New York-based CarbonAI has raised approximately $12 million to develop AI agents that automate underwriting, due diligence and loan monitoring for commercial real estate lenders.
The funding was reported by Axios on October 15, with Cowboy Ventures among the investors. Other backers include Timber Grove Ventures, Bling Capital and Max Ventures.
The $12 million figure represents funding reported to date. The sources reviewed do not establish that the entire amount was raised in a single newly closed round.
CarbonAI was co-founded by CEO Jeff Saul and Josh Kaufman, who bring experience in commercial real estate lending.
What CarbonAI Actually Automates
Commercial property lending requires analyzing extensive financial and legal documentation.
Before financing a property, lenders review rent rolls, leases, operating statements, property-level expenses, borrower information and market assumptions.
These documents may arrive in incompatible formats, and important information can be buried across hundreds of pages.
Underwriters must convert that material into financial models to estimate property income, debt service capacity and credit risk.
CarbonAI uses agents designed specifically for those workflows.
Its system extracts relevant information, validates figures against source documents and populates underwriting models according to a lender's established templates.
It also prepares investment memoranda and assists with transaction due diligence.
The objective is not simply summarizing documents.
CarbonAI attempts to execute complete sequences of financial work and present finished outputs for human review.
From Underwriting to Asset Surveillance
After a loan closes, lenders must monitor the underlying property.
Changes in rent collections, vacancy, operating expenses or debt covenants can affect the risk of a loan.
These reviews are often performed quarterly or when significant information becomes available.
CarbonAI extends its technology into ongoing asset surveillance.
The system can process updated financial statements, compare performance with previous periods and prepare standardized reports.
It also monitors required documentation during the lending process.
According to CarbonAI, its platform has processed more than 40,000 documents.
The company says its outputs preserve a traceable connection between extracted figures and the underlying source material.
That feature matters because loan committees must be able to verify financial assumptions before approving transactions.
Customers and Commercial Position
CarbonAI identifies financial organizations including Greystone, Lightstone, Fulcrum and SitusAMC among its customer relationships.
Greystone has publicly described using the platform to improve due diligence collection.
The company's website also reports SOC 2 Type II compliance and says customer data is not used to train AI models.
These controls are particularly important when software processes confidential borrower and property information.
CarbonAI has not disclosed annual recurring revenue, customer concentration or average contract values.
Competition in Real Estate Finance Software
CarbonAI competes with established commercial real estate software and newer AI document-processing companies.
Its differentiation is its focus on end-to-end lending workflows rather than individual document extraction tasks.
Generic AI tools may be capable of reading a financial statement, but commercial lending requires specialized treatment of operating income, loan covenants, property valuations and transaction assumptions.
CarbonAI is betting that domain-specific agents can perform these workflows with less customization and more reliable outputs.
However, incumbent software platforms already occupy important positions within lenders' technology environments.
CarbonAI will need to integrate with those systems while maintaining detailed audit trails.
The Accuracy Problem
Commercial real estate lending involves consequential financial decisions.
An incorrectly classified expense, missing lease obligation or mistaken cash-flow assumption can affect a loan's underwriting.
Even when AI automates document processing, lenders remain responsible for validating the information and making credit decisions.
CarbonAI's product must therefore demonstrate low error rates, reliable exception handling and consistent performance across different property types.
Its opportunity is to make commercial real estate finance teams more productive. Its next test is proving that agents can complete complex financial workflows accurately enough to become part of institutional lending operations.