Robotera Ships Thousands, Humanoid Soldiers Head to Ukraine, and Scaling Laws Come for Dexterous Hands
Funding
ROBOTERA Raises $200M and Begins Thousand-Unit Deliveries in Q2 2026
This is the commercialization signal the space has been waiting for. ROBOTERA isn't just raising money; it's actually shipping robots at scale, with 95% of core components built in-house and deployments live across 10+ logistics centers with China Post and SF Group. The production-at-scale question just got a real data point.
Mind Robotics Hits $3.4B Valuation as Kleiner Perkins Leads $400M Round
The Rivian spinoff crossed $1B in total funding with this round and is now valued at $3.4B in under a year. The Kleiner Perkins lead is a signal that tier-1 West Coast VCs are back underwriting physical AI bets at serious scale, with Rivian as the anchor customer giving the deployment story more credibility than most humanoid plays at this stage.
News
The first claimed deployment of a humanoid robot to an active conflict zone is a genuine threshold moment for defense robotics, even if the technology is still early. The skeptical read: a CEO who previously ran a bankrupt fintech is asking for $500M at $3B+ valuation while claiming a 250x production scale-up on roughly $21M in funding. Worth tracking hard, but verify before buying the narrative.
Amazon's Vulcan Named RBR50 Robot of the Year Ahead of Robotics Summit Boston
Amazon's Vulcan, the first warehouse robot with a genuine sense of touch, just won the industry's most-watched annual innovation award. It's operational across Amazon fulfillment centers, stowing at roughly human speed, and has a 2026 global rollout planned. Getting to parity with average human speed after 500,000+ real stows is a practical benchmark most robotics teams should be studying.
China's ROBOTERA Reports 300%+ Growth as It Expands Beyond Logistics Into Auto and Electronics
Beyond the headline raise, ROBOTERA's claim that hardware from its stack has been adopted by NVIDIA, Apple, and Boston Dynamics for their own robotics R&D is worth flagging. That's a supply chain play on top of a deployment play, which is a smarter moat than most Chinese humanoid startups are building right now.
Research
EgoScale: NVIDIA Shows Dexterous Manipulation Follows LLM-Style Scaling Laws
This is one of the most important research results of 2026. NVIDIA trained a VLA on 20,854 hours of egocentric human video and found a near-perfect log-linear scaling law between data volume and robot performance, meaning more human video data predictably produces better dexterous robots. If this holds, whoever accumulates the most egocentric human data has a structural, compounding advantage in physical AI.
Analysis
KraneShares: The Humanoid Robotics Industry Is in a 'GPT-2.5 Moment'
A solid mid-year state-of-the-industry read covering JAL's Haneda airport deployment, BMW's 30,000-vehicle production milestone with Figure AI, and why McKinsey calls the robotics supply chain the most underappreciated constraint on scale. The framing from Bessemer Venture Partners that we're in a "GPT-2.5 moment" is useful shorthand: capabilities are real, but the 99.9% reliability gap for production deployment is still very wide.
Stay Ahead
Delivered each morning.