Kodiak AI has announced a strategic agreement with Bosch to scale the manufacturing of production-grade autonomous trucking hardware, marking a significant step toward large-scale deployment of driverless trucks. The collaboration was revealed ahead of CES 2026, where a Kodiak Driver-powered autonomous truck will be displayed at Bosch’s booth in Las Vegas.

The partnership focuses on building a redundant, automotive-grade platform that integrates hardware, firmware, and software interfaces required to deploy Kodiak’s AI-powered virtual driver at scale. By combining Kodiak’s autonomy software with Bosch’s manufacturing expertise and supply chain capabilities, the companies aim to move autonomous trucking beyond pilots and into sustained commercial operations.

Scaling Physical AI for Heavy-Duty Trucks

Kodiak’s autonomous system, known as the Kodiak Driver, is designed as a unified platform that blends AI-driven perception and planning software with modular, vehicle-agnostic hardware. The system can be integrated either directly on a truck production line or through aftermarket upfitters, giving fleet operators flexibility in how autonomous capability is deployed.

Under the agreement, Bosch will support the development of a redundant autonomous hardware platform, supplying key components such as sensors, steering systems, and other vehicle actuation technologies. These components are designed to meet automotive-grade reliability standards, a critical requirement for long-haul trucking applications where uptime and safety are paramount.

“Advancing the deployment of driverless trucks and physical AI requires not only robust autonomy software, but also manufacturing experience and a resilient supply chain,” said Don Burnette, founder and chief executive of Kodiak. He emphasized that Bosch’s industrial scale and system-level integration expertise are essential for commercial success.

From Commercial Pilots to Industrial Scale

Kodiak has already deployed trucks operating without human drivers in commercial service, positioning the company as one of the few autonomous trucking developers with real-world revenue-generating operations. The new agreement is intended to build on that foundation by enabling higher-volume production and standardized hardware configurations.

Bosch’s role extends beyond component supply. As the world’s largest automotive supplier, the company brings decades of experience in industrialization, quality assurance, and global manufacturing. This expertise is expected to help Kodiak transition from limited deployments to repeatable, scalable production suitable for fleet-wide adoption.

Paul Thomas, president of Bosch in North America and president of Bosch Mobility Americas, said the collaboration allows Bosch to deepen its understanding of real-world autonomous vehicle requirements while contributing production-grade systems to the broader autonomous mobility ecosystem.

CES 2026 and the Push Toward Autonomous Freight

Autonomous trucking emerged as a key theme at CES 2026, with increasing emphasis on commercialization rather than experimental prototypes. Kodiak and Bosch used the event to highlight how Physical AI systems are moving into operational environments where reliability, redundancy, and cost efficiency matter as much as technical performance.

The Kodiak Driver-powered truck on display demonstrates how the integrated platform brings together sensors, compute, and vehicle control into a single autonomous system. Unlike many earlier demonstrations, the focus is on readiness for deployment rather than future concepts.

Industry analysts view the partnership as a sign that autonomous trucking is entering a more mature phase, where partnerships with established automotive suppliers are essential to overcoming manufacturing and regulatory hurdles.

Broader Implications for Autonomous Logistics

For Kodiak, the deal supports its long-term vision of becoming a trusted provider of autonomous ground transportation across commercial and public-sector applications. The company has also positioned its technology for use in government and national security contexts, where reliability and safety standards are especially stringent.

The collaboration underscores a broader trend in robotics and automation, where autonomy developers increasingly rely on established industrial partners to bridge the gap between software innovation and large-scale deployment. As Physical AI systems move from test routes to highways and supply chains, the ability to manufacture and support hardware at scale becomes a decisive competitive advantage.

With CES 2026 as the backdrop, the Kodiak-Bosch agreement signals growing confidence that autonomous trucking is transitioning from experimentation to infrastructure, setting the stage for wider adoption in the years ahead.

Automation, Business & Markets, News, Robots & Robotics

Humanoid robots were the headline attraction at the Silicon Valley Humanoids Summit, where startups and researchers showcased rapid progress in mobility, perception, and manipulation. Demonstrations highlighted robots walking autonomously, handling objects, and interacting with human-built environments. Despite the enthusiasm, discussions repeatedly returned to unresolved challenges around cost, reliability, and commercial readiness.

The summit reflected a broader surge of interest in humanoid robotics driven by advances in artificial intelligence, sensors, and actuators. Investors, engineers, and corporate buyers attended sessions focused on how humanoid form factors could operate in warehouses, factories, and service environments. Yet many participants cautioned that impressive demonstrations do not always translate into scalable products.

Progress Meets Practical Constraints

Several companies presented humanoid robots designed to work in logistics and manufacturing, emphasizing their ability to navigate spaces built for humans without infrastructure changes. Developers argued that bipedal robots could eventually replace or support workers in tasks ranging from material handling to inspection. The appeal lies in flexibility, with a single robot potentially performing many roles rather than one specialized task.

However, experts at the summit noted that humanoid robots remain expensive to build and maintain. Power consumption, mechanical wear, and software robustness continue to limit continuous operation. While some robots can perform short demonstrations reliably, sustaining performance across long shifts in unpredictable environments remains a significant hurdle.

There was also skepticism about whether humanoid robots offer clear advantages over existing automation. In many warehouses and factories, wheeled robots, conveyors, and fixed automation already deliver higher efficiency at lower cost. Critics argued that humanoid designs may only make economic sense in narrow use cases where human-like mobility is essential.

Market Expectations and Deployment Reality

The summit highlighted growing tension between investor expectations and deployment timelines. Several startups predicted rapid adoption within the next few years, pointing to pilot programs and early commercial agreements. Others urged caution, warning that widespread deployment would likely take longer due to safety certification, workforce integration, and total cost of ownership considerations.

Labor dynamics were a recurring theme. Proponents suggested humanoid robots could help address worker shortages and reduce injury risk in physically demanding roles. Skeptics countered that deploying complex robots introduces new maintenance and oversight requirements that may offset labor savings, at least in the near term.

Regulatory uncertainty also surfaced during discussions. Humanoid robots operating alongside humans raise questions about liability, workplace standards, and certification processes. Industry observers noted that clear regulatory frameworks will be critical before large fleets can be deployed in active industrial settings.

A Sector at a Crossroads

By the end of the summit, humanoid robots had clearly captured attention, but consensus remained elusive. The technology is advancing rapidly, and real-world pilots are expanding, yet doubts persist about near-term scalability and economic viability. Many attendees described the current moment as a transition from hype-driven excitement to a more sober evaluation of practical constraints.

The discussions underscored that humanoid robotics is no longer a speculative concept, but neither is it a solved problem. As companies continue to test deployments and refine designs, the coming years are likely to determine whether humanoid robots become a mainstream industrial tool or remain a niche solution reserved for specific environments.

Automation, News, Policy & Regulation, Robots & Robotics

Researchers at Texas A&M University have developed an AI-powered robotic dog designed to operate in complex, unpredictable environments using memory-driven navigation and human-like decision-making. Built by graduate engineering students, the robot is capable of seeing, remembering where it has been, and responding dynamically to new situations. The system is aimed primarily at search-and-rescue and disaster response missions, where conditions are often chaotic and GPS signals are unavailable.

Unlike conventional robotic systems that rely on pre-mapped environments or simple obstacle avoidance, the robotic dog integrates vision, memory, and language-based reasoning. It understands voice commands, analyzes camera input in real time, and plans routes autonomously. The developers say this combination allows the robot to behave more like a human responder than a traditional machine.

Memory-Driven Navigation With Multimodal AI

At the core of the system is a memory-driven navigation architecture powered by a custom multimodal large language model. The model interprets visual data captured by onboard cameras and combines it with stored environmental memory to make navigation decisions. This enables the robot to recall previously traveled paths and reuse them, improving efficiency and reducing redundant exploration.

A hybrid control structure allows the robot to balance reactive behavior with high-level planning. It can quickly respond to immediate hazards, such as avoiding collisions, while simultaneously reasoning about longer-term navigation goals. According to the research team, this mirrors how humans navigate unfamiliar spaces by combining instinctive reactions with deliberate planning.

“Some academic and commercial systems have integrated language or vision models into robotics,” said Sandun Vitharana, an engineering technology master’s student involved in the project. “However, we haven’t seen an approach that leverages MLLM-based memory navigation in the structured way we describe, especially with custom pseudocode guiding decision logic.”

The robot’s navigation system was designed specifically for unstructured and unpredictable environments, such as disaster zones or remote areas. Traditional autonomous navigation methods often struggle in these conditions due to changing layouts, debris, and limited visibility.

From Disaster Response to Broader Applications

The project was led by Vitharana and Sanjaya Mallikarachchi, an interdisciplinary engineering doctoral student, with guidance from faculty at Texas A&M University. With support from the National Science Foundation, the team explored how multimodal AI models could be deployed at the edge, rather than relying on cloud-based processing.

“Moving forward, this kind of control structure will likely become a common standard for human-like robots,” Mallikarachchi said.

Beyond search-and-rescue operations, the researchers see broader potential applications for the technology. The robot’s ability to navigate large, complex spaces could make it useful in hospitals, warehouses, and other industrial facilities. Its memory-based system may also assist people with visual impairments, conduct reconnaissance in hazardous areas, or support exploration tasks where human access is limited.

Dr. Isuru Godage, an assistant professor in the Department of Engineering Technology and Industrial Distribution, emphasized the importance of deploying advanced AI directly on robotic platforms. “The core of our vision is deploying MLLM at the edge, which gives our robotic dog the immediate, high-level situational awareness previously impossible,” Godage said. “Our goal is to ensure this technology is not just a tool, but a truly first responder-ready system for unmapped environments.”

The robot was recently demonstrated at the 22nd International Conference on Ubiquitous Robots, where the team presented experimental results and system design details. The work highlights how advances in multimodal AI are beginning to reshape autonomous robotics, moving systems closer to adaptive, human-like behavior in real-world conditions.

Artificial Intelligence (AI), News, Robots & Robotics

Engineers at MIT have demonstrated an aerial microrobot capable of flying with speed and agility approaching that of real insects, overcoming a long-standing limitation in micro-scale robotics. The insect-sized robot can execute aggressive maneuvers, including repeated midair somersaults, while maintaining stability even in windy conditions. The breakthrough was enabled by a new artificial intelligence-based control system that dramatically improves flight performance.

Tiny flying robots have long been viewed as promising tools for applications such as search-and-rescue operations, where they could navigate through narrow gaps and unstable environments inaccessible to larger drones. Until now, however, aerial microrobots have been constrained to slow, smooth flight paths due to limitations in control systems and onboard computation.

AI-Based Control Unlocks Agile Flight

The MIT team designed a two-part, AI-driven control framework that balances high performance with real-time efficiency. Compared to previous versions of the robot, the new system increased flight speed by approximately 450 percent and acceleration by about 250 percent. In testing, the robot completed 10 consecutive somersaults in just 11 seconds while staying within a few centimeters of its intended trajectory.

“We want to be able to use these robots in scenarios that more traditional quadcopter robots would have trouble flying into, but that insects could navigate,” said Kevin Chen, an associate professor of electrical engineering and computer science at MIT and co-senior author of the study. “With our bioinspired control framework, the flight performance of our robot is comparable to insects in terms of speed, acceleration, and pitching angle.”

The microrobot itself is roughly the size of a microcassette and weighs less than a paperclip. It uses flexible artificial muscles to rapidly flap oversized wings, generating lift and enabling sharp directional changes. While recent hardware improvements made more aggressive flight possible, earlier versions of the robot relied on manually tuned controllers that limited overall performance.

From Predictive Planning to Real-Time Action

To overcome these constraints, Chen’s team collaborated with researchers led by Jonathan P. How in MIT’s Department of Aeronautics and Astronautics. Together, they developed a two-step control strategy that combines predictive planning with machine learning.

The first component is a model-predictive controller that uses a mathematical model of the robot’s dynamics to plan optimal flight maneuvers. This controller can account for uncertainties in aerodynamics, external disturbances such as wind, and physical limits on force and torque. While powerful, it is too computationally demanding to run continuously in real time.

To address this, the researchers used the predictive controller to train a deep-learning-based policy through imitation learning. The resulting AI model captures the behavior of the high-performance planner in a compact form that can run extremely fast, allowing the robot to respond instantly to changing conditions.

“If small errors creep in and you try to repeat a flip multiple times, the robot will crash,” How said. “We need robust flight control.”

Toward Real-World Deployment

The robot’s agility allowed it to demonstrate flight behaviors commonly seen in insects, including rapid pitch-and-stop maneuvers known as saccades. These movements help insects stabilize their vision and orient themselves in space, and similar capabilities could support future onboard sensing.

“This bio-mimicking flight behavior could help us when we start putting cameras and sensors on board the robot,” Chen said. “This work signals a paradigm shift. It shows that we can build control architectures that are both high-performing and computationally efficient, even at insect scale.”

The research team plans to focus next on adding onboard sensors and autonomy so the robots can operate outside laboratory motion-capture environments. Coordinated flight among multiple microrobots and collision avoidance are also areas of active investigation.

The study was published in the journal Science Advances and highlights how advanced AI control methods are enabling microrobots to move beyond experimental demonstrations toward practical, real-world use.

Artificial Intelligence (AI), News, Robots & Robotics, Science & Tech