We were 12 months off.
While most of us were scrolling humanoid backflip videos, the real story was happening somewhere less photogenic — procurement meetings and construction sites.
Four signals from the last 30 days:
→ Schaeffler signed a binding deal for up to 2,000 humanoid robots across its global plants (supplied by UK startup Humanoid).
→ Japan Airlines started trialling
@UnitreeRobotics humanoids at Haneda — baggage handling and cabin cleaning.
→
@Apptronik closed a $520M Series A extension. Total raised now past $935M.
→ Tesla broke ground on a dedicated Optimus factory at Giga Texas. 5.2M sq ft. Target: 27,000 units a day. First line online by August.
Six months ago the industry called humanoid procurement a "2027 conversation."
The buyers didn't wait.
The robots work. The open question now isn't if, it's which foundation model ends up running them. Helix, SkildAI , π0, GR00T, or one of the vertically integrated stacks (Tesla, Boston Dynamics, Apptronik, 1X, AgiBot).
That's why we built
@MACHINASUMMIT .
July 7,
@joinstationf . One room:
@BerntBornich (
@1x ), Carolina Parada (
@GoogleDeepMind Robotics),
@jeffcardn (
@Apptronik ), Marc Raibert (
@BostonDynamics / RAI Institute), Jonathan Hurst (
@agilityrobotics ), Abhinav Gupta (
@SkildAI ) - and 20 more.
The people whose deployment decisions will define what factory floors look like for the next decade.
Then July 8-9, Carrousel du Louvre, for
@RaiseSummit .
The procurement window just opened.
Genuine question for the robotics people here: which model ends up running the floor - Helix, π0, GR00T, Skild, or a vertically integrated stack? Reply with your bet.