FigureAI Makes Previous-Generation Robots “Throw Themselves Into Lava” for Decommissioning
FigureAI Makes Previous-Generation Robots “Throw Themselves Into Lava” for Decommissioning
FigureAI decommissioned Figure 02 robots by having them jump into molten steel after Arnold Schwarzenegger suggested melting them, releasing a video that recreates a Terminator scene.
In late August, the CEO of the AI humanoid robotics startup announced that Figure 02 robots would be decommissioned because headquarters was running out of space as development of the next-generation humanoid began. He asked X followers how best to do it.
Arnold Schwarzenegger appeared in the comments and replied: “You should melt them down.”
Figure took the advice literally and even invited Schwarzenegger to participate. Yesterday the company released an epic video showing robots jumping into a ladle of molten steel and recreating the famous Terminator thumbs-up scene.
The robots were trained to autonomously drop themselves down. A new motion model based on stunt performers was trained in simulation, and the robots were then made to jump from a second floor onto air cushions.
The entire operation had to be carried out in Finland because no foundry in the US or Mexico agreed to melt robots containing lithium-ion batteries.
Figure will make a limited series of commemorative souvenirs from the recovered metal.
Why it matters
AnalysisThe case shows how AI robotics companies can turn hardware retirement into viral storytelling. It also highlights the safety and recycling challenges posed by robots with lithium-ion batteries.
Discuss in community
Share your questions and insights with developers
Related in Ecosystem
Practical guides, core terminology, and agent skills
How to Turn Articles into Branded Pinterest Pins with Claude Code
1. What the Tool Does and Which Problem It Solves...
Prompting in 2026: ChatGPT, Claude, image generation and work assignments
Most people use AI as a search engine: type a short query, skim the answer and close. For reference it works. For everything that is really important - no.
RAG (Retrieval-Augmented Generation)
An architectural pattern for corporate AI that dynamically enriches the model's context window with relevant verified knowledge from external repositories (vector databases, graphs, full-text indexes) before generating the final response.
Text Summarization (How to Condense 50 Pages into 5 Key Points)
A methodology for effectively condensing large volumes of information (Summarization). It explains how to transform 50-page reports into concise executive summaries using clear prompt templates following the structure: 'Core Insight + Risks + Next Steps'.