Four years ago, Ramp was a scrappy challenger selling corporate cards to startups that wanted slightly better rebates than Brex. Today it is worth $60 billion, pulls in $2 billion a year, and serves 85,000 companies—making it, by private-market valuation, the most valuable expense-management software firm ever built. The transformation owes less to plastic than to language models: Ramp bet early and heavily on AI-powered automation, and the bet has paid off with unusual clarity.
The new funding round, details of which emerged this week, lands at a moment when most fintech unicorns are nursing down-rounds or hunting for acquirers. Ramp's trajectory is conspicuously different. Its revenue has roughly doubled in the past twelve months, and the customer count has tripled since early 2024. That growth has persuaded investors to mark the company up rather than down, even as interest rates remain elevated and the IPO window stays mostly shut.
Why AI changed the unit economics
Expense software has always been a grudge purchase: nobody loves submitting receipts. Ramp's insight was that large language models could turn grudge into delight—or at least into indifference. The platform now auto-categorizes transactions, drafts policy-compliant expense reports, flags anomalous spending, and negotiates vendor contracts on behalf of customers. Each of those tasks used to require either headcount or user friction; each now runs in the background.
The result is a product that sells itself on savings rather than features. Ramp claims its median customer cuts software and subscription costs by double digits within six months. For CFOs under pressure to show efficiency gains, that pitch lands harder than a new dashboard ever could.
The competitive moat question
Skeptics will note that Brex, Airbase, and even legacy players like SAP Concur are all racing to ship similar AI features. Ramp's lead, however, is measured in data as much as code. Every transaction processed trains the model; every negotiated contract refines the benchmark. That flywheel is difficult to replicate without comparable scale, and Ramp's 85,000-customer base now generates a volume of spend data that smaller rivals cannot match.
The company has also expanded upmarket, winning mid-cap and even enterprise accounts that once defaulted to incumbents. Those contracts are stickier and higher-margin, which helps explain why investors are willing to pay a revenue multiple that would have seemed absurd two years ago.
Our take
Ramp's valuation is aggressive, but it is not irrational. The company has found a genuine productivity gain—AI that actually reduces labor—and wrapped it in a business model that aligns incentives with customers. If the IPO market reopens in 2027, Ramp will be near the front of the queue, and it will deserve to be.
