Running an Automotive Open Source Program at Scale
Talkat Open Source Summit Japan + AI_dev: Open Source GenAI & ML Summit + Automotive Linux Summit 2025 · 2025-12-08..12-10·
summary
Describes operating an SDV open source program that tracks 1000+ components, enforces hermetic, reproducible builds and supply-chain provenance, and uses a GitHub “meta-repo” plus copybara for mirroring and CI orchestration. Gives a concrete cost example: building LLVM takes about 325 minutes on paid cloud runners (~$0.25/min), so cloud costs scale rapidly and motivate evaluating self-hosted runners versus total cost of ownership. Covers policy, license checks, security scans, and developer enablement workflows.
AI-written summary · about
chapters (16)
- 0:00Automotive open source intro
- 1:02Program goals
- 2:27Legal and licensing
- 3:32Security and supply chain
- 4:16Safety-critical operations
- 5:17Internal policies
- 6:40Strategic benefits
- 7:02In-house software issues
- 8:29Training and policies
- 9:10Tools and automation
- 10:15GitHub automation
- 11:40Mirroring costs
- 12:23Cloud build costs
- 13:27Self-hosted runners
- 14:52Summary
- 15:12Q&A
related talks
13mJust how far can you go with FOSS?
19mDriving Open Source
39mPanel Discussion: Open Source as a Path for a Competitive Automotive Industry
19mDriving Automotive with Open Source: From Compliance to Collaboration
22mSteering into Tomorrow: Embracing Software-Defined Vehicle in an Open-Source Appro...
32mOpen Source: The Engine of Innovation in the Digital Age
30mSoftware-Defined Vehicles and FOSS
35mDriving the automotive industry into the future with Open Source