Lagrange x NVIDIA: A Vision for Safe AI

Over the last year, artificial intelligence has evolved from a frontier technology to a global phenomenon. 78% of businesses reported using AI in 2024, up sharply from just 55% a year earlier. As AI becomes the backbone of every industry we interact with—from finance and defense to healthcare and consumer apps—so does one fundamental question: can we trust what AI systems are doing?

Trust in AI currently relies on policy and promise. In order to ensure that AI aligns with humanity’s best interests, the outputs of AI models must be proven.

Lagrange provides the cryptographic foundation for AI you can prove. Lagrange’s core technology, DeepProve, enables anyone to verify that an AI model ran correctly and produced the right results—without exposing the model or the input. This makes it possible to verify AI activity across the key industries that can benefit most from machine learning.

This month, Lagrange became the first company at the intersection of AI and crypto to join the NVIDIA Inception Program. This marks a major milestone for verifiable AI, and signals the direction that AI is headed.

“Every transformative technology needs its trust layer. For the internet it was https, for AI it’s cryptographic verification. Our collaboration with NVIDIA brings us closer to a future where AI systems are secure and trustworthy. By joining the NVIDIA Inception Program, we aim to make verifiability the standard for AI, enabling a future where safety and accountability are built into every AI system from the ground up.” — Ismael Hishon-Rezaizadeh, Co-Founder & CEO of Lagrange. 

Why This Matters: Lagrange’s DeepProve

AI systems are increasingly making high-stakes decisions: which transactions to block, which health diagnoses to suggest, which military targets to identify. Yet we still lack the technical standard needed to verify that these systems are functioning correctly.

Lagrange’s core product, DeepProve, provides the standard we need. DeepProve generates zero-knowledge proofs for AI inferences—meaning we can now mathematically prove that an AI ran on the right model and produced the correct results. 

This opens up new possibilities for AI, including:

• Safe deployment of powerful AI models in critical domains (health, defense, autonomous vehicles)
• Privacy-preserving inference for proprietary or regulated data
• On-chain AI execution with verifiability in decentralized environments

However, scaling verifiable AI requires serious compute power. That’s where NVIDIA comes in. Through NVIDIA Inception, Lagrange gains access to NVIDIA’s core technology, training resources, and ecosystem support to scale DeepProve across real-world use cases. With the world leader in AI compute, Lagrange plans to unlock high-value use-cases for verifiable AI.

Joining the NVIDIA Inception Program is just the beginning. Lagrange and NVIDIA share a bold vision for a world where AI is not just intelligent, but accountable—where trust is mathematically guaranteed and every AI system can prove that it is correct.

About Lagrange

Lagrange is a leader in zero-knowledge proof generation for safe and secure AI. Its flagship product, DeepProve, is the fastest zkML system in existence, enabling AI verification through zero-knowledge proofs via the Lagrange Prover Network. These capabilities are made possible by Lagrange's world-class research team, whose innovations in proving systems have made verifiable AI both practical and scalable for real-world applications.

About the NVIDIA Inception Program

The NVIDIA Inception Program helps startups during critical stages of product development, prototyping and deployment. Every Inception member gets a custom set of ongoing benefits, such as NVIDIA Deep Learning Institute credits, preferred pricing on NVIDIA hardware and software, and technological assistance, which provides startups with the fundamental tools to help them grow.