How Cursor Passed a Clean Big 4 Audit

Core results: 

‍

• $4B in ARR supported by three-person Revenue Accounting Team 
• Clean Big 4 Audit Opinion GAAP-compliant self-serve revenue from day one of the audit
• $1M+ in free trial abuse recovered through systematic Stripe fraud identification
• CY24 restatement completed in under two months substantially below the external consultant costs

‍

Background

‍

When Kirill joined Cursor as their first dedicated accounting and finance leader, the company was already one of the fastest growing companies in history, 75 employees, a product that had captured the developer market, and the self-serve revenue engine processing tens of millions of Stripe Transactions each month. The accounting and finance function was being built up from the ground up, intentionally and quickly.

‍

The revenue accounting team mandate was clear: establish GAAP-compliant reporting at scale, support and pass an imminent Big 4 audit, and do it without adding the headcount the business didn’t yet need.

‍

The Problem

‍

Cursor’s self-serve revenue stream runs entirely through Stripe; including monthly and annual subscriptions as well as usage-based. At the transaction volumes Cursor operates at, Stripe reporting starts to break down. Summary level exports were the only practical output, which was insufficient for ASC 606 compliance, audit support and month-end close. 

‍

The team's options were straightforward: build internally, use the generalized AI approach or implement a purpose built tool.

‍

An internal build was seriously evaluated. At that time, the engineering investment required to handle Cursor and Stripe API complexity and transaction volume was better directed toward the product — with the expectation that the in-house solution would follow in time.

‍

Generalized AI was also considered and set aside. Revenue recognition requires deterministic, auditable logic. Cursor uses AI extensively across operations and accounting, but applying it to a high-volume, rule-intensive accounting workflow where a wrong answer has audit and financial statement consequences and wasn't the right direction at that time. 

‍

HubiFi’s tools provided what was needed: a deterministic rules layer on top of Stripe's transaction data, purpose-built for the revenue recognition use case.

‍

Audit-Ready Reporting, Delivered Quickly 

‍

“Hubifi provides the deterministic layer on top of the Stripe data that enables us to investigate and resolve issues, reconcile and close the books quickly .” 
- Will Smerdon, Head of Revenue Accounting & Order to Cash at Cursor

‍

Working with HubiFi, Cursor's revenue accounting team established:

‍

• ASC 606-compliant recognition at the individual transaction level, across Stripe's ~35 endpoint types and transaction states
• Automated three-way reconciliation across Stripe, HubiFi, and NetSuite — replacing a manual, error-prone close process
• A two-day self-serve close with full ability to investigate and substantiate the numbers
• Fraud detection workflows that identified and corrected free trial abuse, now monitored on a monthly basis
• Audit-ready revenue reporting the foundation for a clean Big 4 opinion

‍

Bringing on deterministic automation to provide accurate Stripe reporting was a game changer. With HubiFi, the team can trust their reported numbers and prepare for a clean, fast audit.

‍