Eka Robotics Hits $1.45B Valuation With Force-Based Manipulation Model Showing Superhuman Dexterity
Eka Robotics, founded by MIT professor Pulkit Agrawal and ex-DeepMind researcher Tuomas Haarnoja, raised a $150M Series A led by a16z and F-Prime at a $1.45B valuation. Its vision-force-action model, trained with large-scale simulated reinforcement learning, lets a force-sensitive gripper pick about 1,500 objects per hour with over 95% reliability and adapt to new objects without retraining. Experts praised the demos, but the system still has to prove itself in real industrial deployments and on tasks beyond pick-and-place.
Skynet Chance (+0.01%): Stronger, faster physical manipulation expands the ways AI systems could act in the real world, slightly raising the stakes of misaligned embodied AI. However, the system is task-specific industrial automation with safety-conscious force control, so the added risk is small.
Skynet Date (+0 days): Faster progress in robotic dexterity modestly speeds up the arrival of capable embodied systems that could pair with general planners like LLMs. The effect on the overall risk timeline is minor.
AGI Progress (+0.02%): The model narrows the generality-performance gap in manipulation, a long-standing bottleneck for physical intelligence. Its simulation-driven RL approach exceeds what human demonstrations alone can teach and could serve as a low-level physical layer for general AI planners.
AGI Date (+0 days): Simulation can generate thousands of hours of robot experience in minutes, easing the data bottleneck in robotics. Major venture backing further speeds the arrival of embodied capabilities relevant to AGI.
<< All AI news for October 20, 2026
[ Get the daily index digest on Telegram → ]Related AI News
- Elorian Launches Glimpse, a Visual Reasoning Model That Revisits Images Mid-Reasoning 2026-10-16
- Google DeepMind Launches Gemini 4 Argon Frontier Model with Leading Agentic, Scientific, and Embodied Reasoning Capabilities 2026-10-13
- Nvidia Expands Halos Safety Architecture to Accelerate Physical AI and Autonomous Robotics 2026-10-08
- Efficient AI Masters Long-Horizon Imperfect-Information Strategy in Stratego 2026-10-01
- AMD Buys Fei-Fei Li's World Labs for $8.2 Billion to Push Into World Models and Physical AI 2026-09-29