Blue Origin Builds AI Agent Platform for Lunar Hardware Design

Blue Origin deployed 2,700+ AI agents with 70% company-wide adoption, achieving a 90% reduction in hardware development time using Amazon Bedrock.

Impact

2,700+

AI agents deployed

70%

Company-wide adoption

3.5 million

Monthly interactions

90%

Hardware dev time reduction

6x faster

Analysis workflow acceleration

10x improvement

AI app development speed

Challenge

Specialized aerospace knowledge was siloed in expert minds, and traditional development cycles took years. Base LLMs lacked training in industrial manufacturing and specialized domains.

Solution

Built BlueGPT, a secure LLM gateway and multi-agent orchestration platform using Amazon Bedrock, Strands Agents SDK, EKS, OpenSearch, and GPU-accelerated EC2 instances.

Tools & Technologies

What Leaders Say

Small teams of 2-3 working with AI agents delivering work of dozens, orders of magnitude faster.

Will Brennan, VP Enterprise Technology
Get the full context.

Sign up to read complete case studies, access detailed metrics, and unlock all use cases.

Full Story

Blue Origin, the aerospace company founded by Jeff Bezos, needed to accelerate lunar hardware development while overcoming siloed specialist knowledge and traditional development cycles that took years. Standard LLMs lacked training in industrial manufacturing, thermal battery design, and regolith extraction.

The company built BlueGPT, a secure LLM gateway and multi-agent orchestration platform powered by Amazon Bedrock and the Strands Agents SDK. Running on Amazon EKS with OpenSearch for RAG knowledge bases and EC2 P5/G5 instances for GPU-accelerated physics simulations, the platform enabled small teams of 2-3 people working with AI agents to deliver work previously requiring dozens of engineers.

The results were remarkable: 2,700+ agents deployed across the company with 70% adoption, 3.5 million interactions monthly, and a 90% reduction in hardware development time. The platform produced TEAREx (Thermal Energy Advanced Regolith Extraction), the world's first AI agent-designed lunar hardware, going from concept to 3D-printed part in days rather than years.

Similar Cases

LM
Lockheed Martin
50%
reduction in data and ai tools

Lockheed Martin consolidated 46 disparate data management systems into a single integrated platform with IBM, reducing data and AI tools by 50% and enabling 10,000 engineers to build and deploy AI solutions using IBM Granite models.

Aerospace & Defense
MD
Missile Defense Agency (MDA)
1000x
increase in available threat data per scenario

The U.S. Missile Defense Agency partnered with C3 AI to deploy a generative AI platform for missile threat modeling and simulation. The solution delivers a 1000x increase in available threat data and reduces data generation time from weeks to minutes. This capability enables MDA to stress-test missile defense systems at unprecedented scale in secure, classified environments.

Aerospace & DefenseCAC3 Agentic AI PlatformCAC3 AI Parametric Threat Generative Modeling
P
Pinterest
17% YoY
revenue growth

Pinterest built an AI-powered discovery engine on AWS processing 18TB daily, delivering 10 million AI recommendations per second across 10,000+ GPU instances, driving 17% revenue growth and 70% AI-driven discovery.

TechnologyAEAmazon EKSASAmazon SageMaker
A
Albemarle
80%
support tickets resolved without it back-and-forth

Albemarle, the world’s largest lithium manufacturer with 8,300 employees across four continents, deployed Moveworks as its AI assistant ALbot within Microsoft Teams to provide multilingual, 24/7 employee support. The platform resolves 80% of support tickets without IT back-and-forth and has cut average resolution time by 49%, giving Albemarle’s growing global workforce consistent, instant support regardless of language or time zone.

ManufacturingMMoveworks
MI
Mondelez International
2–3x
engineering productivity gain

Mondelez International, one of the world’s largest snack companies with products in 150+ countries, selected AWS as its strategic cloud provider in late 2024 and deployed Amazon Q Developer as its agentic AI coding assistant for global engineering teams. The result: engineers experience 2–3x productivity gains, new environment setup dropped from days to under an hour, and server provisioning shrank from 7–10 days to 20 minutes. New engineers who previously took months to become effective team contributors now reach full productivity in days.

ManufacturingFood & BeverageAQAmazon Q Developer
RG
Renault Group
€1 million
value recovered in first 3 months

Renault Group, a global automaker with 98,000 employees across 36 countries, deployed the Celonis Process Intelligence Platform to eliminate inefficiencies in its Procure-to-Pay operations. By combining rapid wins in Accounts Payable with a DMAIC-driven transformation strategy, the company recovered €1 million within three months and €15 million in the first year.

ManufacturingAutomotiveCPCelonis Process CopilotCCelonis
WV
West Valley Construction
97%
safety events reduction

West Valley Construction, a California underground utilities contractor operating 600 vehicles across 12 locations, deployed Samsara’s Connected Operations Platform including AI Dash Cams and Fleet Telematics to replace disconnected manual systems. Within four months, safety events dropped 97% and the company went from award to full implementation in under 100 days.

ManufacturingSCSamsara Connected Operations PlatformSFSamsara Fleet Telematics
P
Petrobras
$120M
tax savings identified

Petrobras applied generative AI and intelligent automation to its tax department, feeding 150 pages of Brazilian tax regulations and three months of financial data into a model built on Automation Anywhere, Amazon Bedrock, and Amazon SageMaker. In three weeks, the system identified $120 million in tax savings and filed taxes within three days—the first time in 15 years the team avoided weekend work during tax season. The company projects over $1 billion in total savings by year-end.

EnergyASAmazon SageMakerABAmazon Bedrock