Ramp

IWhat it does

Ramp is not listed on any stock exchange. You cannot buy its shares on the open market. It is a private company, owned by its founders, employees and venture and growth investors, and the figures on this page come from the company’s own announcements and press reports, not from audited filings¹¹⁸.

Ramp started in 2019 as a corporate card for startups. A corporate card is a card a company issues to its staff so that spending is tracked and capped by finance. Ramp’s card is a charge card, which means the balance is paid in full and the company does not charge interest on carried balances⁵. The card is the front door. Around it Ramp has built software for expense reports, bill payment, buying approvals (procurement), accounting sync and, in 2026, accounts receivable, which is the money customers owe you¹.

The pitch to a customer is simple: spend control and time saved. Ramp says its customers save over $1 billion a month, that median savings rose to 7.5% of spend from 5% a year earlier, and that it saved 1.5 million hours in September 2026¹². These are company claims that I cannot check. They are marketing numbers, but they explain why finance teams switch.

In 2026 Ramp also pushed into AI. It launched a tool that tracks what a company pays AI model providers, a router that sends AI requests to cheaper models (Ramp says savings of up to 40%), and Glass, an AI workspace¹. Ramp’s CEO argues that AI model makers have little reason to help customers spend less, so an independent tool has a role⁸. Whether this becomes a real business or a feature is not yet visible in any disclosed number.

Scale is the other part of the story. Ramp says it serves more than 85,000 businesses, including Visa, Uber, Shopify, Robinhood and Notion, and that customers paying it more than $100,000 a year grew 113% year on year¹. In June it counted 3,200 such customers³. It started 2026 with about 50,000 customers, per Bloomberg as relayed by PYMNTS⁹. Adding 35,000 customers in under a year is fast, which is why the numbers draw attention and why they deserve care.

IIHow it makes money

Ramp’s own help page names three sources. First, interchange: every time a card is used, the merchant pays a fee, and Ramp keeps a portion while Visa and the card-issuing bank share the rest⁵. Second, Ramp Plus, an optional paid plan with advanced features⁵. Third, other services such as international and same-day payments⁵. The basic product is free to customers, so the card does most of the work.

Ramp has not disclosed how revenue splits between these. Outside estimates disagree. Sacra put the interchange take at about 1.6% of card volume in August 2025 and described a move toward subscriptions, at $15 per user per month for Ramp Plus, plus payment fees, travel fees and interest on deposits¹¹. That is an estimate, not a reported figure, and it is more than a year old.

A simple check shows why the split matters. Ramp reports over $275 billion of annualized purchase volume and over $2 billion of annualized revenue¹. Dividing the two gives roughly 0.7% (my calculation). If interchange were 1.6% of all that volume, revenue would be above $4 billion, so either the volume figure includes lower-fee flows such as bill payments, customer incentives are deducted, or the rate is lower now. Ramp says its revenue figure includes interchange, subscription, interest and other revenue, and excludes customer incentives¹. I cannot reconcile this without disclosure.

The consequence is that revenue grows with spend. When customers spend more on the card, Ramp earns more. That has helped while companies spend heavily on AI tools and cloud, and it would hurt if spend fell or if card fees were cut by law, which I cover under risks.

On profit, the company says cash flow is positive¹³. It has not said what its operating margin is, how much it loses to unpaid balances, or how much it spends on stock pay. Positive cash flow can sit alongside accounting losses, so I treat it as a good sign but not as proof of profit.

Revenue and customer figures as announced. Reported means announced by the company.

IIIThe market and the competition

Ramp sells to the finance team of a business. Its market is the money companies spend on cards, supplier bills, travel and software, plus the fees on moving that money. The company says its volume is above $275 billion a year¹. I have not found a reliable figure for total market size, so I do not give one.

The closest rival was Brex. In January 2026 Capital One agreed to buy Brex for $5.15 billion in cash and stock, about 60% below Brex’s 2021 valuation of $12.3 billion, per PitchBook as cited by Banking Dive¹². Pulse 2.0 reported the deal completed by April 2026¹³. Brex was said to serve about 25,000 companies¹². The Brex sale shows two things. A card-first spend business can be sold for much less than a venture price, and a large bank now owns a direct competitor with a balance sheet, a card network relationship and customers of its own.

Other rivals come from different sides. BILL Holdings sells payables and receivables software to small businesses: 479,300 customers, $1.65 billion of fiscal 2026 revenue, growing 13%⁶. Navan sells travel and expense software and is listed¹⁶. Corpay sells fleet and corporate payments and is far more profitable⁷. Sacra also lists Mercury (banking with cards and bill pay), Rippling and Deel (payroll with spend) as competitors, and says Coupa, Navan and Zip partnered with Brex¹¹. Large banks and card networks sell corporate cards too.

Ramp’s edge in this crowd is speed of product and a single product that spans cards, bills, buying and accounting. Its weakness is that cards are easy to copy, and customers can hold more than one. Comparison articles written by vendors, including Ramp’s own, say competitors are card-first or travel-first; I treat those as marketing and do not rely on them.

IVThe moat and how it could be disrupted

I rate the moat as contestable. That means the advantages are real but may not last. It is the honest middle of the scale, not a slight.

What protects Ramp today. Switching costs rise once a finance team has connected its accounting system, approval rules, vendor list and card limits to one product. Scale gives data on spending that Ramp says it uses to catch savings and to judge credit. Growth in larger customers, up 113% year on year, suggests the product is passing tests at bigger firms¹. A brand built on speed also helps recruiting and sales.

What does not protect it. A card is a commodity: the same Visa network and similar rewards are available from banks. Ramp’s fee income comes from interchange set by networks and, in the US, shaped by regulation. A customer can usually add a second card from another provider in an afternoon. Software and AI make it cheaper for any rival to build the same features.

How it could break. A bank with a large balance sheet and ready customers can match Ramp on price. The Capital One purchase of Brex is the plain example¹²¹³. Rules that route card payments over other networks could lower fees for merchants and so the fee Ramp earns¹⁴. AI agents could take over the work of coding and approving spend, and if software becomes cheap, the software layer loses pricing power. Last, a customer who does not need Ramp’s card could use a cheaper tool that connects to any card.

Two things would change my rating. If Ramp showed that software and payments revenue, not interchange, are most of its profit, and that customers stay even when the card is not the cheapest, I would move toward strong. If a bank-owned rival started winning larger customers on price, I would move toward thin.

VThe numbers we can see

Everything below is company-announced or press-reported. Ramp has not published audited accounts, and I found no registration statement with the SEC¹⁸. That search was narrow, a text search on the company name, so a confidential filing would not appear.

Revenue has grown very fast. Sacra estimated about $476 million of annualized revenue in August 2024 and about $1 billion in August 2025¹¹. Ramp now says over $2 billion¹. Its blog says the $2 billion mark was passed over the summer of 2026². The numbers do not line up neatly. In June the company’s release said over $1 billion³, while Bloomberg reported over $1.5 billion for the same month⁹. The October release also says growth is 25% faster than a year earlier, which I read as an acceleration but cannot pin down from the wording¹. Annualized revenue means a recent month or quarter multiplied up to a year. It is not the same as revenue booked in a calendar year, and it can overstate a full year if growth has been recent.

Valuation has moved quickly. Ramp was valued at $22.5 billion in July 2025²⁰, then $32 billion in November 2025 with a $300 million primary round plus an employee tender offer¹⁹. In June 2026 it raised $750 million at $44 billion, led by ICONIQ, GIC and Ontario Teachers’ Pension Plan³⁸. On 15 October 2026 it announced $1.85 billion at $60 billion, led by Dragoneer¹. The valuation therefore rose about 36% in four months. Bloomberg’s reporting, relayed by PYMNTS, says $60 billion excludes the new money, implying a pre-money value of $60 billion and about $61.85 billion after⁹. The company release does not say which it is¹.

Against public peers, $60 billion looks rich. It is about 30 times annualized revenue (my calculation from s1). BILL trades at about 2.5 times revenue with 13% growth¹⁵⁶. Navan trades at about 7.2 times revenue¹⁶. Corpay trades at about 6.9 times revenue and 12.3 times EBITDA, which is earnings before interest, tax, depreciation and amortization¹⁷. Ramp is growing much faster than all three, which is the case for a premium. But Ramp has not shown that it keeps much of each dollar of revenue as profit.

Valuation steps and public comparables. Public multiples are as of the 2 Nov 2026 close.

VIWho runs it and who owns it

Eric Glyman and Karim Atiyeh founded Ramp with Gene Lee in 2019¹⁰. Capital One bought their earlier price-tracking app, Paribus²¹. On 25 June 2026 the company made Glyman and Atiyeh co-CEOs and named Rahul Sengottuvelu chief technology officer⁴. Glyman wrote that Atiyeh already managed risk, operations and marketing⁴. Risk sits under Atiyeh, which matters for a business that extends credit.

A co-CEO structure can work, and it can blur who decides. I cannot judge how well Ramp’s structure works from outside.

Ownership is spread among large investors. Founders Fund was an early backer and went on to lead five of Ramp’s rounds²¹. Rounds since 2025 added Lightspeed, ICONIQ, GIC, Ontario Teachers’, Dragoneer, Thrive, Sands Capital, Fidelity, T. Rowe Price and many others¹³. Ramp says it has raised about $5 billion in equity¹. Stake sizes, the share count and the preference terms (who gets paid first in a sale) are not public. That matters because late investors in recent rounds may hold terms that early holders and employees do not. A Ramp share price today is not the same as a share of the $60 billion headline.

Employees were allowed to sell some stock in a tender offer in November 2025¹⁹. I found no report of a tender in the October 2026 round. Headcount was about 1,200 in 2025 and I have no newer figure²¹.

VIIValuation: is the last price supported?

Equity value in US$ billions. Not a share price: a private company has none. Latest mark: primary round, pre-money per Bloomberg reporting (company does not state it).

What the last valuation assumes

At $60 billion before new money, or about $62 billion after, an owner must believe the company can earn enough by 2031 to justify that price after a 12% yearly return and 2% yearly dilution. That means about $117 billion of equity value in 2031. At 25 times earnings and a 28% operating margin, that needs roughly $22 billion of revenue, or about 61% growth a year for five years from $2 billion (my calculation). At 30 times earnings and a 35% margin, it needs about $15 billion, or about 49% a year. Both sit at or above my bull path on growth.

Bear, base and bull

• Bear. Annualized revenue of $2.0 billion grows 20% a year to $5.0 billion by end 2031 as fee pressure and bank rivals bite. Operating margin is 15%, tax 25%, so net income is about $0.56 billion. I value that at 15 times earnings, in line with Corpay and BILL on adjusted earnings. Discount rate 13%, dilution 2% a year. Equity value today about $4 billion.
• Base. Revenue grows 38% a year to about $10 billion by end 2031, a decaying path from today’s roughly 100% pace. Operating margin 28%, tax 25%, net income about $2.1 billion. Terminal multiple 25 times earnings, a premium to the 13 to 15 times implied for Corpay and BILL, because growth is still above 15% in 2031. Discount rate 12%, dilution 2% a year. Equity value today about $27 billion.
• Bull. Revenue grows 55% a year to about $18 billion by end 2031, with software and payments pushing operating margin to 35%. Net income about $4.7 billion at 30 times earnings. Discount rate 11%, dilution 1.5% a year. Equity value today about $78 billion.

Comparables used

Where I come down

The latest valuation is well above my base of $27 billion and above my margin-of-safety value of $20 billion. It sits between my base and bull values. To be right, Ramp must keep growing near or above 50% a year for five years, earn software-like margins, keep its fee income, and still be rewarded with a premium multiple. The market could be right even if this looks high. Revenue doubled in under a year, volume is up 175%, big customers are up 113%, and the company says cash flow is positive¹. Investors including Dragoneer, Thrive and Founders Fund have seen more than I can, such as margins, losses and cohorts, and they paid this price. My model uses public data only and is sensitive to the margin and multiple I assume. A 5-point change in margin moves my base by about a fifth.

What I cannot judge

I cannot judge Ramp’s true operating margin, credit losses, revenue split between interchange and software, the share of revenue from AI-related customers, preference terms, or the share count behind a per-share price. I also cannot test the annualized revenue figure against audited accounts. Public data is thin enough that I would hold a wide range of 4 to 78 billion rather than trust one point. This is research, not investment advice.

VIIIThe IPO

No IPO plan reported. Ramp is private. A narrow SEC EDGAR text search found no S-1 naming Ramp Business Corporation¹⁸. The October 2026 release and blog do not mention an IPO¹², and the CNBC and Bloomberg-based pieces on the last two rounds do not either⁸⁹. Some web aggregators claim Ramp told investors it would be IPO-ready by the end of 2026; I could not find that in a primary source, so I do not rely on it. No exchange, ticker, banks or size have been reported. A confidential filing would not show on EDGAR, so absence of a filing is not proof of absence of plans.

What to watch

• A public S-1 on EDGAR, which would give audited revenue, margins, credit losses and the revenue split
• Press reports of bank selection or a confidential filing
• Another tender offer or secondary sale, which would give a fresh price
• Any Senate floor vote on the Credit Card Competition Act¹⁴
• Whether Capital One uses Brex to cut prices on corporate cards¹³
• Whether the company publishes margin or loss figures

IXWhat could go wrong

The question I ask first is what most likely destroys the value. For Ramp I think it is a price war on card fees combined with a fall in growth, at a valuation that already assumes neither.

Fee rules. The Credit Card Competition Act would require large card issuing banks to allow at least one card network other than Visa and Mastercard. It was reintroduced in January 2026 with a public endorsement from President Trump, and three more senators signed on by August 2026¹⁴. I found no floor vote. If it passed, merchants could pay less, and the portion Ramp keeps could fall. Ramp’s bank partners and the exact rule are unknown to me, so I cannot size the effect.

Credit risk. Ramp gives companies a spending limit and gets paid later. If many startups fail in a downturn, unpaid balances rise. The company has not disclosed loss rates. A charge card limits some risk because balances are due in full, but it does not remove it⁵.

Growth slowing. The valuation depends on growth staying very high. Revenue doubled over the last year on the company’s own numbers¹¹¹, and compounding at that rate gets harder as the base grows. Some named customers are AI companies, such as Cursor¹. If AI funding slows, their spend slows, and so does Ramp’s interchange. I cannot say how large that share is.

Competition from banks. Capital One owns Brex¹³. Other large banks could respond. Bank-owned rivals can fund price cuts that Ramp’s investors might not want to match.

Disclosure. All numbers come from Ramp or from reporters who talked to its backers. Annualized revenue excludes customer incentives, yet the definition of incentives is Ramp’s¹. Without audited accounts, I cannot test margins, losses or the revenue split.

What has to go right together. For my base case, revenue must compound near 38% a year for five years, margins must reach the high 20s, fees must hold, credit losses must stay low, and public investors must pay a premium multiple in 2031. If I give each of those even a better-than-even chance, say 0.6, 0.6, 0.8, 0.8 and 0.7, the joint chance is about 16% (my estimate, not a measured probability). That is why I do not treat the base case as likely, just as a central path.

The next best use of the same dollar is a listed company that already earns what it reports. Corpay is an example of a public payments business trading at about 12 times EBITDA with audited results¹⁷⁷. That is not a recommendation. It shows the choice an owner faces: pay for growth that is not yet shown in profit, or pay less for profit that is.