel, treasury and accounting automation round out the suite, and the whole thing reconciles itself into your ERP.

The unfashionable through-line is automation that removes a job rather than relocating it. AI codes the transactions, matches the receipts, and flags the subscription you forgot you were paying for. Less glamorous than a rewards multiplier. Considerably more useful at 11pm on the last day of the month.

The numbers do the arguing

Skepticism is the correct posture toward any fintech with a nine-figure raise and an AI story. So look at what is measurable. Ramp processes more than $200 billion in annualized purchase volume. Its customers have collectively saved upward of 27 million hours and $12 billion. The names on the roster - Visa, Uber, Shopify, Anduril, Figma, Notion, Cursor - are not the kind that switch finance tooling for fun.

The plumbing underneath matters too. Ramp's global card program is built on Stripe's issuing infrastructure, which is part of how a New York startup ended up issuing cards across dozens of countries without becoming a bank. Visa carries the transactions and, in a small piece of irony the company enjoys, is also a paying customer. The accounting side plugs straight into NetSuite, QuickBooks, Sage Intacct, Xero and Workday - the unglamorous integration work that decides whether a finance team can actually leave its old system or just bolt a new one on top.

Then there is the part the company is least shy about: it eats its own cooking on AI. Ramp's engineers build with tools like Anthropic's Claude and Cursor, and the same impulse shows up in the product. AI reads the invoices, extracts terms from vendor contracts, and categorizes expenses without a human in the loop. The claim that finance can mostly run itself is easier to believe from a company visibly trying to automate its own work first.

Selling time, charging for trust

The business model is the tell. Ramp earns from interchange and from software tiers that companies choose to upgrade into, not from convincing anyone to spend more. The core card and software stay free for most customers. That alignment - the vendor wins when the customer wastes less - is rare enough in finance that it reads almost like a stunt. It is not. It is the entire strategy.

The $750 million Series F, led by ICONIQ, GIC and Ontario Teachers' Pension Plan, is aimed squarely at pushing AI deeper into the finance stack. Investors hungry for fintechs with a credible AI story found one with revenue and free cash flow to back the narrative, which is more than most can claim.

The night the spreadsheet stays closed

Return to that Tuesday night. The finance team that used to lose it to expense reports has the evening back. Multiply that across 70,000 businesses and 27 million hours, and the thing Ramp is really building comes into focus. Not a card. Not even software, exactly. A standing argument that the busywork of money was never necessary, just unexamined.

The bet now is that AI keeps eating the manual middle of finance - reading the invoice, coding the expense, catching the waste - until the close is something that happens in the background rather than the foreground. If Ramp is right, the most valuable thing a finance platform can hand back was never points or cashback. It was the Tuesday night. And the spreadsheet, for once, stays closed.