Fearn Raises $5.5M to Build an AI-Native Patent Firm for Startups

San Francisco-based startup Fearn has raised $5.5 million in funding to build what it describes as an AI-native patent prosecution firm designed specifically for startups.

The round was led by Kindred Ventures, with participation from a16z Speedrun, Designer Fund, and Essence VC.

Fearn is taking a different approach from AI companies that sell patent drafting software to existing law firms. Instead, the company is building a full-service patent firm around its own artificial intelligence infrastructure, combining AI software with human patent experts.

The company was founded by Han Kim and Angela Gao, bringing together legal and artificial intelligence experience from Big Law and Google Research.

Building a patent firm around AI

Patent applications can require significant amounts of legal work.

According to Fearn, a conventional patent application can consume 30 to 40 hours of lawyer time and cost between $18,000 and $40,000.

The startup believes AI can fundamentally change that cost and time structure.

Rather than developing another software product for lawyers, Fearn has created its own patent firm and uses its technology internally to handle the drafting and prosecution process.

The company's platform is called FearnOS.

The system is designed to treat a patent as more than a conventional document. Fearn describes it as a graph that connects individual claims with the figures and technical material supporting them.

The company says this architecture allows its system to move from a founder explaining an invention to a review-ready patent draft in as little as 24 minutes.

The approach is intended to reduce the amount of manual work required during the early stages of patent preparation.

Human experts remain part of the process

Despite its AI-native model, Fearn is not removing humans from the patent process.

Every application is reviewed by a human patent expert before it is filed.

The company says these reviewers include former Big Law patent professionals who evaluate the AI-generated work before submission.

That human review layer is important because patent prosecution involves more than simply generating technically accurate text.

Applications must meet legal and procedural requirements and ultimately interact with patent examiners.

Fearn's model therefore combines automated drafting and analysis with professional review.

The company says provisional applications can typically be filed within three business days.

Fixed pricing instead of traditional legal billing

One of Fearn's biggest differences is its pricing structure.

Rather than charging traditional hourly legal fees, the company uses fixed prices.

A provisional patent application costs $2,500.

A non-provisional application costs $9,000, including USPTO fees.

Fearn also says it puts its own fee at risk through a guarantee attached to its non-provisional service.

If a non-provisional application comes back with no allowed claims, the company says it refunds the $9,000 fee.

The startup reports gross margins above 80%.

This pricing model is designed to make the financial benefit of AI visible to customers.

The company's argument is that if AI reduces the amount of work required to prepare a patent application, customers should see that efficiency reflected in both turnaround time and price.

Founders combine legal and AI backgrounds

Fearn's founding team combines patent law experience with artificial intelligence research.

Co-founder and CEO Han Kim previously worked on patent drafting and prosecution at Morrison & Foerster.

He also began a PhD in computational neuroscience at Caltech before leaving to build Fearn.

Co-founder Angela Gao completed a PhD in AI at Caltech and previously worked on models at Google Research.

The combination of legal and technical experience forms an important part of Fearn's AI-native approach.

Rather than applying a general-purpose AI system to an existing law firm workflow, the company is building its legal operations and software infrastructure together.

Competing with AI patent software companies

Fearn is entering a growing market for AI-powered intellectual property tools.

Other startups are taking different approaches to applying artificial intelligence to patent work.

Solve Intelligence raised a $40 million Series B in December 2025, bringing its total funding to $55 million, according to the source.

The company sells AI-powered patent drafting and prosecution software to law firms and in-house legal teams.

Patlytics also raised a $40 million Series B in April 2026, bringing its total funding to approximately $65 million.

Its platform covers areas including invention harvesting and litigation support.

These companies primarily provide technology to existing legal teams.

Fearn's strategy is different.

Instead of selling software to lawyers and allowing law firms to incorporate AI into their existing operations, Fearn wants to operate as the legal team itself.

That distinction is central to the company's business model.

Selling the service, not just the software

The difference between Fearn and traditional legal technology startups is therefore not simply the AI technology.

Fearn is combining the software, lawyers, pricing structure, and customer relationship into one service.

A startup founder does not necessarily need to purchase software and then find a patent lawyer to operate it.

Instead, Fearn aims to handle the patent process directly.

The company believes this structure allows it to capture more of the efficiency created by AI and pass part of that benefit to customers through fixed pricing and faster turnaround.

Kindred Ventures founder and managing partner Steve Jang highlighted this combination of technology, legal expertise, and business model when discussing the investment.

The challenge ahead

Fearn's model is still entering an early stage.

The company now needs to demonstrate that its AI-native operating model can perform consistently as the number of patent applications increases.

Patent prosecution involves interactions with patent examiners and can require multiple rounds of amendments and responses.

That means the company's ability to generate an initial draft quickly is only one part of the overall process.

Its fixed-fee structure also creates a specific operating challenge.

As the volume and complexity of patent applications increase, Fearn will need to maintain the economics behind its pricing while continuing to provide human review and legal expertise.

The company's $9,000 guarantee for non-provisional applications makes those economics particularly important.

A new model for startup intellectual property

Fearn's $5.5 million funding round gives the company capital to expand its AI-native patent firm and develop FearnOS further.

The company is betting that startups will increasingly expect legal services to work more like modern software products: faster, more transparent, and based on predictable pricing.

Its approach combines AI automation with professional patent expertise rather than attempting to replace lawyers entirely.

For startup founders, intellectual property can be an important part of protecting technology and building long-term enterprise value, but traditional patent processes can be expensive and time-consuming.

Fearn is attempting to change that equation by building the legal service around AI from the beginning.

The next test will be whether its 24-minute drafting workflow, fixed-fee pricing, and human review model can scale from an early-stage operation into a high-volume patent firm while maintaining the quality required for real-world patent prosecution.