XPeng bets on end-to-end VLA 2.0 to power its GX Robotaxi commercial rollout
bookwormengr · x · 2026-09-08
XPeng's self-driving stack is built on its proprietary end-to-end VLA 2.0 (Vision-Language-Action) model. Unlike rigid hand-coded driver-assist rules, it perceives, reasons, and drives like a human, working across cities, highways, and tight parking structures without expensive pre-mapped HD infrastructure. The same software foundation is driving the commercial rollout of its custom GX Robotaxi fleet into autonomous ride-hailing.
More from Embodied
- Robotics researcher to AI hypesters: debug a controller before making grand claims — GeorgiaChal · 2026-09-08
- 'Armored Core' robot demo video makes the rounds on Reddit — Human_Journalist_758 · 2026-09-08
- Kid-focused AI computer Daso rebrands as Kale, built for creation not scrolling — NERDDISCO · 2026-09-08
- New article argues robots will create more retail jobs by 2035, not fewer — broodsugar · 2026-09-08
- RL-X adds BeyondMimic-style G1 motion tracking in MuJoCo and MJX Warp — Jan_R_Peters · 2026-09-08
- XPeng's Turing chip strategy: 3,000 TOPS Robotaxi, 2,250 TOPS humanoid robot — bookwormengr · 2026-09-08