Lasso launches AI guardrail LEAP & raises USD $30m

Lasso Security has launched LEAP, a CPU-based AI guardrail for monitoring AI requests and actions, and disclosed a USD $30 million funding round.

The product is designed to inspect AI traffic without relying on graphics processing units, which are commonly used to run large language models for security checks. LEAP is already in production with large corporate customers and parts of the US federal government, according to Lasso.

The funding round was led by ClearSky, with participation from Entrée Capital, iAngels, Singtel Innov8, Mindset and Swish Data. Entrée Capital previously led Lasso's seed round and increased its investment in the latest financing.

Lasso described LEAP as a transformer-free system that can make a decision in under five milliseconds and handle a far higher volume of traffic than existing guardrails. It said the product is intended to lower the cost of screening AI requests by avoiding the need to run a reasoning model on every interaction.

The approach targets a growing issue for companies deploying AI agents in customer-facing and internal workflows. Security checks based on large models can become expensive when every prompt or action must be reviewed, especially in systems that run continuously and at scale.

According to Lasso, many companies respond by inspecting only part of their AI traffic, leaving some activity unchecked and increasing operational risk.

"Two years ago, the question was whether enterprises would put AI in front of their customers. That question is settled," said Elad Schulman, Co-Founder and Chief Executive Officer, Lasso Security. "What they need now is security that keeps up with agents that act autonomously and at scale. This is what LEAP was built for -- inspecting not only what agents say, but also what they do, at enterprise scale and cost-effectively."

How it fits

LEAP is one of two engines in the company's platform. The second, called RAPID, is a self-hosted system used for the smaller share of decisions that require more complex policy judgment in natural language, according to Lasso.

Traffic is routed first to LEAP for an immediate decision, while cases requiring deeper analysis are escalated to RAPID. In effect, Lasso is trying to reserve more computationally intensive review for a narrower set of situations rather than applying it to every AI request.

Ophir Dror, Co-Founder and Chief Product Officer, said existing approaches are poorly suited to continuous monitoring. "The people building security at the largest scale have started saying out loud what the architecture makes obvious: reasoning models are expensive and built for deliberation, not for the continuous, real-time and low latency monitoring that real AI security demands," said Dror.

He also criticised the idea of using one agent to oversee another when costs are counted at production scale. "The concept of a guardian agent or agents overseeing other agents is great, unless you are the one paying the bill," said Dror. "In real world deployments with millions of prompts and actions, this idea is just not feasible."

Customer base

Lasso said it sells AI security tools to customers in financial services, insurance, hospitality and automotive, as well as to the US federal government. It named BMW, Leonardo Defence, Kaufman Rossin, eToro and Fiverr among organisations using its technology, and said its systems protect tens of thousands of agents and process billions of requests and actions each month.

Revenue grew by more than 500% over the past 12 months, according to the company, which did not provide a revenue figure.

Lasso's pitch is based in part on offensive security work. Its red-teaming group tests customer AI systems with attacks ranging from single prompts to multi-turn autonomous adversaries, then feeds those results back into product development, the company said.

In one engagement with a global healthcare provider, Lasso said its team extracted patient records and altered prescriptions delivered to patients. It used that example to argue that AI systems need monitoring that covers both language outputs and agent behaviour.

Customers and partners pointed to visibility and operational oversight as key factors. "Lasso gives KR the visibility into how Artificial Intelligence is being used across the Firm, which enables faster security decisions with empirical data," said Vladislav Rudnitsky, Chief Information Security Officer, Kaufman Rossin. "That visibility allows us to enable practical guardrails to support data loss prevention and in-line feedback to professionals as they work."

Protiviti also described the relationship in operational terms. "Working with Lasso has been a genuinely collaborative experience," said Kim Bozzella, Global Leader of CIO & CISO C-Suite Solutions, Protiviti. "Its platform allows us to deliver AI security to our clients at scale, backed by a team that's responsive, technically sharp, and easy to build with. We're excited to keep building on this momentum together."

ClearSky's Ryan Crum linked the investment to the spread of AI agents across large companies. "AI agents are going into production across every major enterprise, and most of them have never been properly tested for security. Lasso fixes that. As a former CISO, that's a problem I've felt firsthand. ClearSky led the round because this is exactly the kind of emerging attack surface we invest in," said Crum.