Ramp, the New York-based corporate spend management platform, raised $1.85 billion at $60B pre-money on Thursday, pairing the financing milestone with a set of products that gave a specific name to a line item most finance teams have only recently begun tracking: the cost of AI agents making decisions and purchases on their company's behalf. The round was co-led by Dragoneer Investment Group, Thrive Capital, Founders Fund, Sands Capital, and Avenir Growth, with participation from Capital International, Fidelity Management & Research, Khosla Ventures, T. Rowe Price, and a broader institutional syndicate, bringing Ramp's total capital raised to $5 billion since its 2019 founding.

The metrics that accompanied the announcement made the investor appetite legible. Annualized revenue has crossed $2 billion — up from a reported $1.5 billion just four months ago — and revenue growth is running 25 percent faster than it was at the same point last year. Ramp now serves 85,000 businesses, up from 70,000 in June and 50,000 at the start of 2026. Annualized purchase volume processed on the platform $275 billion, up 175 percent year-over-year.

Ramp's Valuation Climbed From $32B to $60B in Under Twelve Months

The speed of Ramp's valuation arc over the past year is unusual even in the current AI-driven funding environment. The company was valued at $32 billion after a $300 million round in November 2025. It reached $44 billion in June 2026, when it raised $750 million in a round that included ICONIQ Growth, GIC, and the Ontario Teachers' Pension Plan. Now, four months after that, it has added another $16 billion in implied enterprise value — a 36 percent step-up on an already elevated mark, backed by metrics that make the progression feel operationally grounded rather than speculative.

The $60 billion figure is pre-money, meaning the newly raised $1.85 billion is additive on top. At the post-money implied value, Ramp is priced at a level that comfortably exceeds almost every publicly traded fintech company outside of Stripe and would rank among the largest US fintech outcomes of all time if and when a liquidity event arrives. Revenue above $2 billion, triple-digit transaction volume growth, and stated positive free cash flow are the combination that makes an S-1 filing look plausible rather than aspirational, and several reports suggest IPO conversations have been held, though no timeline has been announced.

For the competitive field, the round repriced the entire AI-powered spend management category in a single announcement. Ramp's closest former rival, Brex, was acquired by Capital One for $5.15 billion earlier this year — a gap of roughly twelve times between two companies that, just a few years ago, were routinely discussed in the same breath.

From Corporate Card to AI-Native Financial Operations Platform

Ramp launched in February 2020 with a straightforward pitch: a free corporate card that made money when its customers spent less, not more. While incumbents profited from maximizing interchange and encouraging spending, Ramp aligned its business model with customer cost reduction. Eric Glyman and Karim Atiyeh, who had previously co-founded Paribus — a price-tracking app Capital One acquired in 2016 — brought that same spend-reduction logic to the corporate card market. Gene Lee completed the founding team.

That foundational alignment became the surface on which Ramp built a far more ambitious platform. The company now covers expense reporting, bill payments and accounts payable automation, procurement, vendor management, travel, accounting integrations, and AI-powered fraud detection. Its customer base expanded beyond the tech startups that were its early adopters to include mid-market and enterprise clients — Shopify, Anduril, Notion, and CBRE among them.

The most consequential expansion of the last 18 months, however, has been into AI-native financial workflows. Ramp has embedded AI agents directly into its procurement, expense management, accounting, and budgeting products — moving from software that surfaces data and makes recommendations to software that takes action autonomously. Through a deepened partnership with Visa, Ramp has extended that agent capability to autonomous corporate payments with real-time controls, meaning AI systems can, within configured governance frameworks, spend money on a company's behalf.

In June, Glyman framed the company's longer arc in terms that position it as infrastructure rather than software. "For 500 years, business ran on two pillars of spend: people and vendors. Ramp is the infrastructure for the third pillar," he said at the Series F announcement. The third pillar, in Glyman's framing, is AI agents — systems that incur costs, make purchases, and manage vendor relationships without direct human initiation for each transaction.

Router.com Routes API Calls Across Eight LLM Providers, Matching Cost to Task Complexity

The product most emblematic of Ramp's new positioning is Router.com, an AI model routing service Ramp launched on August 20, 2026. Router.com functions as a unified API gateway that lets companies route large language model requests across eight providers — OpenAI, Anthropic, DeepSeek, Moonshot, Minimax, Nvidia, xAI, and Z.ai — from a single interface, applying routing logic to minimize cost without degrading output quality.

The routing strategies Router.com supports are four in type: preference-based routing, which directs calls to a user-specified provider or tier by default; benchmark-based routing, which selects the model performing best on up to three user-specified benchmarks; difficulty-based routing, which sends complex prompts to more capable and expensive models while directing simpler queries to cheaper ones; and A/B testing mode, which allows teams to switch test models without modifying their code. A dashboard surfaces token spend, cost per call, latency, and fallback attempts in real time. Router.com retains model inputs and outputs by default for one year, with an opt-out available; Ramp states that personally identifiable information is removed before any data is used for product improvement.

What distinguishes Router.com from standalone LLM routers such as OpenRouter is the integration advantage that Ramp's existing platform provides. Because Ramp already processes corporate card transactions, vendor invoices, and accounts payable for 85,000 businesses, it can correlate AI model spend with business transactions directly—linking the cost of an LLM API call to the deal it helped close, the support ticket it resolved, or the engineering sprint it accelerated. That AI-spend-to-business-outcome correlation is what "measuring AI ROI" actually requires, and it belongs to whoever controls both the payment layer and the outcome data simultaneously. No standalone LLM router, operating independently of a company's financial operations platform, can provide it.

"Ramp is going to be the primary way people manage their token spend, how they optimize it, how they measure ROI," Keith Rabois, who has backed Ramp through both Founders Fund and Khosla Ventures since the company's early rounds, said in prior commentary on Ramp's positioning.

Router.com launched free for the rest of 2026, with new users receiving a $26 credit. It is currently available in the US only.

Complementing Router.com is Glass, an AI automation workspace Ramp launched in October 2026 that collects and retains AI automations, prompts, and skills that employees build — creating institutional memory for AI workflows so each team member's work can be reused across an organization rather than rebuilt from scratch. Internally, Ramp itself uses a product called Inspect — an AI coding agent the company has said writes more than two-thirds of all internal code — signaling the extent to which the company's own engineering operates on the AI-native stack it is selling to customers.

Use of Proceeds: Product, Engineering, and Enterprise Scale

According to reporting and investor communications around the round, Ramp plans to deploy the capital across three priorities. First, accelerating product development in its AI-agent ecosystem — particularly the agentic financial workflow tools that have driven its recent enterprise growth. Second, expanding engineering headcount to support that product buildout. Third, scaling enterprise sales and customer success capacity to capture the momentum represented by more than 3,200 customers with $100K-plus accounts, as disclosed in its June 2026 metrics. That enterprise cohort was growing at a 100 percent year-over-year rate as of the June round.

Ramp also entered the accounting firm market earlier in 2026 through a product called Stack, launched June 3, 2026, which it describes as financial operations infrastructure for accounting practices rather than just the companies they serve. The strategic logic is a distribution multiplier: accounting firms that deploy Ramp's tooling bring their entire books of client companies into the ecosystem, accelerating customer acquisition at a cost structure that direct-sales enterprise motions cannot match.

The government sector has also emerged as a meaningful surface for Ramp's ambitions. In January 2025, Ramp published what it called "The Efficiency Formula" — a positioning document framing its spend management platform as a tool for reducing government waste, with the company reportedly considered for a General Services Administration SmartPay charge card pilot. Thursday's announcement did not specify whether the current round accelerates that public-sector effort, but it represents a distribution channel with a materially different customer acquisition profile than the company's startup and enterprise segments.

How Does Ramp's AI Model Router Differ from Other LLM Cost Management Tools?

The market for controlling AI API expenditure has attracted a range of tools since enterprise AI adoption accelerated in 2024. LLM observability platforms such as Helicone, Langfuse, and LLMonitor track token costs and provide dashboards. OpenRouter offers a unified API across dozens of models. Dedicated FinOps-for-AI tools have emerged specifically to optimize inference spend.

What separates Ramp's Router.com from this category is not primarily routing capability — OpenRouter supports more models — but platform context. A standalone LLM router sees only the token cost. Ramp sees the token cost, the corporate card transaction that may have resulted from an agent's decision, the vendor invoice the agent approved, and the travel booking the AI system submitted. That cross-category visibility enables a form of AI expenditure governance that tools outside a company's financial operations platform can't match.

Christian Jensen, a partner at Dragoneer Investment Group, summarized the investor thesis for Ramp's expanding financial operations platform: "In just a few years, Ramp has become a much broader financial operations platform, with AI deeply embedded." That deepening is the strategic rationale for why an LLM routing product belongs on a spend management platform rather than as a standalone product — and why the integration moat, not the routing algorithm, is what the round is pricing.

Rippling, Coupa, and the Consolidating Competitive Field

Ramp now competes across a wider set of categories than it did two years ago. In corporate cards and expense management, its traditional competitors include Navan (formerly TripActions), Mercury, and Airbase (now part of Paylocity). In procurement and accounts payable automation, it competes with Coupa and SAP Ariba. In agentic finance broadly, it competes against general-purpose AI workflow platforms increasingly acquiring financial capabilities.

The competitive landscape, however, has shifted meaningfully in Ramp's favor. TechCrunch's October 15 reporting identified Rippling — the HR and payroll company that has moved into corporate cards, expense management, and bill payments — as Ramp's most direct current overlap competitor. Capital One's acquisition of Brex removes the most closely matched independent challenger. Navan has focused its positioning more tightly around travel. Mercury has staked its identity primarily on banking. None has matched the breadth or pace of Ramp's product expansion into the AI-native financial operations category.

Investor Roster Signals a Bet on AI Infrastructure and Enterprise Finance Converging

The spend management software market was estimated at approximately $29 billion in 2026 and projected to reach roughly $46 billion by 2030 at a 12 percent compound annual growth rate. Ramp's current processing volume and customer base suggest it holds a low-single-digit percentage share of the $2 trillion-plus corporate card market — a figure that implies substantial addressable headroom even before counting the adjacent categories it is actively entering.

Globally, fintech startups raised $28.6 billion in the first half of 2026 — a 22.7 percent year-over-year increase, per Crunchbase — but that capital concentrated among fewer, larger companies with demonstrated AI-native growth stories. Ramp's $1.85 billion round is the clearest expression of that dynamic: institutional capital has grown selective, and the companies still attracting it at scale are ones that have earned the "AI-native" label through product rather than positioning.

The structure of the round carries implications beyond Ramp's own trajectory. The five co-leads — Dragoneer, Thrive Capital, Founders Fund, Sands Capital, and Avenir Growth — represent a cross-section of the most sophisticated long-duration fintech investors in the market, joined by institutional names including T. Rowe Price, D1 Capital Partners, Greenoaks, and General Catalyst, among others.

Thrive Capital's participation is particularly notable given its simultaneous position as one of OpenAI's largest backers. The overlap signals a conviction that AI infrastructure spending and AI-powered business tooling are two sides of the same investment supercycle — that the companies managing AI costs will be as consequential as the companies building the models generating those costs.

The round's confirmation came after what September Bloomberg reporting had telegraphed — a pattern in which secondary market data tends to surface late-stage private round activity several weeks before formal announcement. The formal announcement October 15 drew corroboration from TechCrunch and Axios Pro Fintech Deals within hours, each with independently reported details on investor composition and operating metrics.

For enterprises managing AI expenditure as a meaningful budget category for the first time, Ramp's AI token spend management and agentic payment products address a pain point that did not exist two years ago and has no obvious incumbent solution. The ability to govern and optimize AI costs through the same platform that handles all other corporate spend is a genuinely novel position — and one that grows more defensible as AI expenditure grows as a share of total company operating costs.

Whether Ramp can sustain its trajectory as an independent platform company — and on what timeline a public market listing might arrive — remains open. What the $60 billion round established on Thursday is that the institutional conviction required to support that trajectory is present, and that the company's operating momentum has, at least for now, grown faster than the multiple its valuation implies.

Frequently Asked Questions

What is Ramp's AI model router and how does it save companies money?

Router.com is Ramp's unified API gateway for large language model access, launched in August 2026. It routes API calls across eight LLM providers — including OpenAI, Anthropic, DeepSeek, and xAI — applying one of four strategies: preference-based routing (default provider), benchmark-based routing (model ranked by specified benchmarks), difficulty-based routing (complex prompts go to expensive models; simple ones to cheaper ones), and A/B testing. The cost savings come from the difficulty-based strategy in particular: tasks that don't require a frontier model's capability get routed to a cheaper model automatically, without engineering intervention. Router.com is currently free through the end of 2026 and US-only.

How does Ramp's AI spend management differ from standalone LLM cost-tracking tools?

Standalone LLM cost-tracking tools see only what happens inside the model API layer — token counts, cost per call, latency. Ramp's integration advantage is that it already processes corporate card transactions, vendor invoices, and accounts payable for 85,000 businesses on the same platform. That means it can correlate AI model expenditure with business outcomes — the deal a sales AI helped close, the invoice an AP agent approved — in ways a standalone router cannot. This cross-category correlation is the practical basis for measuring AI return on investment, rather than just tracking AI spending in isolation.

Why did Ramp's valuation grow from $32 billion to $60 billion in less than a year?

Three operating metrics drove the repricing: revenue crossing $2 billion annually (up from $1.5 billion four months earlier), a 175 percent year-over-year increase in annualized transaction volume to $275 billion, and customer count growing from 50,000 to 85,000 since January 2026. The acceleration of revenue growth — running 25 percent faster than the prior year — signals that Ramp's expansion into AI-native workflows and enterprise accounts is compounding its core business rather than cannibalizing it. The departure of Brex from the independent company landscape, following Capital One's acquisition, also narrowed the competitive comparison set for investors.

Is Ramp planning to go public?

Ramp has not announced an IPO timeline or filed any public offering documents as of this report. Multiple media outlets have noted that IPO conversations have occurred, and the company's metrics — $2 billion in revenue, positive free cash flow, 85,000 customers — would support a public filing. The October 2026 round extended Ramp's runway and reduced any urgency a cash constraint might have created, suggesting the company retains optionality on timing.