$44 billion. That is the valuation the company commanded when it closed a $750 million Series F round this month — nearly triple the $16 billion it was worth less than a year ago.
The corporate spend management platform now processes over $100 billion in annualized spend across 50,000+ companies. But the number that matters more than the valuation multiple is this one: Ramp's AI agents now handle expense approvals with 99% accuracy, and Ramp Inspect — an internal coding agent — merges more than half of all production pull requests.
This is not a fintech company that happens to use AI. It is an AI-native financial operations platform whose product is the AI.
From hundreds of agents to one framework
The company began by letting individual teams build their own agents. The result was hundreds of specialized bots doing similar work in four different ways. After major model releases in early 2026, the company consolidated around a single agent framework with thousands of skills, unified through a conversational interface called Omnichat.
The architecture shift mirrors what every AI-native company discovers as it scales: general-purpose agents beat swarms of specialists when the underlying model is capable enough.
Ramp's Policy Agent is the clearest example. It reads expense policies written in natural language — "no first-class flights for domestic trips under six hours" — and enforces them automatically. The agent does not match keywords. It interprets intent. That is the difference between rules-based automation and agentic AI.
$42.6 billion in Q2 — and AI finance leads
The broader market validates the bet. AI funding hit $42.6 billion across 312 rounds in Q2 2026, with agentic AI startups capturing $20 billion — nearly half. Financial services is the largest buying center for agentic software, and the company occupies the intersection of AI infrastructure and finance operations.
The company's Applied AI Solutions division, launched June 10, embeds dedicated engineers with enterprise finance teams to automate workflows that have resisted automation for years: token spend management, accounts payable, procurement, close, accounts receivable.
What Ramp proves about AI agents
Three things. First, that production AI requires iterating from simple use cases upward — Ramp started with transaction classification before moving to multi-step agentic workflows. Second, that human-in-the-loop labeling sessions and extensive evaluation frameworks are not overhead but core infrastructure. Third, that the companies winning with AI are not the ones with the best models but the ones with the best context: vendor history, contract terms, GL codes, approval chains.
Context is the moat. OpenAI provides the reasoning. Ramp provides the data the reasoning operates on.
The $44 billion valuation reflects that distinction. Investors are not betting on Ramp's corporate card. They are betting on Ramp's ability to turn messy financial data into autonomous operations.