Lasso Security launches CPU-only AI guardrail, raises $30M

What happened

Lasso Security launched a new AI guardrail product called LEAP that runs solely on standard server CPUs, eliminating the need for GPUs. Alongside the launch, the company raised $30 million in fresh funding. LEAP acts as a secondary safety check by processing every AI request a production system receives, providing a layer of oversight without relying on costly or specialized hardware.

Why it matters

Most AI safety tools require GPUs to handle the heavy compute demands of scanning and moderating AI outputs, which inflates costs and complicates deployment. LEAP’s CPU-only design reduces hardware expense and broadens access for enterprises running AI at scale. Companies can add guardrails to AI workflows without a full GPU stack, which lowers the barrier to safer AI adoption. Additionally, having a dedicated safety layer that runs independently on commodity hardware limits risks tied to vulnerabilities inside the primary AI models or infrastructure.

What to watch next

It will be important to see how LEAP performs in real-world workloads where latency and accuracy of guardrails matter. The product’s effectiveness depends on balancing thoroughness with speed on CPUs alone. Watch if other AI safety startups follow Lasso’s chip-agnostic approach or if major cloud providers integrate similar CPU-based solutions. Investors and enterprises will also track how Lasso deploys its new capital to sharpen the product and expand market presence, especially against GPU-dependent incumbents.

AI Quick Briefs Editorial Desk