Accelerating Real-World Autonomy to Secure Global Market Leadership
How a National Certification Framework Scales Domestic Innovation
This past fall, we ran the inaugural Knudsen Industrial Strategy Fellowship, training the next generation of industrialists to navigate policymaking in Washington, D.C. Our inaugural cohort included an impressive line-up of founders, engineers, and investors who are building in the private sector but eager to learn the ins and outs of industrial strategy.
After 10 weeks of seminars, readings, and a December field trip to the capitol, we’re excited to share a series of policy memos written by our fellows. While this represents the first foray into policy for many of them, they bring valuable real-world experience to the table, identifying the specific challenges holding back the U.S. industrial base.
Summary
The United States leads the world in autonomy research but lacks a scalable path to commercial certification. Certification is critical because regulators and buyers need a clear, evidence-based way to assess the safety and reliability of autonomous systems (e.g., unmanned aerial systems, advanced robotics, and self-driving vehicles) before purchase. While American companies are hindered by a slow, bespoke regulatory environment, competitors like China are actively accelerating commercialization of frontier autonomy technologies and defining the global standards that will govern this market.
This proposal calls for a National Autonomy Certification Framework (NACF) to replace one-off exemptions with standardized testbeds and portable safety data, unlocking domestic growth and ensuring US regulatory leadership. The framework has three parts:
Standardized Testbeds: Shared physical or virtual infrastructure to validate performance.
Digital Safety Cases: A unified portable, data architecture for safety evidence.
Performance-Based Rules: Adaptive regulations that allow systems to scale based on proven past reliability.
Challenge and Opportunity
The Deployment Gap: The U.S. is the clear leader in early-stage autonomy research, but this leadership stalls when systems move from controlled lab or synthetic environments to real-world deployment. This “deployment gap” threatens vital sectors like logistics, transportation, and manufacturing. The friction is visible across domains: from flying drones beyond visual line of sight (BVLOS) to fielding unmanned maritime systems alongside manned vessels.
The Certification Bottleneck: A primary obstacle to scaled deployment is certification. Currently, most commercial BVLOS operations rely on rule waivers and regulatory exemptions rather than a codified national standard (FAA BVLOS 2025 report). This fragmentation creates severe regulatory drag. Recent FAA data shows that fewer than 1% of BVLOS waivers are approved annually (FAA Aerospace), underscoring a structural failure. This issue extends beyond aviation; maritime and industrial robotics face similar domain-specific hurdles where safety cases are not portable, and definitions of reliability vary by region.
Downstream Effects: This uncertainty creates a vicious cycle across the entire ecosystem.
Wasted resources: Companies burn development budgets re-running the same demonstrations to satisfy slightly different regulations, rather than deepening core capabilities.
Increased sector risk: Investors discount deployment potential because there is no predictable path from prototype to scale, making autonomy a riskier asset class.
Innovation stagnation: Emerging startups with strong technical innovations struggle to compete with incumbents who can afford the expenses of navigating protracted, bespoke certification processes.
The Global Threat: Global competitors are rapidly defining the rules and infrastructure that will govern autonomy. China has established dozens of national-level vehicle test zones alongside nearly 20 pilot cities, creating streamlined pathways for technical verification (UN Economic Commission for Europe 2023). Japan and Europe are pursuing similar strategies with unified maritime and automotive safety frameworks (UN Economic Commission 2020).
Across all three regions, the pattern is the same: federally sponsored centralized test infrastructure that produces standardized safety evidence, paired with clear certification pathways that guide private capital flow. Without a comparable US framework, American innovations will be forced to compete under standards set by others.
Plan of Action
To translate technical leadership into scaled deployment, the federal government should establish a National Autonomy Certification Framework (NACF). This framework links standardized testbeds, digital safety case data, and adaptive deployment rules to anchor global standards around US norms.
Create a network of standardized autonomy testbeds: A testbed is a single environment where autonomous systems can be evaluated using common rules, shared data formats, repeatable scenarios, and consistent simulation and instrumentation infrastructure. NACF should accredit a network of these facilities to replace the current fragmented landscape. Today, testing is siloed across private ranges, FAA sites, and university facilities, forcing developers to duplicate validation work. This approach mirrors the DoD’s Replicator initiative, which uses shared infrastructure and uniform benchmarks to evaluate systems quickly. The NACF would adapt this model for the broader needs of civilian logistics, infrastructure, and transportation, allowing developers to generate safety evidence once and have it trusted nationwide.
Establish a unified digital safety-case architecture. The NACF must standardize how autonomy systems demonstrate reliability. A digital safety case is a structured body of evidence (spanning simulation results, live test data, and subsystem validation) that holistically supports a claim that a system is safe. While the United States utilizes high-quality isolated standards (e.g., ISO 26262 for functional safety and UL 4600 for autonomy), it lacks a unified architecture to package this evidence. A NIST-hosted safety case registry would allow simulation traces, scenario coverage metrics, hardware-in-the-loop results, and real-world data to be submitted in a consistent format. This creates a transparent foundation for regulators and investors to evaluate system safety, and a single portable set of evidence for companies to reuse.
Implement performance-based deployment rules. The NACF should replace waiver-based approvals with adaptive rules. Agencies like the FAA and NHTSA would use NACF safety-case evidence to define specific performance thresholds (e.g., demonstrated interoperability in a detect-and-avoid scenario). Systems that meet these thresholds would automatically qualify for expanded operating envelopes, while those that fall short would be required to re-present evidence. This creates a data-driven regulatory loop that promotes technical evolution while preserving accountability.
While the NACF creates the regulatory rails, partnerships across government, industry, investors, and research universities will drive the velocity.
Universities and startups contribute populate testbeds with edge-case data and build shared scenario libraries.
Investors sponsor challenge programs and use NACF safety progression as a standardized diligence metric, de-risking investment in autonomy.
Government coordinates the standards to ensure consistency and authority.
Building the NACF will bridge the gap between the lab and the real world, transitioning the United States from a research pioneer to an industrial deployment power.
About the Author
Divya is a Product Manager at Applied Intuition, where she builds both advanced autonomy systems and the simulation, command-and-control, and evaluation tooling to deploy them across all domains. Her background spans startups, big tech, and incubators - most recently at Google X (Google’s Moonshot Factory) - which gives her a rare perspective on how breakthrough technologies move from prototype to large-scale deployment. She holds a BS and MS in Computer Science from Stanford. These views are her own.
 | A guest post by
|