Pentagon wants miniature deep-ocean drones

The Pentagon’s top research arm, Defense Advanced Research Projects Agency (DARPA), launched a new research program on April 23, 2026, calling on industry and academia to develop autonomous underwater vehicles capable of reaching full-ocean depth — the deepest points on the planet — at a fraction of the size of anything that exists today.

The program, called Deep Thoughts, is a solicitation from DARPA’s Tactical Technology Office, released under funding opportunity number DARPA-PS-26-05. It is a single-phase, 24-month, hardware-centric effort. DARPA is not looking for studies or white papers. The agency wants functional prototypes — hardware that can actually dive to full-ocean depth by the time the program concludes. Multiple awards are anticipated, and the instruments used will be Other Transaction agreements for prototype development. Abstracts are due May 21, 2026, with oral proposal packages tentatively due July 20, 2026. Estimated program start is November 2026.

The solicitation divides work into two technical areas. Technical Area 1, designated TA1, involves research conducted at the Controlled Unclassified Information level. Technical Area 2, designated TA2, requires SECRET-level facility and personnel clearances, meaning the work there touches classified military applications. Proposers may submit to one technical area or the other — not both. Participation is restricted to U.S. citizens and U.S. organizations. All proposers must comply with export control laws, including the International Traffic in Arms Regulations and the Export Administration Regulations, and must implement a program security plan covering operational security, foreign supply chain risks, and public affairs.

To receive the full solicitation package — including classified addenda, cost proposal templates, and task description documents — interested parties must submit a formal request no later than May 14, 2026. Requests submitted after that date will not be accepted. Proposers who fail to request the package are considered non-compliant and cannot submit.

What DARPA is actually asking for is genuinely ambitious. Full-ocean depth means the ability to operate at the deepest points of any ocean on Earth — pressures that crush conventional structures and defeat most engineering approaches that work fine at shallower depths. Current state-of-the-art autonomous underwater vehicles that can operate at extreme depths tend to be large, expensive, slow to produce, and difficult to field in meaningful numbers. DARPA wants to break all of those constraints simultaneously.

The program targets several interlocking technical challenges. Pressure vessel design is central — building a hull or enclosure that survives the crushing pressures of full-ocean depth without becoming so heavy or large that it defeats the purpose of the vehicle. Materials science is equally critical. Novel alloys, composite materials, and structural geometries that do not exist in current production systems are explicitly called out as areas of interest. The agency is not asking industry to optimize existing approaches. It is asking for fundamentally different ones.

Advanced manufacturing methods — the kind that allow complex geometries to be produced quickly and at lower cost — are a focus area alongside the materials themselves. DARPA’s emphasis on producibility signals that the end goal is not a one-of-a-kind demonstrator. The agency wants vehicles that could eventually be built in quantity, which means the manufacturing process has to scale.

An AUV that can reach full-ocean depth is only useful if it can carry sensors, communications equipment, or other mission-relevant hardware and operate them reliably under extreme conditions. The solicitation calls for non-traditional approaches to subsystem and component architecture — specifically approaches that enable what DARPA describes as free-form design, structural consolidation, and multi-functionality. In plain terms: the agency wants performers to rethink how the internal components of the vehicle are arranged and integrated, rather than simply miniaturizing existing designs.

Deep Thoughts calls for a multi-level secure digital engineering ecosystem that supports collaborative design across classification levels, continuous integration and development, intellectual property protection, and rapid iteration. The design-build-test-learn cycle DARPA envisions is fast-paced by intent. Performers are expected to develop new concepts, test them to their limits, learn from the results, and apply those lessons to the next iteration — all within the 24-month performance window.

Submarine cables carrying global communications and financial data run across ocean floors. Adversary nations have invested heavily in undersea surveillance, mine-laying capabilities, and submarine-launched weapons. The ability to access and operate at full-ocean depth — autonomously, affordably, and at scale — carries significant implications for intelligence gathering, infrastructure monitoring, mine countermeasures, and undersea domain awareness. DARPA’s Deep Thoughts program sits squarely at the intersection of those priorities.

The 24-month timeline is aggressive for hardware development at this level of technical difficulty. DARPA programs are deliberately set up that way. The agency funds high-risk, high-reward research precisely because the conventional acquisition system moves too slowly to tackle problems at the frontier of what is physically possible. Deep Thoughts is asking performers to compress years of development into two years of hard engineering — and to deliver hardware, not promises, at the end.