Stilta, a Stockholm and New York-based startup building agentic AI for patent enforcement and defense, announced a $10.5 million seed round on Tuesday led by Andreessen Horowitz. Y Combinator and individual investors from OpenAI, Legora, Sana, and Lovable also participated, according to Legaltech News.

The company’s platform deploys networks of AI agents that work in parallel to automate patent infringement analysis. A user inputs a patent number and relevant context, and the agents search for conflicting patents, flag similar IP that could apply, and pull filing and court histories. The output, according to CEO Oskar Block, is “litigation-grade: a report and claim charts with pinpoint citations to every piece of evidence,” as he told TechCrunch.

The Founders and the Problem

Stilta was founded earlier this year by four McKinsey veterans: Block, Petrus Werner, Tobias Estreen, and Oscar Adamsson. The idea originated when Estreen’s father, a patent attorney, described a workflow that hadn’t changed in thirty years, according to TechCrunch. Block had previously worked at an autonomous trucking company where he observed how manual and slow the patent process was.

The team participated in Y Combinator’s Winter 2026 cohort. Stilta intends to serve both enterprises and law firms, targeting the analytical bottleneck that has made patent litigation slow and expensive.

The Competitive Signal

“When one company starts using AI for patent enforcement, every competitor has to follow,” Block said in a press release. That framing positions patent agents not as a nice-to-have but as an arms race dynamic: once one company in a sector deploys autonomous IP enforcement, the cost of not doing so becomes untenable.

Block also pointed to a large pool of dormant IP. Many companies hold patents they’ve “never enforced, never licensed, never even analyzed properly because the cost of doing so was prohibitive,” he told TechCrunch. Lowering that cost barrier could unlock new licensing revenue streams for companies sitting on underutilized portfolios.

The startup joins a growing field that includes Solve Intelligence and DeepIP, per TechCrunch.

Andreessen Horowitz has been building a concentrated legal tech portfolio over the past 18 months. The firm has invested across multiple rounds in Harvey, backed plaintiffs firm AI developer Eve through two rounds, and led pre-seed funding for communication firewall provider ZeroDrift, according to Legaltech News. Y Combinator has been equally active, recently investing in Parlai, PointOne, and Legora.

The funds will go toward hiring engineers, patent specialists, and go-to-market team members in Stockholm and New York City, per Legaltech News.

The Agent Architecture Bet

The distinction Block draws is between AI as a search tool and AI as a collaborative team. Stilta’s agents “reason in parallel and converge the way a room full of specialists would, but at a scale no human team can match,” he told TechCrunch. The lawyer remains in the driver’s seat, guiding the analysis rather than ceding it.

That architecture choice matters. Patent enforcement is one of the few domains where AI agent output needs to survive adversarial scrutiny in court. If Stilta’s claim charts hold up under litigation pressure, it validates the broader thesis that agentic workflows can replace not just search and summarization, but the structured analytical work that has historically required teams of domain specialists.