Ramp is interesting to me not just because of what it does but because of how it does it. Corporate cards and expense management are a commoditized category on the surface. The pitch from most players in this space is points and perks. Ramp's pitch is that they will save you money. That repositioning, from a rewards product to a cost-reduction product, is a product decision that runs all the way through the engineering.
The numbers tell a fast-moving story. From a $7.65B valuation in April 2024 to $32B in November 2025. Revenue doubling year over year. $57B in total payments volume. Cash-flow positive. Four funding rounds in 2025 alone. The company is not slowing down.
The Market
Commanding 1.5% of an enormous, mostly untouched market.
Ramp reportedly commands 1.5% of the addressable US corporate spend management market. That means 98.5% of the market is still on legacy tools, still using SAP Concur, still using expense reports submitted on paper receipts, still tolerating the friction. The corporate card market in the US runs into the hundreds of billions in payment volume annually. The software layer on top of it, which is where Ramp operates, is a smaller but rapidly growing slice as companies realize the card is not just a payment method but a data source.
The competitive pressure from incumbents is real but slow. SAP Concur has distribution and enterprise contracts. American Express has brand and float. Neither of them was built for the kind of real-time AI automation that Ramp is shipping. That creates a window.
The Landscape
Where legacy players leave the table.
automation
anomaly detection
platform
UX
● Purpose-built◐ Partial / legacy○ Not a focus
Brex is the closest competitor on product quality, but its focus has shifted toward enterprise and away from the SMB/growth-stage market where Ramp built its base. SAP Concur has distribution but its UX reflects a product designed in an era before real-time data was an expectation. Expensify does expense reports. None of them was built from the start around the intelligence layer that Ramp treats as the core product.
The Engineering
The hard parts that make the product feel easy.
Processing millions of transactions in real time, enforcing spend policies at the point of purchase, detecting anomalies without creating false positives that frustrate users. These are legitimately hard distributed systems problems. The receipt parsing and auto-categorization alone require a reliable OCR and classification pipeline across wildly inconsistent input formats. At scale, a small error rate in categorization is not a rounding error, it is millions of misclassified transactions per year.
I spend my time on this class of problem. At Movius I cut API response times 70% and server load 30% through stress testing and query optimization at scale. At Amazon I built anomaly detection for pricing mismatches across Amazon Go. Doing this reliably, without the false positive rate that makes users work around the system, is the unglamorous engineering that separates a product people advocate for from one they just tolerate.
Ramp made a product that people inside companies actually advocate for. In enterprise software, that almost never happens.
The Frontier
From spend management to autonomous finance.
Ramp launched AI agents for fraud detection and bill payment automation in October 2025. The company frames this as the beginning of "autonomous finance," where AI handles the routine financial operations layer entirely and the finance team focuses on decisions rather than data entry. At $57B in payments volume annually, the dataset to train these agents on is substantial.
The risk is competitive gravity. A $32B valuation draws attention. Brex is not standing still. Stripe is a natural competitor if they chose to move into expense management seriously. American Express has the distribution and the balance sheet. What Ramp has is product quality and a head start on the AI layer, both of which are real but not permanent advantages.