All in on AI at Astra
Summary
Astra, a space company that went private in 2024 at less than 1% of its peak multi-billion dollar SPAC valuation, is aggressively implementing an AI-centric operational model. Following multiple Rocket 3 failures and financial struggles, co-founder Chris Kemp reported nearly \$50 million in revenue in 2025 from electric thrusters and expects the first Rocket 4 launch by year-end. The company, based in Alameda, California, is eliminating all purchased enterprise software, instead developing its own AI model trained on all internal data, from budgets to technical specifications. This AI is intended to assist engineers and ensure consistency in work instructions. Astra's strategy involves a drastic reduction in its workforce, from a peak of 400 to 110 employees, with a goal to replace most human roles with AI and automated manufacturing to cut costs and enable mass production. Kemp predicts engineering team leaders will become the last members of their teams, with robots and AI agents performing the work.
Key takeaway
For Directors of AI/ML evaluating operational efficiency, Astra's aggressive AI adoption presents a stark model. You should assess whether your organization's culture and technical infrastructure can support a "no purchased software" policy and a mandate for all employees to contribute to internal AI. This approach, while potentially yielding extreme cost reductions and output increases, also carries significant risks regarding talent retention and the adaptability of AI in highly specialized, risk-averse fields like space.
Key insights
Astra is pursuing an extreme AI-driven operational model to eliminate software costs and drastically reduce its workforce.
Principles
- AI can replace, not just augment, human roles in complex engineering.
- Mandate employee contribution to internal AI systems or face termination.
- Aggressive cost reduction through automation enables mass production.
Method
Train a proprietary AI model on all company data (budgets, specs) to automate operations and rewrite work instructions, eliminating purchased software.
In practice
- Develop internal AI tools to replace expensive enterprise software.
- Integrate AI into engineering workflows for consistency and efficiency.
- Implement automated manufacturing with AI agents for cost reduction.
Topics
- Artificial Intelligence
- Workforce Automation
- Space Industry
- Rocket Propulsion
- Cost Reduction
- Enterprise Software
Best for: CTO, VP of Engineering/Data, Executive, Director of AI/ML, Entrepreneur, Consultant
Related on AIssential
Editorial summary, takeaway, and curation by AIssential. Original article published by SpaceNews.