OLogic

10/19/2026

Spotlight on our Thermal Design service!

Our team delivers tailored thermal analysis and cooling strategies built around your product’s performance and reliability targets. From conduction, convection, and airflow modeling to CFD-driven evaluation with board- and system-level layout guidance, we engineer thermal solutions for efficiency, manufacturability, and long-term stability.

See all services at OLogic: https://na2.hubs.ly/H08t8X10

10/16/2026

RoboBusiness 2026 is next week, and OLogic CEO and Co-Founder Ted Larson will be on stage in Santa Clara!

The biggest constraint in robotics today isn’t hardware. It’s data. In his session, “Egocentric Robot Learning: Teaching Robots Through Human Experience,” Ted will break down how first-person human demonstrations are giving robots the real-world skill data they need, and what that shift means for teams building commercial robots.

Join us Wednesday, October 21 at 10:45 AM in Room 203.
https://na2.hubs.ly/H08qSCm0

10/15/2026

Our CEO Ted Larson has a new blog post on a practical design question for learned robot systems: when a robot makes a decision, what should it remember?

When a warehouse robot stops short of a tote, the video only shows what happened physically, not whether the cause was perception, grasp planning, a force sensor reading, or model uncertainty. Ted makes the case for a synchronized "decision record" that captures detections, confidence, rejected paths, safety constraints, and model and software versions. Since robots can't record everything, he outlines approaches like a rolling buffer that preserves the moments around a fault, e-stop, or human takeover. He also stresses that this has to be designed into the sensors, compute, storage, and software architecture from the start rather than added later.

Read the full post: https://na2.hubs.ly/H08gsQj0

How robot decision-making really works and why capturing the right evidence is key to diagnosing failures and improving fleets.

10/14/2026

New Stanford research finds that care robots in Japanese nursing homes helped ease staffing shortages without displacing workers, according to a study published in the inaugural issue of Health Affairs Review.

The researchers, working with economists from the University of Tokyo and Notre Dame, analyzed survey data from roughly 860 nursing homes and used regional differences in government robot subsidies as a natural experiment. Facilities that adopted robots, including transfer aids, mobility robots, and monitoring and communication systems, saw fewer retention problems, along with a 28% increase in care workers and a 39% increase in nurses. That employment growth came entirely among flexible, nonregular staff, and the authors say it suggests robots can complement caregivers, particularly by supporting physically demanding tasks, rather than replace them.

Read more: https://na2.hubs.ly/H08gsQd0

10/13/2026

Runway has introduced Praxis-1, an open-weight world action model that turns the company's large-scale video pretraining into control for robots.

Because robot demonstration data is scarce and expensive to collect, Praxis-1 learns mostly from third-person video, so its performance scales with how much video it can learn from rather than how many robot demonstrations exist. Runway reports that simulating robot policies inside its world model predicts real-world results with a 0.95 correlation. Early partners including Noble Machines, Standard Bots, and Ultra are running it on their own hardware, and Runway says a single model has adapted across very different embodiments, from bimanual arms to humanoids, with light fine-tuning. A public release with open weights is planned in the coming months.

Read more: https://na2.hubs.ly/H08gsQ70

When Runway does release Praxis-1 publicly, it plans to ship it with open weights rather than as a closed model.

10/12/2026

Finding a signal integrity problem after fabrication is an expensive way to learn. 📡
OLogic's mixed signal simulation catches SI, PI and EMI issues before your board ever goes to fab. Using Cadence Sigrity and Clarity 3D tools, our team brings Altium designs into simulation to verify compliance for DDR, MIPI, HDMI, USB 3.1, PCIe and more, plus full PDN, IR drop and power ripple analysis.

The result: fewer hardware revisions, faster timelines, and more confidence heading into manufacturing.
Learn more: https://na2.hubs.ly/H08gsVD0

10/12/2026

Carbon Robotics, a developer of physical AI for agriculture, has launched its Carbon Autonomy Ready program, with Great Plains Manufacturing joining as the first partner.

The program sets a technical specification and process for integrating any manufacturer's implements directly into the Carbon Autonomy platform. Tractors running Carbon Autonomy will connect to real-time data from Great Plains equipment and make autonomous adjustments on the fly, including implement depth, down pressure, and leveling, the same way an operator would from the cab. The goal is to let farmers choose the equipment that works best for their operation without giving up autonomy. The integration will be showcased at the 2026 Farm Progress Show on a Great Plains HT1100 Terra-Max equipped with integrated sensors and a Carbon implement kit.

Read more: https://na2.hubs.ly/H08gsNS0

Carbon Robotics, a developer of physical AI for agriculture, has announced the launch of its Carbon Autonomy Ready program. Great Plains Manufacturing, a manufacturer of seeding and soil management…

10/11/2026

Inbolt, a Paris-based company that gives industrial robots real-time 3D vision, has raised $12.5 million in a round led by Shift4Good, bringing its total funding to $34 million.

The system pairs 3D vision with a hardware-agnostic AI software layer that lets robots perceive misalignment and part variation and adjust their control loops on the fly. It runs on major platforms including Fanuc, ABB, Kuka, Yaskawa, Comau, and Universal Robots, and can be added to existing production lines without replacing installed equipment. Inbolt is already deployed on 200+ robots across 100+ factories, with customers including Bosch, Ford, Stellantis, and Toyota. The new funding will support expansion in the U.S. and APAC and entry into data center and electronics manufacturing.

Read more: https://na2.hubs.ly/H08gsN30

Inbolt, which gives industrial robots the ability to see, think, and adapt in real time, has announced $12.5 million in funding. The round was led by new investor Shift4Good, a venture capital firm…

10/10/2026

In a blog post on our site, our CEO Ted Larson argues that the next wave of robotics may be won by whoever builds the strongest data infrastructure for AI model training rather than the most impressive humanoid.

Unlike language models trained on internet-scale text, robots have to learn through physical interaction, which makes high-quality first-person (egocentric) data one of the industry's biggest bottlenecks. Ted points to research like Georgia Tech's EgoMimic and Meta's Project Aria showing that human demonstrations captured from a first-person view can meaningfully improve robot learning, and to a new category of "robotics data factories" built to capture and operationalize that experience at scale. The post also covers the learning flywheel that deployments can create and why governance, privacy, and IP protection will matter as much as the technology itself.

Read the full post: https://na2.hubs.ly/H08gsNF0

Discover why embodied data and AI model training are reshaping robotics' competitive advantage and future innovation.