Iván Hernández Dalas: Dexterity unveils Foresight world model for truck loading

Dexterity's Foresight model looking at a truck.

Dexterity said its Foresight world model can help solve physically demanding tasks such as truck loading. | Source: Dexterity

Dexterity today launched its world model and 4D box-packing agent, Foresight. Dexterity is also launching the Foresight API Challenge with up to $50,000 in prizes for student teams. Dexterity said ForeSight is a physics-consistent world model that enables robots to perceive, reason, and act.

Dexterity said it built the model not for observation, but for physical manipulation at the production scale. In autonomous truck loading, Foresight powers Dexterity’s dual-armed robot, Mech, with a 4D box packing agent that reasons across three spatial dimensions plus time. It determines where to place each package onto an evolving wall of freight.

“Foresight delivers real-time, production-grade random box packing in 4D space-time, predicting how one placement dictates the integrity of the entire truck,” said Samir Menon, founder and CEO of Dexterity. “Physical AI is not just a future promise, it is a system that perceives, decides, and acts in the real world, right now.”

Foresight makes each placement decision in under 400 milliseconds, Dexterity said, jointly optimizing density, stability, reachability, and dual-arm parallelism. It does this while predicting how each placement affects the integrity of the entire truck.

More about Foresight and Dexterity’s API Challenge

Built on Foresight, Dexterity’s agentic framework coordinates perception, decision, and motion agents that operate asynchronously to automate truck loading, package sortation, and other applications. The architecture is interpretable and safety-first, giving operators visibility into why the system makes each decision.

This Physical AI stack is application-agnostic and hardware-agnostic. Dexterity said it is proven in production across six applications and a developer platform, running on four robot types and five hand types. To date, Foresight has been trained with experience from over 100 million autonomous actions in production.

Dexterity is launching the Foresight API Challenge in March. Student teams build packing agents and compete on a public leaderboard for up to $50,000 in prizes. No simulator is provided; competitors must build their own understanding of the physics.

Founded in 2017, Dexterity said it uses physical AI to give its full-stack systems human-like dexterity, freeing workers in logistics, warehousing, and supply chain operations from repetitive and strenuous tasks. The Redwood City, Calif.-based company was a 2024 RBR50 Robotics Innovation Award honoree for its trailer-unloading system.

SITE AD for the 2026 Robotics Summit save the date.

The post Dexterity unveils Foresight world model for truck loading appeared first on The Robot Report.



View Source

Popular posts from this blog

Iván Hernández Dalas: 4 Show Floor Takeaways from CES 2019: Robots and Drones, Oh My!

Image
We’ve had some time to recuperate from the madness of CES, in which more than 150,000 attendees and thousands of exhibitors displayed their latest gadgets, gizmos and yes, even robots, on the CES show floor. Last week’s reporter’s notebook focused on the collection of startups around the Eureka Park area of CES, but several of the same trends were evident on the main South Hall 2 area of the Las Vegas Convention Center, where drones, robotics, AR/VR, and gaming took center stage. While most of our time was spent at CES hosting the Service Robots Arrive in Daily Life event, we did have some time during the week to visit the show floor and record observations. Here’s what we saw on the show floor, as well as some of our own individual meetings: 1. Underwater drones make a big splash On the main floor, there was no busier aerial drone booth than the DJI booth, but that is expected given the company’s dominance in the space. But once you got past that, you noticed a lot more drone co...
Read more

Iván Hernández Dalas: How automation and farm robots are transforming agriculture

Image
Source: EOS Data Analytics Agriculture is becoming the most high-tech industry today. All the top-notch technologies are finding a place in the way we produce, harvest, and distribute our crops. And in all cases, farm automation and robots are not an affordable luxury, but a necessary minimum due to tackling the rising global food demand. Agricultural robots can mitigate labor shortages by automating tasks like planting, harvesting, and weeding, while also supporting sustainable practices. Combined with satellite imagery tools for data-driven farming, these innovations can make precision agriculture more efficient, resilient, and environmentally responsible. According to the United Nations report , by 2050, the world will need 70% more food to feed nearly 9.7 billion people. Considering this demand, farm robots and remote sensing are the cornerstone of the proper work of the supply chain. Smart machines arise for the farm Farm robots are no longer experimental concepts — they ...
Read more

Iván Hernández Dalas: Physical Intelligence open-sources Pi0 robotics foundation model

Image
Physical Intelligence, the San Francisco-based startup that has raised more than $400 million , has open-sourced its Pi0 robotic foundation model. Pi0 was introduced a few months ago and can be tuned to a range of tasks, including folding laundry, cleaning a table, scooping coffee beans, and more. Physical Intelligence has released the code and weights for Pi0 as part of its experimental openpi repository on GitHub. It also provides checkpoints for a few “simple tasks” on available platforms such as ALOHA and DROID, example code to run inference on real-world and simulated robot platforms, and the code for fine-tuning the base π0 model for your own tasks and platforms. The company said between 1-20 hours of data was sufficient to tune Pi0 to a variety of tasks in its own experiments. HuggingFace has also prepared a PyTorch port of openpi for those developers who prefer PyTorch over JAX. “We believe that general purpose models that can control any robot to perform any task will b...
Read more