In April 2026, the Defense Advanced Research Projects Agency (DARPA), through its Microsystems Technology Office (MTO), released a Request for Information (RFI) titled “Materials for Physical Compute in Untethered Robotics” (DARPA-SN-26-76). This RFI signals DARPA’s early-stage interest in advancing the next generation of autonomous robotic systems capable of operating in communication-denied and dynamically changing environments. At a high level, DARPA is exploring how embedding intelligence directly into materials, rather than relying on centralized computing architectures, could fundamentally transform robotic performance and enable faster, more adaptive, and energy-efficient systems for national security applications.
RFI Seeks Insight Into “Physical Intelligence”
At its core, the RFI seeks insight into “physical intelligence,” defined as the encoding of sensing, actuation, learning, adaptation, and decision-making directly within a robot’s physical structure. DARPA highlights the limitations of current robotics approaches, which depend heavily on increased sensor density and data transmission to centralized processors.
These approaches introduce latency, increase power consumption, and reduce operational security. In contrast, this effort emphasizes material-centric innovation, where intelligence is distributed throughout the system at the component or material level. Respondents are encouraged to propose disruptive concepts that move beyond incremental improvements and instead demonstrate transformative potential in how robots perceive, adapt, and act in unconstrained environments.
RFI Outlines Two Primary Research Focus Areas
First, Actuation and Sensing centers on developing advanced materials and architectures that integrate sensing, processing, and actuation into unified systems. This includes innovations in soft robotics, embedded proprioception (self-awareness of system state), and energy-efficient materials capable of harvesting, storing, and utilizing energy.
Second, Dynamic and Adaptive Closed-Loop Compute focuses on embedding computational capabilities directly within materials to enable real-time, low-latency decision-making without reliance on traditional central processing units (CPUs) or graphics processing units (GPUs). DARPA is particularly interested in multi-modal systems capable of processing diverse environmental inputs such as pressure, temperature, and light simultaneously, as well as systems that can adapt instantaneously to changing conditions.
Together, these areas reflect DARPA’s broader goal of enabling embodied intelligence, where perception, computation, and action are inseparable and co-designed at the hardware level. DARPA also explicitly noted that incremental hardware improvements, purely software-based AI approaches, and conventional CPU/GPU-centric architectures are not of interest under this RFI, underscoring the agency’s focus on transformative hardware-enabled physical intelligence.
Several key dates are associated with this RFI:
- April 27, 2026: RFI released by DARPA MTO
- May 27, 2026 at 2:00 PM ET: Deadline for submission of RFI responses
- June 10, 2026: Anticipated date for workshop invitations to selected respondents
- June or July 2026: DARPA-hosted in-person workshop in Arlington, Virginia to discuss submissions and future research directions
As an RFI, this announcement does not include formal funding allocations, award sizes, or defined periods of performance. However, it is important to recognize that DARPA RFIs often precede formal Broad Agency Announcements (BAAs), which may result in multiple awards supporting multi-year research efforts. The scope and ambition outlined in this notice suggests that any future program could involve significant investment in early-stage, high-risk, and high-reward research, likely spanning multiple performers across academia, industry, and government laboratories.
Ultimately, the technologies envisioned in this RFI have the potential to extend far beyond defense applications. Advances in material-based computation and embodied intelligence could enable new classes of autonomous systems for disaster response, medical robotics, and assistive technologies, where adaptability, resilience, and real-time decision-making are critical to improving human safety and health outcomes. Organizations with capabilities in advanced materials, robotics, and embedded systems should view this RFI as a strategic opportunity to shape a future DARPA program.
If your company is considering applying for federal funding, your journey starts here. EverGlade is a national advisory firm helping innovators navigate the federal funding ecosystem. We support companies across the funding lifecycle, from early-stage strategy through proposal development, negotiations, and post-award execution, connecting breakthrough innovation with non-dilutive funding. We support programs across federal agencies including ARPA-H, BARDA, DARPA, DTRA, NIAID, CPE CBRND, DOE, and DFC, among others.
