Ramp
The $44 Billion Pre-IPO Deep Dive — revenue doubling to $1B, an AI-agent finance stack, and the cleanest fintech IPO setup since Stripe.
Ramp is the fastest-compounding name in B2B fintech and the most credible near-term fintech IPO: gross revenue doubled to ~$1B in 2025, annualized revenue hit ~$1.5B by May 2026 (+89% YoY), the company is free-cash-flow positive, and it just raised $750M at $44B (June 2026 — ICONIQ, GIC, Ontario Teachers') after nearly 3.5×'ing its valuation in fifteen months. The strategic landscape broke Ramp's way twice this year: rival Brex agreed to sell to Capital One for $5.15B — a fraction of Ramp's mark — leaving Ramp as the category's only independent at scale, and its AI-agent stack (expense approval, accounting, procurement — built on frontier reasoning models) gives it the AI-native story public investors are paying up for. The bear case is the multiple: ~29× run-rate revenue for interchange-heavy economics, against Amex at a mid-single-digit multiple and Bill.com in the single digits. Business Insider reports Ramp telling investors it's preparing to go public around a $1.4B ARR milestone — the S-1 is the catalyst.
The Numbers at a Glance
1 · Business Overview
Seven years from anti-points corporate card to the AI finance-operations platform
| Metric | Value |
|---|---|
| Founded | 2019 · Eric Glyman & Karim Atiyeh (previously sold Paribus to Capital One) |
| HQ | New York City |
| CEO / CTO | Eric Glyman · Karim Atiyeh |
| Round Valuation | $44B · Series F, June 2026 ($750M — ICONIQ, GIC, Ontario Teachers') |
| 2024 Gross Revenue | ~$500M |
| 2025 Gross Revenue | ~$1B · doubled YoY (Forbes Fintech 50) |
| Annualized Revenue | ~$1.5B · May 2026, +89% YoY |
| Profitability | Free-cash-flow positive since Nov 2025 · underlying profitability +153% YoY |
| Customers | 70,000+ businesses · 2,200 paying $100K+/yr (+133% YoY) |
| Payments Volume | $80B+ annualized · $57B in 2024 |
| Equity Raised | ~$3B+ total · Founders Fund, ICONIQ, Lightspeed, Thrive, General Catalyst, Khosla |
| IPO Status | No S-1 filed · telling investors it is preparing to go public (BI, April 2026) |
Ramp launched in 2020 with a contrarian pitch in a market Amex and Brex had trained on points: a corporate card designed to help companies spend less, not earn perks. The card was the wedge; the business is the workflow layer on top — expense management, bill pay / AP automation, procurement, travel, and treasury — sold as one platform that gets stickier with every product adopted. Half of customers already use two or more products, over 30% of contribution profit now comes from software and services beyond cards, and the 2025–2026 arc is the AI turn: agents that approve expenses, code the books, and run procurement — moving Ramp's pitch from "save money on spend" to "automate the finance department."
2 · Product Suite & Moat
The card is the wedge — the agents are the lock-in
| Product | Description | Strategic Role |
|---|---|---|
| Corporate Cards | Physical + virtual cards with policy enforcement per card/vendor/category | Revenue engine (interchange) · acquisition wedge |
| Expense Management | Auto-capture, text-a-receipt, AI approval agents (99% reported accuracy) | Daily-active engagement · lock-in |
| Bill Pay / AP | Invoice OCR → coding → approval → payment; 85%+ fields auto-correct first pass | Tripled YoY · Bill.com attack vector |
| Accounting Agent | End-to-end bookkeeping automation · up to 3.5× more auto-coded transactions, books closed ~3× faster | The AI-native story |
| Procurement Agents | Vendor research, RFx, compliance, renewals · ~16% reported vendor-spend savings | Coupa/Ariba attack vector |
| Ramp Treasury | Institutional cash management · $1.5B+ AUM since Jan 2025 launch | Balance-sheet relationship deepener |
| Travel (+ Juno) | Employee + guest travel booking with policy control | Navan/Concur attack vector |
| International (+ Billhop) | UK/EU-authorized payments · London & Stockholm offices · UK/EU onboarding from summer 2026 | TAM expansion beyond the US |
The moat thesis rests on four pillars:
- Spend-data flywheel — Ramp sees what tens of thousands of companies pay for software, services, and vendors. That corpus powers price intelligence and agent accuracy competitors can't replicate without the volume.
- Multi-product lock-in — cards create daily engagement; each added product (AP, procurement, treasury, travel) raises switching costs. Half of customers run 2+ products; enterprise accounts ($100K+) grew 133% YoY.
- AI-agent head start — agents built on frontier reasoning models are shipped and measurable (99% approval accuracy, 3.5× auto-coding, 16% procurement savings) rather than roadmap slides — and a renewed multi-year Visa agreement plugs Ramp agents into Visa Intelligent Commerce for autonomous B2B payments.
- Last independent standing — with Brex selling to Capital One and Divvy long absorbed by Bill.com, Ramp is the only at-scale independent in corporate spend — scarcity value for public investors wanting pure-play exposure.
3 · Financials & Unit Economics
Doubling revenue while turning cash-flow positive
Ramp's gross revenue ran ~$500M (2024) → ~$1B (2025), with annualized revenue at ~$1.2B exiting 2025 and ~$1.5B by May 2026 (+89% YoY). Underneath: total payments volume of $57B in 2024 (vs $22.3B in 2023) and $80B+ annualized today — implying a blended take across interchange and software of roughly ~1.7–1.9% of TPV. Unusually for a hypergrowth fintech, the profitability arrow points the same way: free-cash-flow positive as of November 2025, with underlying profitability up 153% YoY and >30% of contribution profit from non-card software and services.
The customer mix is climbing upmarket without losing the SMB engine: 70,000+ businesses as of June 2026 (from 50,000 at the start of the year), with 2,200 customers contributing $100K+ in annualized revenue — up 133% YoY. Bill Pay more than tripled YoY; Treasury crossed $1.5B AUM within a year of launch and recently added ~$1B in deposits via the Moment partnership. Ramp has raised ~$3B of equity in total — modest dilution for the scale, reflecting a business that increasingly funds itself.
4 · IPO Status & Timeline
No S-1 yet — but the company is telling investors it's coming
Ramp has not filed an S-1, but the signals stack up: in April 2026, Business Insider reported Ramp telling potential investors it's on track for ~$1.4B in annual recurring revenue "as the company prepares to go public." CEO Eric Glyman has said publicly that Ramp intends to go public rather than sell. The June 2026 Series F brought in exactly the shareholder register you assemble before a listing — ICONIQ, GIC, Ontario Teachers', Goldman Sachs Growth, Morgan Stanley Investment Management — and repeated tender offers (March 2025, November 2025) have already given employees liquidity, easing pre-IPO pressure.
The private market has already run the IPO
From the March 2025 tender at $13B to the June 2026 Series F at $44B, Ramp's mark compounded ~3.4× in fifteen months — each round validated by a revenue base that roughly kept pace (annualized revenue ~$800M → ~$1.5B over the same window, keeping the multiple in the high-20s throughout).
- The Series F syndicate — sovereign wealth (GIC), pension capital (Ontario Teachers'), and crossover desks (Goldman Growth, Morgan Stanley IM) — is the classic pre-IPO register: buyers who anchor order books later.
- Repeated employee tenders (Mar + Nov 2025) mean no pent-up insider-liquidity pressure forcing a rushed listing.
- The Brex sale to Capital One at $5.15B removed the perpetual "Ramp vs Brex" discount debate — there is now exactly one independent at scale.
Base case: a confidential S-1 in H2 2026 against the ~$1.4–1.5B ARR milestone, pricing in the first half of 2027. The catalyst chain: S-1 flip on EDGAR → the first clean look at net revenue vs gross, cohort economics, and agent-driven margin.
5 · Competitive Landscape
Last independent standing — aiming at Amex's $1.5T
| Company | Status | Positioning vs Ramp |
|---|---|---|
| Brex | Selling to Capital One · $5.15B | The direct rival, exiting at ~1/8th Ramp's mark — under a bank's balance sheet and distribution now |
| American Express (AXP) | Public | The real target: $1.5T network volume, ~$500B billed SME business — 20× Ramp+Brex combined |
| Bill.com (BILL) | Public | Owns SMB AP/AR (and Divvy); single-digit revenue multiple frames the bear case |
| Corpay (CPAY) | Public | Corporate payments consolidator; value-multiple comp |
| Navan | Private | Travel + expense; the T&E flank |
| SAP Concur / Coupa / Ariba | Incumbent | The legacy enterprise estate Ramp's agents are underpricing |
| Mercury / vertical SaaS cards | Private | Neobanks + embedded-card players nibbling at the SMB base |
The strategic frame most investors miss: Ramp and Brex were never really fighting each other — the prize is displacing Amex and the legacy T&E/AP stack. With Brex now inside Capital One, that fight has two shapes: bank-owned scale (Capital One/Brex) vs independent platform velocity (Ramp). Public-market scarcity favors Ramp — after this deal, there is no other way to buy pure-play corporate-spend software at scale.
6 · The AI-Agent Bet
From spend platform to autonomous finance department
Ramp's AI story is unusually concrete for the category. Its agents — built on frontier reasoning models — are deployed against the three most expensive workflows in the finance back office, with published performance claims:
- Expense approval agents — policy enforcement and fraud prevention with a reported 99% accuracy in expense approvals; controllers stop reviewing receipts.
- Accounting Agent — end-to-end bookkeeping: up to 3.5× more transactions auto-coded than legacy tools at 90%+ accuracy; finance teams report closing books ~3× faster and saving 40+ hours a month. The agent learns customer-specific logic over time.
- Procurement agents — vendor research, RFx generation, compliance checks, renewal intelligence: customers report ~16% annual savings on vendor spend and purchases completed ~3× faster.
- The Visa rails — a renewed multi-year issuing agreement plus integration with Visa Intelligent Commerce and the Trusted Agent Protocol positions Ramp as default infrastructure for agent-initiated B2B payments — the plumbing for autonomous finance, not just a chatbot skin.
The flywheel argument: every transaction Ramp processes trains better price intelligence and agent behavior, which saves customers measurable money, which pulls more spend onto Ramp. It's the same data-compounding logic that justified premium multiples for CrowdStrike in security or Datadog in observability — applied to the CFO stack. The June 2026 round was explicitly priced off this story; the S-1 will have to quantify how much of revenue is agent-attached software versus interchange.
7 · Valuation Framework
~29× run-rate — priced like software, monetized like payments
| Mark | Implied Value | vs Annualized Rev (~$1.5B) | Context |
|---|---|---|---|
| Series F · Jun 2026 | $44B | ~29× | Priced off ~89% growth + FCF-positive + AI story |
| Nov 2025 round | $32B | ~27× (then ~$1.2B) | Multiple roughly constant — growth carried the mark |
| Brex → Capital One | $5.15B | — | The strategic floor for the category's #2 |
| Bill.com (BILL) | — | single-digit × | The public bear anchor for SMB payments software |
| American Express (AXP) | — | mid-single-digit × | Scaled interchange economics price like a financial |
The bull math: hold ~80–90% growth for two more years and $44B becomes ~12–15× 2027 run-rate — defensible for FCF-positive compounding software. The bear math: if public investors classify Ramp as payments (interchange) rather than software (agents), the comp set is BILL and AXP, and the multiple debate starts at single digits. The S-1's disclosure of net revenue mix — interchange vs software/services (already >30% of contribution profit and rising) — will decide which shelf Ramp sits on. That mix shift is the whole IPO thesis.
8 · Key Risks
What public-market diligence will price in
| Risk | Severity | Mitigant |
|---|---|---|
| Multiple compression at IPO (~29× run-rate) | High | Growth + FCF + software-mix shift argue for a premium shelf; the S-1 mix disclosure is the swing factor |
| Interchange dependence / classification risk | High | >30% of contribution profit already non-card and rising; agents monetize as software |
| Capital One + Brex | Medium | Bank balance sheet + distribution vs Ramp's product velocity; integration typically slows challengers |
| Credit / macro cycle | Medium | Charge-card model with short duration; SMB spend contracts in downturns — TPV is cyclical |
| AI-agent claims vs audited reality | Medium | Performance stats are company-reported; the S-1 and public quarters will test them |
| Interchange regulation | Low–Medium | Commercial-card interchange has been stable; diversification reduces exposure over time |
| International execution (UK/EU 2026) | Low–Medium | Billhop authorization + ~half of customers already transacting internationally de-risk entry |
9 · Pre-IPO Exposure Routes Today
Indirect vectors before the eventual listing
| Route | How | Caveats |
|---|---|---|
| Private secondary marketplaces | Ramp common via secondary platforms | Accredited only · post-Series-F marks near the $44B round · ROFR/company consent apply |
| Crossover funds | Fidelity-style funds and pre-IPO vehicles holding Ramp | Diversified, no control of entry price |
| BILL / CPAY | Public SMB-payments and corporate-payments beta | The bear-case comps — cheap for a reason |
| AXP · V · MA | The rails Ramp rides (and the incumbent it attacks) | Heavily diluted exposure; Visa is a direct partner via Intelligent Commerce |
| Patience | Wait for the S-1 on EDGAR | Cleanest entry · first real look at net-revenue mix and cohort economics |
Bottom Line
Ramp is the cleanest large fintech IPO setup in the pipeline: ~$1.5B annualized revenue growing ~89%, free-cash-flow positive, 70,000+ customers, and the category's only independent at scale after Brex's $5.15B exit to Capital One. The June 2026 Series F at $44B — anchored by sovereign, pension, and crossover capital — reads as the final private round before a listing the company is already telling investors is coming. The debate is not quality; it's shelf classification: at ~29× run-rate, Ramp is priced as AI-native software, while its revenue still leans on interchange that public markets price in single digits. The S-1's net-revenue mix will settle it. For growth portfolios, Ramp is a top-tier IPO-watchlist name — with the cleanest entry at the S-1 roadshow, the Brex takeout as the category floor, and the agent-attach rate as the number that decides whether $44B was the discount or the top.