Selling Autonomy Not Robot Cells
Path Robotics
This shifts competition from selling whole welding cells to selling an autonomy layer that can sit on top of robots a plant already owns. That matters because many large fabricators already have ABB or Yaskawa arms on the floor, so Novarc can pitch a cheaper, faster upgrade path, while Path still has to prove that its tighter hardware and software stack delivers enough extra yield and labor savings to justify replacing or bypassing existing equipment.
-
Path already uses third party robot arms inside its own cells, including Yaskawa sourced hardware, then adds its own sensors, compute, weld planning, and remote support around that stack. Novarc is attacking that same installed base from the other direction, by bringing adaptive welding software to ABB and Yaskawa cells already in service.
-
The practical buyer difference is procurement. Path sells a full production system and often wraps it in a subscription, which means Path owns more of uptime and performance. Novarc and Cohesive can work through robot OEMs and integrators, which fits how many factories already buy automation, one robot cell upgrade at a time.
-
Cohesive shows where this market can go next. Its Argus OS is built to run not just welding, but also sanding and polishing on common robot arms. That makes the software layer more like a plant wide automation standard, instead of a single purpose machine, which is a different value proposition from Path’s welding first stack.
The next phase of this market is likely to center on who becomes the default intelligence layer for messy metalwork, not who ships the most robot hardware. If retrofit platforms keep improving, Path will need to extend Obsidian across more robot formats and show that its deeper integration still wins on first pass yield, setup time, and support in the hardest welding jobs.