Impact and effort: A menu of AI and autonomy options for AUKUS Pillar II

Executive summary

• Public commentary around AUKUS Pillar II has recently coalesced around the notion that the initiative remains a “solution in search of a problem” with an agenda too vast to progress the “quick wins” it was designed to deliver. Furthermore, some current and former officials have suggested the scope of Pillar II’s eight priority areas be narrowed to focus on a specific project or mission.
• To be sure, AUKUS Pillar II is making progress. Vital work has already been undertaken by the three participating governments to align export controls, procurement and certification settings, conduct electronic warfare and maritime innovation challenge competitions and deepen interoperability between the three countries, including in artificial intelligence and autonomy. Meanwhile, the announcement of a marquee project to demonstrate greater ‘proof of concept’ for trilateral codevelopment has been tipped as imminent.
• Notwithstanding these successes, AUKUS Pillar II outcomes should urgently address three major operational challenges confronting the three partners in the Indo-Pacific: buying back mass to offset China’s numerical advantages; achieving resilient Combined Joint All-Domain Command and Control (CJADC2); and supporting efficient logistics in a contested environment.
• This policy brief seeks to define and then match high-impact operational challenges with existing or near-term technological solutions available through AUKUS Pillar II. It focuses on matching existing or high technology readiness level (TRL) AI and autonomy solutions to these challenges, with the aim of presenting a menu of high-impact, reduced effort options for Pillar II’s ‘growth stage’.

This brief seeks to define and then match high-impact operational challenges with existing or near-term technological solutions available through AUKUS Pillar II.

Introduction

AUKUS Pillar II was conceptualised to deliver deterrent effects far quicker than AUKUS Pillar I (the nuclear-powered submarine component), which will not enter service until the early 2030s at best. Though significant work has been completed to align export control regimes and regulatory settings (among other initiatives), a growing number of experts and industry professionals have suggested that Pillar II remains a “solution in search of a problem,” with an agenda too vast to deliver the “quick wins” intended. Similarly, some current and former officials have suggested the scope of Pillar II’s eight workstreams — six technological and two functional — be temporarily narrowed to focus on a specific project or mission, which could better galvanise progress, or at least that which is publicly visible.

To support greater progress during Pillar II’s ‘growth stage’, this policy brief outlines a menu of project options, drawing inspiration from existing initiatives to drive timely wins. The brief borrows a conceptual ‘impact-effort’ prioritisation framework from business practice, sometimes used during a company’s product development phase to support the efficient allocation of attention and capital. In doing so, it aims to match high-impact operational challenges with reduced-effort technological solutions — in other words, operational challenges likely to be faced in any hypothetical Indo-Pacific contingency, as well as technological solutions that already exist or have a high technology readiness level (TRL).

The brief begins with the challenge of buying back mass to effectively compete with the People’s Republic of China (PRC), followed by resilient Combined Joint All-Domain Command and Control (CJADC2) and finally, supporting efficient logistics. The recommendations suggest how the AUKUS partners could operationalise a menu of options for AUKUS Pillar II’s ‘AI and autonomy’ workstream. These are intended to enhance the “speed and precision of decision-making processes to maintain a capability edge and defend against AI-enabled threats.” This approach by no means suggests Pillar II’s seven other priority areas should be eschewed in favour of an AI and autonomy-only approach. Rather, it has been adopted because of the relevance of AI/autonomy technologies to the severity of the strategic environment and associated operational challenges, and the evident interest displayed to date by the three countries in prioritising this area of collaboration.

Identifying the operational challenges and targeted technological approach

Buying back mass

The first challenge is the need to enhance mass, resilience and qualitative advantage against a potential adversary with far greater quantity, owing to the largest military buildup in 80 years. Over the last two decades, the People’s Liberation Army has undergone significant military modernisation, and while its technological sophistication lags the United States in some areas, quantitatively it has rapidly closed the gap and will likely enjoy additional geographic benefits in an Indo-Pacific contingency that may erode US and allied edge. Official strategies and statements from each of the AUKUS partners have acknowledged this is a distinct challenge. In Australia’s case, the 2023 Defence Strategic Review noted Australia’s “regional technology capability edge,” enabled by the alliance with the United States, has been eroded due to “military modernisation in the region,” arguing that the Department of Defence must focus on asymmetric advantages to achieve “parity or qualitative advantage” in key military technology areas. The 2024 National Defence Strategy doubled down on this theme, echoing that achieving asymmetric advantage would involve innovative methods of pitting “strength against weakness, at times in a non-traditional or unconventional manner.” The development of emerging technologies via AUKUS Pillar II is central to this effort.

An unmanned aerial vehicle sits ready for use during a Defense Innovation Unit Blue UAS Refresh Challenge at Marine Corps Air Ground Combat Center, California, November 2024.Source: US Department of War

In the United States, former officials, including Deputy Secretary of Defense under the Biden administration, Kathleen Hicks, have discussed the need to hedge against the PRC’s advantage in scale, in terms of larger numbers of ships, missiles and forces. In some respects, the Trump administration has picked up where the Biden administration left off, indicating in their interim national security guidance that deterring China remains a priority, and progressing reform efforts to speed up acquisition and innovation to compete militarily. Much as with the Biden-era Replicator initiative — which proposed to deliver unmanned systems at “a scale of multiple thousands, across multiple warfighting domains” — these themes are central to the Pentagon’s July 2025 Memorandum on Unleashing US Military Drone Dominance and November 2025 Acquisition Transformation Strategy. Similarly, the United Kingdom’s 2025 Strategic Defence Review identified PRC military modernisation, including a “vast increase in advanced platforms and weapons systems” as a persistent challenge, requiring “renewed deterrence” via military modernisation and industrial collaboration with key Indo-Pacific allies.

Analysis from current and former AUKUS country officials and experts suggests there are existing or high TRL AI and autonomy solutions that could be progressed in response to this challenge. In April 2025, Admiral Paparo, Commander of US Indo-Pacific Command (INDOPACOM), stated that AI-enabled autonomous systems have already demonstrated asymmetric advantage, overcoming quantitatively superior opponents manned with legacy technology. Specifically, Paparo cited the acute relevance of “autonomous aerial systems, AI-driven undersea vehicles, and enabling technologies” to addressing operational challenges in the Indo-Pacific. Likewise, Michèle Flournoy, former Under Secretary of Defense for Policy has argued that integrating AI-enabled, unmanned systems into US force structure is “the only way we buy back mass in the relevant timeframe of the next four or five years.” These views have been supported by arguments set forth in a recent Center for a New American Security report, which identified affordable autonomous systems — those enabled by AI software and algorithms — as critical to addressing the quantitative disadvantage the United States and allies will face in an Indo-Pacific contingency. The readiness of these solutions has been at least partially tested and proven over the past few years in the Ukraine war.

That the United States has already acquired relevant technologies through the Replicator initiative suggests an AUKUS Pillar II version of this workstream — or an adjacent AI and autonomy effort with a designated acquisition and procurement strategy — could be rapidly progressed, piggybacking off existing workflows. For the AUKUS partners, a Pillar II all-domain attritable autonomy (ADA2) project, described in the context of Replicator as intended to deliver unmanned systems “at a scale of multiple thousands, across multiple warfighting domains,” may ensure that greater defence production capability and platform capacity exists outside of the United States and closer to the Indo-Pacific, while also building redundancy through geographically distributed supply chains. A Pillar II ADA2 project could provide an opportunity for Australia and the United Kingdom to more deliberately support the United States’ drone manufacturing uplift, noting capacity challenges a Pillar II Replicator may present to the US defence industrial base. Such an initiative would be a natural transition from several multi-domain AI and autonomy exercises and demonstrations that the three countries have pursued, including the Maritime Big Play series, the Trusted Operation of Robotic Vehicles in a Contested Environment (TORVICE) initiative, and resilient and autonomous artificial intelligence technology (RAAIT) experiments conducted in the context of the US-led Project Convergence series.

Resilient Combined Joint All Domain Command and Control (CJADC2)

The PRC possesses anti-access area denial weapons in both the kinetic domain and in cyberspace, posing a significant challenge to US and allied command and control in a potential regional contingency. In recent years, this has led each of the AUKUS countries to elevate the concept of ‘resilience’ in thinking about how to deter or at least withstand the growing threat of electronic warfare attacks, cyber warfare attacks and attacks on space assets, reflected in a growing number of exercises, experiments and initiatives focused on those domains. Technological advancements have also continued to compress the timeline for decision-making on the battlefield, with US Department of Defense (DoD) officials predicting that operational decisions will likely be made in seconds — rather than minutes — on the battlefields of the future. These pressures have given rise to the joint all-domain command and control concept (JADC2), which first emerged in 2019 and has been carried forward by a cross-functional DoD team. In 2023, the JADC2 moniker was expanded to account for interoperability with allies, becoming Combined Joint All Domain Command and Control (CJADC2), in line with the Biden Administration’s focus on leveraging allies and partners. CJADC2 refers to an integrated tactical, cloud-like architecture enabling the sharing of intelligence, surveillance and reconnaissance data, which transmits across networks and supports rapid decision-making with allies in contested environments. One of the benefits of this network would be the construction of a common operating picture among allies, which would still preserve each ally’s sovereign control over policy options in response to any given scenario.

The People’s Republic of China possesses anti-access area denial weapons in both the kinetic domain and in cyberspace, posing a significant challenge to US and allied command and control in a potential regional contingency.

AI offers relatively near-term promise for the build out of resilient CJADC2 architecture, even if some officials and departments remain concerned about the maturity and price tag of this technology. Former senior US officials have suggested it is realistic from a technological point of view to expect that AI-enabled CJADC2 architecture will exist within a five-year time period. Looking at relevant workstreams over the past 18 months or so, the US DoD appears to have already made progress on this front. For example, during a February 2024 speech, then-Deputy Secretary of Defense Kathleen Hicks noted the DoD was investing to “enable and scale data and AI across (the) DoD” to achieve CJADC2. Hicks noted “minimum viable capability” had been achieved, including via INDOPACOM’s Joint Fires Network, which uses AI to aggregate battlefield data across the combatant commands to facilitate faster decision making. The initiative has been dubbed a “pathfinder” for CJADC2 and is expected to be operational in 2025. In September 2025, the Trump administration added its own flavour, announcing a project called ‘Mission Network-as-a Service’, which intends to consolidate 17 military networks into one to communicate with allies, in support of CJADC2. Furthermore, the AUKUS countries have also pursued a number of activities focused on building the means for AI-assisted “collective multi-domain awareness” using connected land, air and sea-based manned and unmanned systems. The examples cited above demonstrate ample vision and some progress towards developing AI-enabled CJADC2 solutions. Though these largely remain stove-piped at present, one version of their end-state could be a more deliberate trilateral project.

Contested logistics

During the Second World War, military logistics complicated the US campaign in the Indo-Pacific, prompting General Douglas MacArthur to proclaim, “victory is dependent upon the solution of the logistics problem.” In a future hypothetical contingency, US and allied in-theatre logistics, maintenance and sustainment would be complicated by the PRC’s growing military capacity and power projection capabilities, a challenge often referred to as “contested logistics.” Indeed, as documented in a prior United States Studies Centre report, tackling this problem is one of the primary drivers of the US-led allied industrial integration agenda. Two case studies illustrate the potential benefits existing off-the-shelf AI and autonomy solutions like remote and predictive maintenance portend for this problem, indicating the potential to drive interim progress through Pillar II while more advanced technological solutions are developed.

An operator on the Royal Air Force Voyager Multi-Role Tanker Transport conducts air refuelling with a US Marine Corps F-35B Lightning II during Exercise Pitch Black 2022Source: Australian Department of Defence

Firstly, a 2024 RAND Europe report argued AI will be transformational for military logistics challenges, specifically highlighting opportunities in “last-mile” logistics — that which is closest to in-theatre personnel — reducing the need for crewed equipment during resupply missions. The report also noted opportunities in supply chain and warehouse management, as well as stockpile optimisation, as prime candidates for immediate action. One prominent historical example of the US military successfully implementing such tools is US Transportation Command’s rapid fielding of the Dynamic Analysis and Replanning Tool (DART) during Operation Desert Storm in 1991. DART used AI to run strategic models of different transportation scenarios to the Gulf, which helped to move tanks and heavy artillery to Saudi Arabia three weeks ahead of schedule, granting the US-led coalition a significant logistical advantage during the campaign. This positive example illustrates the kinds of efficiencies that AI-enabled technology can offer logisticians in challenged or contested scenarios.

More recently, Adarryl Roberts, Chief Information Officer of the US Defense Logistics Agency (DLA), discussed various advantages that AI promises or is already delivering to defence logistics, including existing use cases in supply chain risk management, such as revealing or predicting supply chain bottlenecks and recommending alternative supply chain routes. Moreover, Roberts pointed to the DLA’s Long-Term Contracts Negotiations Analytics tool, which uses past data to run advanced simulations that quantify demand variability for the agency and their suppliers, enabling the DLA to place risk-adjusted orders, ensuring war fighters are adequately resourced and suppliers receive sustainable contracts, thereby promoting the long-term health of the defence industrial base. These existing use cases point to near-term AI and autonomy-enabled opportunities for logistics challenges, bolstering the case for its adoption as a Pillar II project. This argument is underscored by the US-led International Joint Requirements Oversight Council’s interest in “enhanced logistical networks” to support “sustainment for operations in contested environments.”

Policy recommendations

To progress a menu of high-impact, reduced effort AI and autonomy options for Pillar II, the AUKUS partners should:

• Compliment Replicator with a similar AUKUS Pillar II effort.
  Establish an AUKUS Pillar II version of the US Department of Defense’s Replicator initiative to field a mass of interoperable, autonomous, attritable capabilities, with an initial focus on advanced swarming techniques. A Pillar II version of Replicator would ensure excess capacity exists outside of the United States and closer to the primary theatre of interest, while also building redundancy through distributed supply chains. Viewed another way, this project could provide an opportunity for Australia and the United Kingdom to more deliberately support US ADA2 manufacturing uplift. This initiative would be a natural progression of the series of unmanned-centric exercises, experiments and other initiatives conducted by the AUKUS countries in recent years, which have demonstrated a growing level of conceptual maturity and platform sophistication across the three nations’ industrial bases. This will require further work to “AUKUS-ify” national systems and processes around acquisition, procurement and production, something which the three partners have sought to progress through their series of innovation challenges.
• Examine how AUKUS Pillar II could be useful to US Indo-Pacific Command’s (INDOPACOM) Joint Fires Network.
  INDOPACOM has already completed substantial work on the Joint Fires Network, an interservice version of CJADC2 designed to use AI as part of an integrated, tactical, cloud-like architecture, enabling the sharing of intelligence, surveillance, and reconnaissance data, which transmits across networks and supports rapid decision-making, including targeting, with allies in contested environments. The Joint Fires Network is a formal acquisition program office and has demonstrated at Exercise Valiant Shield with Indo-Pacific partners — including Australia and the United Kingdom. Building on this momentum, the AUKUS partners should consider how Pillar II could be more deliberately integrated into the slipstream of the Joint Fires Network. One possible outcome could be exercising Joint Fires Network at Maritime Big Play, a Pillar II-specific trilateral experiment and exercise series which already involves greater coordination on maritime surveillance and strike. This would represent progress towards CJADC2 — specifically for AUKUS — and build on AUKUS resilient and autonomous artificial intelligence technology (RAAIT) experiments to support joint domain awareness and targeting for strike missions.
• Prioritise ‘contested logistics’ as an upcoming challenge competition area.
  Given the importance of military logistics to successful campaigns and the challenging environment the AUKUS partners face in this regard in the Indo-Pacific, ‘contested logistics’ should be prioritised as an AUKUS challenge competition area. Indeed, in September 2024, AUKUS defence principals identified the development of “enhanced logistical networks” capable of delivering “persistent support and sustainment for operations in contested environments” as one of four trilaterally determined key operational problems that would inform trilateral capability requirements for AUKUS Pillar II projects. There are some promising startups across the AUKUS countries doing impactful work in this area, such as Gallatin AI, a company that uses AI to compress and support resupply missions. Sending stronger demand signals to industry by spotlighting ‘contested logistics’ as an upcoming challenge competition area will lead to further capital investments, faster product delivery, and the projection of more confident deterrence through exercise.