1,000+ Opportunities
Find the right grant
Search federal, foundation, and corporate grants with AI — or browse by agency, topic, and state.
Artificial Intelligence Exploration is DARPA's high-risk, high-speed on-ramp for new AI concepts, and it exists specifically to solve the problem that conventional defence contracting is far too slow for a field moving at the pace of AI research.
Rather than running a single annual competition, DARPA maintains an umbrella Program Announcement under which individual AIE Opportunities are released on a rolling basis by programme managers across the agency's offices, each targeting a narrow technical question.
The mechanism is aggressive by government standards: DARPA's stated goal is to go from Opportunity announcement to award start in roughly three months, with awards of up to 1,000,000 US dollars supporting 18 months of work structured as a Phase 1 Feasibility Study followed by an optional Phase 2 Proof of Concept.
The technical centre of gravity is what DARPA calls third wave AI, meaning systems that can contextually adapt, explain their reasoning and operate reliably outside their training distribution, as opposed to the statistical pattern matching of second wave machine learning.
Past AIE Opportunities have covered adversarial machine learning, AI for cyber defence, physics-informed learning, symbolic and neuro-symbolic reasoning, and AI for intelligence analysis. The practical implication for applicants is that AIE is not something to prepare for in the abstract: teams should monitor SAM.
gov and DARPA's opportunities page continuously, because each Opportunity has its own short response window and the only way to win is to already have the technical idea in hand when the relevant Opportunity drops.
Get a weekly digest of new grants like this
A free weekly digest of new foundation and federal funding opportunities as they're added to Granted. Unsubscribe anytime.
Or search similar grants →According to the current listing, eligibility includes: AIE is deliberately open to a wide proposer base: US universities, non-profit research organisations, small businesses, large defence primes and non-traditional commercial AI firms may all respond, and DARPA actively encourages performers who have never worked with the Department of Defense before. Foreign participation is generally restricted and individual Opportunities may impose citizenship or facility clearance requirements, but all AIE proposals must be unclassified, which materially lowers the barrier for commercial AI teams without cleared personnel. Every proposal must address both project phases in a single submission, with clearly defined Phase 1 milestones that DARPA can use as the basis for the Phase 2 option decision; a proposal that treats Phase 2 as vague follow-on work will not be competitive. Because most AIE efforts are awarded as Other Transactions for prototype projects under 10 U.S.C. 4022, non-traditional defence contractors are a preferred category and there may be cost-share expectations where only traditional contractors participate. There is no standing deadline for the umbrella Program Announcement itself; applicants respond to individual AIE Opportunities, each of which sets its own abstract and proposal due dates, typically measured in weeks rather than months. Teams should register in SAM.gov well in advance, since the response windows are too short to complete registration reactively, and should read the specific Opportunity carefully because eligibility, classification and cost-share terms vary from one Opportunity to the next. Confirm the full requirements in the official notice before applying.
The current listing shows individual AIE awards are worth up to 1,000,000 US dollars each, so amount_min is recorded as 0 and amount_max as 1,000,000 US dollars; the exact ceiling for any given effort is set in the specific AIE Opportunity document rather than in the umbrella Program Announcement, and some Opportunities set a lower cap. The money is split across two sequential project phases that are proposed together: a Phase 1 Feasibility Study, which is the base effort, and a Phase 2 Proof of Concept, which DARPA exercises as an option only if Phase 1 milestones are met. The whole arc is compressed into 18 months, which is the defining budgetary constraint: this is not a programme where a team can propose a slow ramp, hire over a year and deliver in year three. Budgets should be built around a small senior team working immediately, with compute and data acquisition costs front-loaded into Phase 1 so that the go/no-go decision has real evidence behind it. Because most AIE awards are made as Other Transactions for prototype projects rather than as grants or standard contracts, cost-sharing and intellectual property terms are negotiable in ways that traditional FAR-based instruments do not permit, which is often more valuable to a small AI company than the headline dollar figure. Verify award ceilings, matching requirements, and allowable costs in the official notice.
DARPA Artificial Intelligence Exploration (AIE) Program Announcement for Rapid Feasibility Studies in Third Wave AI is funded by Defense Advanced Research Projects Agency (DARPA). Verify program details on the funder's official page before applying.
Start from the official opportunity page linked in this listing — it carries the sponsor's submission instructions.
The U.S. Army's SBIR/STTR Artificial Intelligence/Machine Learning Open Topic provides non-dilutive funding to small businesses developing AI and machine learning solutions for defense applications such as supply chain management, logistics coordination, target identification, and modeling and simulation. Awards follow the standard SBIR/STTR structure: Phase I establishes technical feasibility (typically up to about $250,000) and Phase II funds development and demonstration (typically up to about $2,000,000), with Phase III for commercialization. Open-topic solicitations accept proposals on a periodic basis.
The Office of Naval Research Young Investigator Program, established in 1985, is the Navy's principal early-career research award and one of the largest such programmes in the US federal government, typically providing 750,000 US dollars over three years to faculty in their first or second tenure-track appointment. Machine learning, autonomous operations and dexterous robotics are consistently represented among funded topics: the FY2026 cohort of 23 awardees, drawn from 22 institutions in 11 states and sharing roughly 17,000,000 US dollars, included work on machine learning, autonomous operations, dexterous robotics, advanced sensors and decision superiority alongside more traditional naval science areas such as ocean acoustics and hypersonics. The FY2027 competition is open under announcement N0001426SF004, with white papers due 30 October 2026. What distinguishes ONR's programme from its Air Force and Army counterparts is both the award size and the tight coupling to programme officer interests: proposals are evaluated against the research areas described by ONR's Science and Technology Departments, and the expected practice is that a prospective applicant contacts the relevant programme officer before writing. For AI and robotics researchers specifically, ONR's interests skew toward autonomy in contested and communications-denied maritime environments, human-machine teaming under uncertainty, multi-agent coordination, and machine learning on sparse, noisy or adversarially corrupted sensor data, which is a meaningfully different framing from the data-rich settings that dominate academic machine learning.
DARPA's TIAMAT (Transfer from Imprecise and Abstract Models to Autonomous Technologies) program develops sim-to-real transfer techniques that train autonomous systems in low-fidelity, abstract simulations and reliably deploy them in unpredictable physical environments. Research targets foundations for rapid sim-to-real generalization, robustness to physics-model error, and transfer learning across morphology and sensor configurations. The program is structured as two 18-month phases and supports applications including ground autonomy, aerial autonomy, and robotic manipulation in adversarial, novel, or contested settings. Awards have included University of Central Florida ($1.2M) and other academic and industry performers. TIAMAT is a core DARPA investment in foundational autonomy R&D.
DARPA's FY26 Release 6 carries seven SBIR and three STTR topics closing October 21, 2026. TRIAGE-X and ICU-in-a-Box are listed at $2 million each, CORPS at $750,000 — not Phase I feasibility money. Meanwhile DoW FY-27 Release 1 pre-released October 7 with 25 topics, zero of them DARPA's, and a new one-proposal-per-topic rule that took effect October 1. Here is how to read both calendars.
Read articleDPA26BZ06-DV023 is a Direct-to-Phase-II SBIR paying $700,000 over 18 months plus a $500,000 option. The physics demands 256x more transmit power than the systems that qualify you to compete, and DARPA will not accept modeling alone as proof. Here is the eligibility wall, the five engineering problems, and who can realistically win it before the October 21 close.
Read articleRelease 6's SBIR topics got the attention. Its three STTR topics — SHIELDER, fuel-flexible electric propulsion, and hypersonic wind tunnel noise diagnostics — are all Direct-to-Phase-II, all require a research institution to perform at least 30 percent of the work, and all close October 21, 2026. The feasibility gates are the real filter.
Read article