Lagrange Labs

This Sunday in Hong Kong 🇭🇰, Lagrange is hosting the Applied AI Summit with Amazon Web Services (AWS) and @metisl2. Expect packed panels and practical conversations on deploying AI in production — monitoring, failure handling, governance, and infrastructure choices that matter once AI leaves the lab. February 8 3:00–6:00 PM 🎟 Limited capacity. 📍 Hong Kong More information and to RSVP:  https://luma.com/3loqk0iz

2025 was a pivotal year for Lagrange. We set out to transform cryptography from theory into strategic infrastructure, bringing verifiable AI to defense, aerospace, and government systems where correctness cannot be assumed. Read our year in review below: https://lnkd.in/eXXuGgUx

Lagrange is pioneering cryptographic verification for defense, aerospace, and government ecosystems. Our mission is to transform cryptography into a strategic asset for the systems that protect us. Modern defense platforms are increasingly defined by autonomy, AI-driven decision-making, and software-defined workflows across surveillance, mission planning, communications, and command-and-control. As these systems take on greater responsibility, performance alone is no longer sufficient. Mission systems must be able to prove correctness, integrity, and compliance at operational speed. Anduril: Within *Anduril’s Lattice SDK™, Lagrange integrated DeepProve into an autonomous reconnaissance pipeline. Each stage—proximity detection, tactical classification (EVADE / INTERCEPT / SURVEILLANCE), and movement calculation—emits a cryptographic proof alongside the decision, demonstrating that autonomy and verification can coexist without sacrificing operational tempo. *Anduril Lattice™ and Lattice SDK™ are properties of Anduril™ Raytheon: Lagrange is now a registered supplier within Raytheon's vendor network, making DeepProve available as a production-ready capability for cryptographic verification of mission data and AI-enabled systems. By attaching proofs to sensor fusion, telemetry, and onboard AI outputs, DeepProve helps verify that inputs were untampered, firmware was uncompromised, and safety constraints were respected—strengthening certification and supporting safer fielded systems. Oracle: Lagrange joined the Oracle Partner Network to build verifiable AI use cases on Oracle Cloud Infrastructure (OCI) sovereign and mission-cloud environments designed for defense, aerospace, and government workloads. By extending DeepProve into OCI’s high-assurance environments, Lagrange enables cryptographic verification of AI inference, sensor provenance, and mission workflows *without* exposing sensitive data or proprietary logic. Lockheed Martin: Lagrange Labs joined Lockheed Martin’s vendor ecosystem as a registered supplier, supporting AI assurance and Zero Trust data integrity across aerospace and defense programs. DeepProve provides cryptographic evidence that AI models used in avionics, mission planning, and ISR fusion operated on approved data and within authorized constraints. General Dynamics: Lagrange Labs joined the GeneralDynamics supplier ecosystem to develop cryptographic infrastructure to support AI assurance and Zero Trust data integrity across government, national security, and defense applications. DeepProve extends Zero Trust principles beyond access control by proving that analytics and AI-driven outputs were executed correctly on authorized inputs, supporting secure communications, C4ISR, and automated mission planning across ground and naval platforms. Read more about our work below: https://lnkd.in/eEGJFFpk

Accountable autonomy will become the standard for AI systems operating in mission-critical environments, but what do we mean by "accountable autonomy"? Autonomous systems are accelerating across defense, aerospace, and national-security operations. Drones decide flight paths, ISR systems fuse sensor streams, mission software recommends actions. But autonomy without verification is uncontrolled decision-making. Accountable autonomy means the system doesn’t just act, it can prove it acted correctly. Not “we believe the model behaved properly" and not “an operator reviewed the logs afterward.” Rather, cryptographic evidence that the model followed authorized logic on unaltered data. This is where Lagrange's DeepProve comes in: Proof ≠ probability. Proof = mathematical certainty. Every AI inference can emit a proof verifying that: • The model executed correctly • Inputs were authentic • Outputs respected constraints • No sensitive data was exposed Why does this matter for defense? Because modern mission systems can no longer rely on implied trust, manual review, or unverifiable ML pipelines. Accountable autonomy gives operators confidence at machine speed. e.g. A model classifies or recommends → DeepProve attaches a cryptographic proof → downstream systems verify immediately. It also enables coalition trust without data sharing and strengthens lifecycle integrity: • Partners can verify each other’s outcomes without exchanging telemetry, model weights, or classified inputs. • From development → deployment → after-action review, proofs create a tamper-evident record of system behavior. Put simply: Autonomy = the system can act. Accountable autonomy = the system can prove it acted correctly. Accountable autonomy is the emerging standard for defense AI and the foundation Lagrange is pioneering with DeepProve.

Lagrange Labs Listed in the Vulcan-SOF Technology Portal We’re excited to share that Lagrange Labs is now listed in the Vulcan-SOF marketplace and technology portal, expanding access to DeepProve, our zero-knowledge proof framework for verifiable AI. This listing increases our visibility across the Special Operations community—including operators, mission enablers, program managers, and acquisition teams exploring technologies that strengthen SOF modernization, data integrity, and AI-assurance. DeepProve brings cryptographic verification to AI outputs and mission data pipelines, supporting use cases where correctness, provenance, and safety must be mathematically proven. Its capabilities align with key defense priorities such as AI assurance, Zero Trust data integrity, and resilient autonomy in sensitive or contested environments. As AI becomes central to mission planning and operational decision-support, the need for provable trust is growing. Our inclusion in the Vulcan-SOF ecosystem marks another step toward enabling secure, responsible, and verifiable AI across national-security missions.

Lagrange Labs is now an approved supplier within General Dynamics’ Enterprise Supplier Portal, enabling teams across the organization to discover and evaluate DeepProve. General Dynamics delivers mission systems across land, sea, air, cyber, and C4ISR—domains that depend on trustworthy data and verifiable correctness. Our inclusion reflects rising demand for technologies that can mathematically prove the integrity, provenance, and safety of AI-enabled defense workflows. DeepProve, our zero-knowledge proof framework, provides cryptographic verification across defense data and AI pipelines, proving that computations were performed correctly on authorized inputs. This strengthens resilience against spoofed telemetry, corrupted internal systems, and adversarial interference. In secure communications and C4ISR environments, DeepProve ensures messages, sensor outputs, and analytic products originate from validated processing chains and remain untampered throughout their lifecycle. For ground and naval platforms, cryptographic proofs offer a tamper-evident record of how each mission-planning decision was generated, verifying adherence to approved constraints without exposing sensitive models. Across the defense industrial base, these proofs support program compliance and integration by enabling auditable vendor handoffs, integrity checks across development stages, and clear evidence for accreditation and mission-readiness reviews. As autonomous and AI-supported systems expand across defense, DeepProve ensures they don’t just function—they can prove they functioned correctly. Our listing within General Dynamics’ supplier ecosystem helps advance accountable, secure, and verifiable autonomy for mission-critical applications.

With AI powering decisions across the defense stack, Lagrange is laying the foundations for cryptographic supremacy to bring verifiable AI to autonomous defense systems. Let’s explore what “cryptographic supremacy” means in the age of Defense AI. 2/ For autonomous defense systems to truly give nations a tactical edge, decisions made by AI must be provably correct, secure, and auditable. This means leveraging zk cryptography as a superpower. 3/ With DeepProve, we can embed cryptographic guarantees for AI Defense systems across drones, ISR pipelines, cybersecurity operations, and other defense tech. This turns cryptographic proof into a powerful deterrent for the tactical edge of defense. 4/ Learn more about how Lagrange is building the new foundation of American strength with DeepProve, the leading zkML for verifiable Defense AI. https://lnkd.in/eYtu5yaJ

Lagrange is now a registered supplier in Raytheon Technologies’ vendor network, allowing Raytheon teams to discover and evaluate Lagrange's DeepProve as part of their supplier ecosystem. Raytheon develops the sensing, missile, and weapons platforms that underpin U.S. and allied security, including Patriot and Tomahawk. Lagrange's DeepProve ca n enable cryptographic verification of mission data & AI-enabled systems across defense platforms. As these systems incorporate AI for detection, targeting, and engagement, the need for verifiable correctness and integrity continues to grow. Many sensor-fusion and fire-control systems depend on precise, trustworthy data. DeepProve attaches mathematical assurance to every AI-driven output, proving that decisions reflect the intended logic path and unaltered inputs. Every inference can carry its own proof of correctness: Defense platforms also rely on complex software, firmware, and telemetry pipelines that must remain uncompromised. Integrating proofs into these pipelines helps reduce the risk of tampered inputs or corrupted firmware, strengthening resilience across distributed operational environments. For airborne and maritime systems, DeepProve provides verifiable evidence that onboard AI adhered to approved safety constraints and mission envelopes *without* exposing sensitive model parameters or proprietary data. Operators gain a durable, cryptographically validated record of model behavior. Defense programs must meet rigorous safety and certification requirements. DeepProve strengthens certification dossiers by providing reproducible proof that: • Mission data was processed correctly • Subsystems followed authorized constraints • No tampering occurred in upstream data flows By embedding proofs directly into mission workflows, telemetry infrastructure, and onboard AI logic, DeepProve can add a new layer of reliability and compliance assurance for Raytheon’s fielded systems. Systems don’t just perform, they can prove they performed correctly. Joining Raytheon’s supplier network marks a meaningful step toward Lagrange's mission to support the long-term goal of bringing accountable autonomy to defense technologies.

Lagrange Labs is now a registered supplier within Lockheed Martin's vendor ecosystem, positioning Lagrange to support AI assurance and Zero Trust data-integrity initiatives across aerospace and defense programs. Lockheed Martin, responsible for the F-35 Lightning II, F-16 Fighting Falcon, and C-130 Hercules, increasingly relies on machine learning across avionics, mission planning, and ISR analysis. As AI becomes embedded in these systems, the ability to verify AI behavior is now mission-critical. Lagrange’s DeepProve attaches a cryptographic proof to every AI model output. This replaces assumption with certainty. DeepProve proves that: • The model used authorized, unaltered inputs • Outputs followed approved safety boundaries • The inference path is reproducible and auditable • No parameters or sensitive telemetry were exposed This unlocks new possibilities for AI assurance and Zero Trust data-integrity: • Avionics & Flight-Critical Systems: For avionics and operator-in-the-loop workflows, correctness must be verified at machine speed. DeepProve ensures every model action can be proven, not guessed. • Mission Planning: Mission-planning tools must show that outputs were produced using approved logic and constraints. DeepProve generates tamper-evident records that strengthen certification, verification, and accreditation. • ISR Fusion: ISR systems depend on sensitive, multi-source data. DeepProve preserves confidentiality while proving the output was produced correctly, critical for intelligence workflows. • Coalition Operations: Coalition missions require trust without sharing classified telemetry. DeepProve enables partners to exchange validated outcomes, not raw data, maintaining security while enabling interoperability. • After-Action Review & Lifecycle Traceability: DeepProve verifies correctness during testing, binds proofs to outputs in deployment, and provides reproducible evidence for after-action review. This creates continuous operational integrity from development → deployment → audit. Becoming a registered supplier in Lockheed Martin's vendor ecosystem advances Lagraneg's mission to scale accountable autonomy for defense. Learn more at the blog here: https://lnkd.in/eTFgXVgc

We had a great time at this week's San Diego Cyber Center of Excellence (CCOE) event, bringing together industry, government, and defense leaders advancing cybersecurity and AI modernization. Pictured: Our Head of BD Brian Novell representing Lagrange as one of CCOE’s newest members alongside Andy Haass, Director at Booz Allen Hamilton (left), Mansi Thakar from NVIDIA (upper right), and Lisa Easterly, President & CEO of CCOE (lower right). ICYMI: Lagrange joined CCOE to support San Diego companies and communities, advancing a shared mission: bringing cryptographic verification and verifiable AI into defense, aerospace, and critical infrastructure. CCOE stands as one of the country’s strongest catalysts for defense innovation. Being part of this network allows Lagrange to collaborate directly on: • Zero Trust for AI • Mission data integrity • Trusted autonomy • National security innovation Lagrange complements CCOE’s initiatives by embedding cryptographic assurance into AI systems, proving that models behave as intended, that inputs aren’t tampered with, and that mission-critical pipelines remain verifiable end-to-end. This is the foundation of verifiable defense AI.