Introduction

Today's most ambitious manufacturers — particularly in defense, aerospace, and advanced industrial sectors — are racing to ramp production at software speed. Designs evolve weekly, products are increasingly modular and configurable, and the pressure to scale output without scaling headcount has never been greater.

Yet on most factory floors, the connective tissue between engineering intent and physical production still depends on static, manual, document-driven processes that were never designed to keep up. McKinsey finds that targeted improvements to existing defense industrial operations could cut production ramp-up times in half and double new capital efficiency. However, the industry continues to struggle with a younger, less experienced workforce where difficult skilled trades take significant time and training to master (read more). 

The result is a coordination bottleneck. Engineering teams move quickly, but every design change ripples into hours or days of manual rework: reconciling CAD models, regenerating work instructions, updating routings in MES, and validating what is actually being built on the line. At scale, this consumes thousands of engineering hours per week that should be going into improving throughput, quality, and ramp speed. The downstream cost is staggering: Bain reports that aerospace and defense program cost overruns have surged to nearly $46 billion, while delivery timelines have stretched from 8 years to 11 years (read more).

A recent BCG global survey of nearly 1,800 manufacturing executives found that while 89% of companies plan to implement AI in their production networks, only 16% have achieved their AI-related targets (read more). To close that gap, Dirac and Amazon Web Services (AWS) are partnering to bring AI-driven, model-based production planning to advanced manufacturers. Through this partnership, Dirac’s BuildOS platform — the AI-driven system of record for model-based production planning — is delivered on AWS infrastructure, including AWS GovCloud (US) for customers handling Controlled Unclassified Information (CUI), ITAR, and other export-controlled data.

In this post, we’ll share how Dirac’s BuildOS, hosted on AWS, helps manufacturers replace document-driven production with an automated, model-based system, as well as how customers like Anduril are already using it to compress work instruction authoring by nearly 90%.

Dirac is a manufacturing technology company building the AI-driven system of record for production planning. Its mission is to rebuild the industrial capacity of the West by turning manufacturing facilities into context-aware, adaptive, dynamic environments.

The Manufacturing Coordination Bottleneck

In manufacturing, work instructions are the atomic unit of information on the factory floor. They are the step-by-step guidance that turns engineering intent into physical production: defining parts, tools, sequences, and fit. When products were stable and product lines changed slowly, static work instructions were workable. That world is gone.

Today, advanced manufacturers face a structurally different challenge:

• Products are increasingly modular, configurable, and updated continuously
• Design changes happen at software cadence, not annual release cadence
• Manufacturing engineers spend a significant portion of their time reconciling CAD, MES routings, and floor reality instead of improving the line
• Document drift between engineering and production introduces quality risk and slows ramp

Manufacturers facing this pattern are forced into a choice: slow design to protect manufacturing, or fundamentally rethink how the coordination layer works. For companies competing on speed, particularly in aerospace and defense, slowing down is not an option.

Snapshot of Dirac’s BuildOS

Rather than treating work instructions as static documents, Dirac's BuildOS maintains a living, model-based representation of the product, the factory, and the relationship between the two. Geometry, structure, variants, physics, assembly logic, stations, tools, and material flow are represented in a single system. Work instructions are then derived automatically from this model, thus significantly reducing the need for manual authoring and re-authoring.

When the design changes, instructions update automatically. There is no painstaking re-authoring, no document drift, and no manual reconciliation between systems.

BuildOS is built on AI that deterministically reasons over structured manufacturing data by interpreting CAD geometry, inferring assembly steps, detecting dependencies, and propagating engineering changes through production plans. Beyond work instructions, BuildOS provides continuous Design-for-Manufacturability (DFM) feedback and earlier insight into cost, tooling, and throughput constraints.

AWS Powers and Secures Dirac BuildOS

Manufacturers, especially those operating in regulated environments, need production processes to be highly available, secure, and compliant from day one. That is why Dirac is powered by AWS.

Running on AWS gives Dirac and its customers:

• Global, elastic infrastructure that scales with manufacturing demand without requiring customers to deploy or operate underlying infrastructure.
• AWS GovCloud (US) to support customers handling CUI, ITAR, and other export-controlled data, with FIPS-validated encryption.
• A mature security and compliance baseline, enabling Dirac to maintain SOC 2 Type II attestations, with controls aligned to NIST SP 800-171 and CMMC 2.0 Level 1 practices. (Dirac is not currently CMMC-certified; certification status will be updated as assessments are completed.)‍
• Defense-in-depth services — including identity, key management, network isolation, monitoring, and threat detection — so customers can deploy quickly while meeting their internal security requirements.

Dirac’s regulated environment is hosted in AWS GovCloud (US) and is used today by multiple defense contractors. CAD files, BOM data, factory models, and production plans are encrypted in transit and at rest, with keys managed through AWS-native key management services. This combination is what allows Dirac to stand up regulated environments in days rather than the months traditionally associated with bringing new tooling into a defense manufacturing program.

Eight Deployment Options, One Platform

Because manufacturers operate under widely different security, compliance, and IT requirements, Dirac on AWS offers eight supported deployment options. Each is built on AWS, and customers select the model that aligns with their data classification, compliance obligations, and internal infrastructure strategy.

This range matters because manufacturing customers do not all sit in the same place on the security and operations spectrum. A commercial industrial customer benefits from Dirac-managed SaaS on AWS Commercial. This offers fastest time-to-value and no infrastructure lift. A defense prime needing strict isolation, CUI support, and customer-owned operations can deploy Self-Hosted in their own AWS GovCloud (US) account. Every option in between exists for customers whose requirements fall in the middle.

For most regulated customers, Dirac typically recommends starting with SaaS on AWS GovCloud. It minimizes IT lift while meeting compliance requirements, and gets customers to production value quickly.

Business Benefits for Manufacturers

Manufacturers deploying BuildOS see impact across several critical dimensions:

• Automated authoring: Work instruction authoring time is dramatically reduced, freeing manufacturing engineers from documentation work.
• Engineering changes propagate in minutes, not days: When a design update lands, the production plan updates with it.
• Earlier DFM feedback: Cost, tooling, and throughput constraints are surfaced before production begins, reducing ramp risk.
• Scalable throughput without scalable coordination overhead: Production scales without proportional growth in engineering headcount.‍
• Production lines that adapt at the speed of design: The factory becomes responsive to engineering rather than a constraint on it.

Customer Spotlight: Anduril Selects Dirac to Power AI-Driven Work Instructions

In January 2026, Anduril selected Dirac as its core partner for AI-driven work instruction authoring in a multi-year deal, after evaluating incumbent enterprise tools and internal builds (read more). Before Dirac, more than 100 Anduril manufacturing engineers spent roughly half their time manually authoring and updating work instructions, equating to thousands of engineering hours per week.

‍

Anduril COO Matt Grimm: "Dirac accelerates Anduril's sales"

“Every serious manufacturer eventually hits the same wall: engineering moves fast, factories move carefully, and coordination becomes the true bottleneck. Dirac is the only team that understood this problem from first principles and how to solve it implicitly. Dirac’s BuildOS is becoming a core enabler of Arsenal OS, Anduril’s digital software ecosystem of manufacturing technologies. With Dirac, Anduril’s factories can be even more adaptive, dynamic, reconfigurable, and context-aware. AI-driven work instructions are the key.”

— Matt Grimm, Co-Founder and COO, Anduril

‍

Deployed in an Anduril-hosted, ITAR-compliant AWS GovCloud environment, BuildOS was operational within days and delivered an 87.5% reduction in work instruction authoring time, therefore collapsing a 12-hour process into 90 minutes. Anduril is now rolling out Dirac enterprise-wide, standardizing BuildOS as core manufacturing infrastructure across the company.

‍

Anduril COO Matt Grimm: "Since implementing BuildOS, we've seen wild, wild improvements" & 87.5% reduction in time to generate factory-ready work instructions from CAD

“The Dirac team understands manufacturing at a system level. Work instructions are the atomic unit of information in a factory, and Dirac’s BuildOS is the first platform we’ve evaluated that actually models that reality correctly. Just as importantly, they execute at an extremely high bar. From both a capabilities and execution standpoint, they were the obvious choice.”

— Cy Sack, Head of Business Systems, Anduril

‍

Why This Partnership Matters

Advanced manufacturing is being redefined by speed, configurability, and the ability to adapt in real time. Achieving that requires automating the production planning layer of the factory where engineering intent meets physical production. Dirac is building that layer; AWS provides the secure, compliant, elastic foundation it runs on.

For manufacturers, the combined offering means:

• Accelerated time to value: Regulated customers can be in production on AWS GovCloud in days, not months.
• Compliance-ready foundation: SOC 2 Type II attestations, controls aligned to NIST SP 800-171 and CMMC 2.0 Level 1 practices, and FIPS-validated encryption.
• Deployment flexibility: Eight supported deployment models on AWS so customers can match the platform to their security posture and operating model.
• Room to scale: from pilot to multi-site, multi-program production without re-architecting.

For Dirac, AWS is the foundation that makes serving the most demanding manufacturers in the country possible. For AWS, Dirac extends the platform into one of the highest-leverage layers of industrial operations: the system that decides, in real time, how things actually get built.

Conclusion

Dirac and AWS are partnering to give advanced manufacturers — from commercial industrials to defense primes — a path off document-driven manufacturing and onto a live, AI-driven, model-based production planning layer. With BuildOS deployed on AWS, manufacturers can collapse coordination overhead, propagate engineering changes through production in minutes, and scale output without scaling headcount.

To learn more about Dirac and how BuildOS can fit into your manufacturing stack, visit diracinc.com or contact the Dirac team at contact@diracinc.com.

Dirac Spotlight

Dirac is building the AI-driven, model-based production planning platform that replaces document-driven manufacturing with a dynamic system, helping advanced manufacturers run as fast as their engineering teams design.

Contributing Authors

‍Vedanth Srinivasan is Head of Solutions Engineering & Design and Go To Market (GTM) at Amazon Web Services, where he partners with technology companies building the future of advanced manufacturing.

This post was co-authored by the Dirac team in collaboration with AWS.

‍