Today’s advances in robotics are often driven by breakthroughs in artificial intelligence, machine learning, and perception. However, in complex and constrained environments, the limiting factor is often hardware rather than software. Systems that rely on continuous data processing, high-bandwidth communications, and centralized computing can face delays, power constraints, and vulnerabilities that limit performance or completely prevent mission success.
DARPA seeks to address these challenges by building intelligence directly into the physical materials of robotic systems. A new request for information (RFI) calls on the research community to help define new classes of materials capable of mixed sensing, adaptation, and real-time operation without relying on continuous external computation or communication links.
While the RFI itself is exploratory, it is a first step toward a more immediate opportunity: an invitation-only in-person workshop scheduled for summer 2026. Selected participants will have the opportunity to present their ideas, collaborate with DARPA, and inform future program direction.
Rethinking where intelligence resides
Commercial robotics is primarily focused on building systems that can work together with humans, often emphasizing familiar shapes and interfaces. National security applications require something different.
Defense robotic systems must operate in extreme, unpredictable, and hostile environments with limited communications and little opportunity for human intervention. In these situations, performance is not defined by the amount of data the system can process, but rather by how quickly and reliably the system can respond.
Meeting these demands requires changing where intelligence resides.
DARPA researches physical intelligence, an approach that embeds sensing, computation, and actuation directly into materials, components, and structures. Rather than routing information through a centralized processor, future systems may respond through physical design, enabling faster, more efficient, and more resilient operation in dynamic environments.
“Today’s robots are often limited by the need to sense, process, and act as separate steps,” said DARPA Program Manager Julian McMorrow. “We’re interested in breaking that loop by building intelligence directly into the hardware, allowing systems to respond in real time without relying on constant movement of data.”
This change has the potential to enable faster, more energy-efficient, and more resilient robotic systems in complex, unstructured environments.
Focus on materials, not machines
The RFI targets fundamental advances at the material, component, and kernel levels and focuses on two areas:
- Actuation and sensing: DARPA is interested in materials and structures that integrate sensing, actuation, and even control elements onto the same physical substrate, allowing robots to perceive and interact with their environments with greater speed, efficiency, and adaptability.
- Dynamic and adaptive closed-loop computation: DARPA is exploring materials that can perform calculations directly, rather than relying on centralized processors or large data flows. Incorporating computing into sensors and actuators has the potential to enable real-time decision-making with minimal latency, reduced power demands, and the ability to continuously adapt to changing conditions.
Together, these areas point to a new class of systems in which perception, decisions, and actions are tightly integrated at the hardware level.
DARPA is not looking for incremental improvements or system-level concepts separate from hardware enablement. Instead, the focus is on breakthroughs that have the potential to fundamentally reshape what robotic systems can do.
Furthermore, while industry has emphasized human-like form factors designed to operate in human environments, what is interesting here are systems that are optimized for mission needs. Depending on the application, this includes smaller, larger, softer, or structurally unconventional designs that prioritize performance and adaptability over familiarity.
