Eka Robotics, a US-based startup with a specialization in robotic dexterous manipulation, has recently made headlines by announcing the successful closure of its Series A funding round, raising an impressive $150 million. This influx of capital has propelled the company's post-money valuation to a staggering $1.45 billion. Co-founded by Pulkit Agrawal, a professor from MIT, and Tuomas Haarnoja, a former researcher at Google DeepMind, Eka Robotics distinguishes itself from mainstream Vision-Language-Action (VLA) models with its innovative technical approach, namely the Vision-Force-Action (VFA) model.

The VFA model represents a significant leap forward in robotic technology by seamlessly integrating visual recognition with force feedback. By leveraging large-scale physical simulation reinforcement learning, the model enables robots to engage in trial-and-error learning within virtual environments. This process allows them to acquire and refine skills that can then be seamlessly transferred to real-world applications. Measured data attests to the model's efficacy, revealing that Eka Robotics' robots are capable of grasping up to 1,500 items per hour—approximately 2.5 times faster than the average human benchmark. Moreover, these robots boast a grasping success rate exceeding 95% and are adept at handling items weighing up to 1 kilogram.

In terms of hardware, the cost of Eka Robotics' gripper remains competitively priced below $1,000, while its designed lifespan extends to approximately 5 million operational cycles. Currently, the company is actively collaborating with a select group of industrial clients on pilot projects. These projects encompass a range of tasks, including grasping, placing, and cable manipulation. Through these collaborations, Eka Robotics aims to validate the adaptability and long-term stable operation capabilities of its technical approach in complex industrial scenarios.