Intervention decline drives Weave margins

This is the core test of whether Weave becomes a software improving robot business or a labor backed service business. Every time Isaac 0 gets stuck, a remote specialist spends five to ten seconds fixing the fold, so gross margin improves only if those handoffs drop fast enough to outrun growth in deployed robots. The reason Weave started in commercial laundry and a dense Bay Area rollout is to collect many repeats, find edge cases quickly, and turn each correction into weekly model gains.

• Weave already has evidence that deployment data can bend the curve. In live use, adding proprietary laundry data to pre training cut missed grasps by 42% and reduced human interventions by 50%, which directly attacks the variable labor cost inside every fold.
  1 sacra
• This is a different margin problem than most home robots. Sunday avoids teleoperation in homes and spends on glove collected training data before launch, while Weave puts robots into homes earlier and pays for remote assistance now in exchange for faster real world learning on the exact task it sells.
  1 sacra 3 sacra
• Consumer robotics history shows why this matters. Roomba scaled because vacuuming was cheap, reliable, and needed little human backup. Laundry folding is a harder manipulation task, so Weave must drive intervention rates down until the robot behaves more like an appliance than an always supervised service.
  1 sacra 2 sacra

The next phase is a race between the data flywheel and the service burden. If intervention rates keep falling with each software release, Weave can widen margin, expand beyond dense local rollout, and use laundry as the first wedge into a broader home manipulation platform. As autonomy improves, the economics shift from staffing operations to selling software enhanced household labor.