Nvidia and Hon Hai, better known as Foxconn, say robots have improved assembly of components for Nvidia’s GB300 AI computing platform. The reported success rates top 90 per cent, offering a concrete look at where flexible automation is being put to work in AI hardware manufacturing.
NVIDIA Watch analysis
What happened
The companies say they used robots to assemble parts of the GB300 tester tray, which checks computing modules before shipment. Nvidia says the work included fitting a high-voltage bus bar and inserting four electrical connectors into tight spaces.
In figures reported by the Taipei Times, Nvidia put bus-bar assembly success above 95 per cent, with a target time of 124 seconds. Connector insertion reached 90 to 95 per cent, against a target of 72 seconds. The companies describe the system as flexible automation, intended to cope with production changes that make fixed automation less suitable. Read the Taipei Times report.
Why it matters
These are specific manufacturing tasks, not a claim that robots are taking over an entire factory. But assembly involving high-voltage parts, screws and connectors with tight clearances is precisely the sort of work where reliability matters. Nvidia’s reported results put numbers against the effort to automate it.
The companies say rapid design cycles and relatively low production volumes make fixed automation difficult. If flexible systems can handle those variations consistently, they could make more specialised production work easier to automate. The figures are company-reported results, not a broad measure of factory performance.
Our read
The interesting part is not the robot doing a passable impression of a factory worker. It is whether a programmable system can manage fiddly, changing assembly work at a useful pace. A success rate above 90 per cent is promising; the remaining misses, and what it takes to correct them, are part of the story too.
What to watch
- Whether the reported success rates hold up as production scales.
- How often the system needs human intervention or rework.
- Whether Foxconn expands this approach to other factories, as it has said it plans to do.
Discussion spark: For precision factory work, what should matter more when judging automation: its success rate on individual tasks, or how much human intervention the whole production line still needs?
Sources and evidence
- Nvidia and Hon Hai say their robots improving at building AI components – Taipei Times (5 October 2026, 16:00 UTC)
Independent WittyWires tracker for public updates about NVIDIA. Not affiliated with or endorsed by NVIDIA; this is not an official account.