The top-down requirements process Ukraine’s military inherited from Moscow in the 1990s kept headquarters analysts employed but left 87 percent of needs unfulfilled. Today, warfighters get the final say in what gets built. Drones that once relied on GPS and luck now use automated navigation and targeting algorithms to overcome operator error and Russian jamming, raising success rates from 20 percent to 70 percent. The newest generation uses fiber-optic cable for communication to eliminate the threat of electronic interference.

The Pentagon’s approach to weapon development looks more like the one used by Soviet apparatchiks. Requirements officers in the Joint Staff and military services try to guess capability gaps and potential solutions years in advance. By the time these analyses emerge from the Joint Capabilities Integration and Development System (JCIDS) two years later, the threat has changed, technology has marched on, and a different solution is likely needed.

By definition, predictive planning systems such as JCIDS cannot work in a dynamic environment. They define performance metrics before testing a single prototype because they assume cutting-edge defense systems can only arise from dedicated government-led research and development. That approach is now obsolete thanks to the rapid advance and broad availability of militarily-relevant commercial technology.

Ukraine’s successes show how the U.S. Department of Defense could unlock the potential of private-sector innovation through collaborative experimentation between engineers and operators. Instead of funneling their needs through a multi-year staffing process, Ukrainian commanders talk with local drone pilots and data scientists to identify problems and reach out to government offices that can pay for solutions.

Under Kyiv’s innovation model, a new uncrewed system concept can reach the battlefield in months, drawing on commercial AI to quickly adapt flight paths or identify targets in thousands of video streams. For example, a volunteer-driven missile team eschewed extensive predictive analysis and prototyped a new cruise missile in a year and a half — an unthinkable timeline under Ukraine’s previous Soviet-model bureaucracy.

Real-time operator feedback is essential to this approach. It defines what is “good enough” and helps program managers cut through the competing equities that often prevent a system from reaching the field. In less than a year, Ukraine’s military created Delta, a situational awareness system like the elusive Joint All-Domain Command and Control concept that the Pentagon has chased for nearly a decade. Coders started Delta with a single battlefield map and added new modules when soldiers asked for them. Now the system ties together thousands of drones, cameras, satellite feeds, and Western cannon and rocket artillery systems.

Instead of waiting for a glacial interagency process to dictate universal interoperability requirements, Delta’s developers iteratively add new elements and test them in the real fight. During NATO interoperability exercises in 2023, Delta proved the value of this bottom-up approach by sharing data via Link 16 with F-16 jets and integrating with Poland’s TOPAZ artillery fire control software. Delta reflects genuine cross-domain synergy, born out of emergent needs and continuous iteration, not years of staff approvals.

Ukraine’s success is not simply a fluke born out of existential desperation; it’s the logical consequence of removing unnecessary processes and letting warfighters shape the pipeline. While we in the United States prioritize box-checking staffing for documents that meet formatting guidelines and have all the right system views and appendices, Ukraine lets demand drive immediate action. This shift from central planning to distributed innovation has not only kept Ukraine in the fight but also opened the door to realizing advanced integrations like real-time targeting.

The Pentagon should take Ukraine’s combat lessons to heart and fund the work to find solutions for today’s problems. Requirements officers should stop trying to predict the future and begin collecting and refining operational challenges to drive experimentation. And acquisition executives should give innovative program managers and their industry partners the decision space to quickly develop systems that deliver relevant capability, use existing components, and can respond to future enemy countermeasures.

The DOD has experimented with new acquisition pathways and innovation initiatives that have these attributes. But “Band-Aid” solutions that speed up paperwork or create more prototypes don’t address the core problem: a requirements system that prioritizes predictive planning over operational results.

The Pentagon should retire centralized requirements processes such as JCIDS. In their place, the U.S. military services should fund focused campaigns of experimentation that test multiple solutions against clear operational problems, enable rapid learning from failure, and scale what actually works in realistic conditions. Until the DOD abandons its Soviet-style faith in headquarters apparatchiks and embraces structured experimentation driven by warfighters, it will continue to fall behind adversaries who are willing to adapt and learn.

Bryan Clark is a senior fellow and director of the Center for Defense Concepts and Technology at the Hudson Institute, and an expert in naval operations, electronic warfare, autonomous systems, military competitions and wargaming. Previously, he served as special assistant to the chief of naval operations and director of the CNO’s Commander’s Action Group, led studies on the Navy headquarters staff, and was an enlisted and officer submariner in the Navy.

Dan Patt is a senior fellow with the Hudson Institute’s Center for Defense Concepts and Technology, where he focuses on the role of information and innovation in national security. Patt also supports strategy at national security technology company STR and supports Thomas H. Lee Partners’ automation and technology investment practice. Previously, he co-founded and was CEO of Vecna Robotics and served as deputy director for the Defense Advanced Research Projects Agency’s (DARPA) Strategic Technology Office.

The post The Pentagon should abandon Soviet-era centralized planning appeared first on DefenseScoop.

The DoD’s problems are not simply financial — they are structural. The DoD’s global responsibilities are growing in every theater while China can focus its efforts on the Western Pacific. Unlike its Reagan-era buildup, the DoD can no longer afford to field more of today’s multimission planes, ships, or vehicles because they have become too costly to operate and crew. Even with better integration like that being pursued by the U.S. Air Force, China can field targets faster and more cheaply than the DoD can create effective shots on target.

Hedging for uncertainty

For nearly a decade U.S. officials have identified China as the DoD’s pacing challenge and an invasion of Taiwan is its most-stressing scenario. Stopping China’s expanding fleet of troop transports may demand more long-range fires than U.S. forces can deploy, even with CJADC2’s vision of AI-enabled decision aids. The People’s Liberation Army’s multiplying rocket forces further complicate the problem by requiring U.S. ships and aircraft to carry more defensive systems or operate farther away and use longer-range weapons that are more expensive and less numerous.

With outright denial likely off the table, deterrence increasingly depends on creating uncertainty for Chinese leaders that an invasion will go smoothly and incur acceptable losses. But pushing forward ships and aircraft to launch torpedoes or missiles against an invasion fleet is the most predictable U.S. approach — and one China is developing capabilities to counter.

For more than a decade, defense officials argued that the PLA would have difficulty fielding a kill chain that could attack aircraft carriers or mobile ground forces more than a thousand miles from China. In addition to the challenge of delivering precision-guided weapons at that range, the sensors and networks China would need to use could be disrupted or deceived by U.S. electronic warfare systems. 

Arguably the shoe is now on the other foot. To stop an invasion, U.S. ships and aircraft operating 500 miles or more from Taiwan will have to find, target, and engage a Chinese force with possibly very little notice. Fighting on its home turf, the PLA has numerous options to break the DoD’s CJADC2 kill chains and increase the confidence of Chinese leaders that an invasion could succeed.

To create uncertainty for the PLA, the DoD will need a more adaptable approach for stopping an invasion than relying on long-range fires alone. In a new report from the Hudson Institute we argue the Pentagon should field specialized forces to hedge against scenarios like the Taiwan invasion in which the one-size-fits-all U.S. military is unable to win at acceptable risk.

Using short-range uncrewed systems in each domain, a hedge force for Taiwan would act essentially as a mobile minefield, damaging transport vessels and warships in narrow “kill boxes” along the Taiwan coast. Just as important, the hedge force would also make long-range fires more effective by slowing the invasion and compelling PLA escorts to activate their defenses, creating targeting opportunities for U.S. forces to exploit through the Joint Fires Network.

Exploiting today’s technology

Although they achieve cutting-edge performance, the DoD’s connectivity initiatives are largely built on existing capabilities, rather than new development and acquisition. CJADC2’s initial instantiation is a descendant of DoD’s Project Maven and the Joint Fires Network implements work matured by DARPA. Going forward, these efforts can take advantage of an expanding range of commercial offerings — like Ukrainian forces who assembled battle networks from Starlink terminals, subscription satellite services, and open source-derived decision support software.

A hedge force for Taiwan could similarly exploit commercial innovation and existing technologies to quickly reach the field. Uncrewed air, surface, and undersea vehicles today incorporate sufficient targeting and guidance automation to find and engage slow-moving troop transports or warships in confined geographic areas. These systems would not need new advancements in AI or sensors and could operate without human intervention beyond programming their rules of engagement.

Small drones like the Switchblade, Altius 700, or A2LE have demonstrated their effectiveness as loitering munitions or by complementing missile attacks in Ukraine. Uncrewed surface vehicles like the Saronic Cutlass or Magura V5 sank more than a half-dozen ships in Russia’s Black Sea Fleet and could pose a lethal threat to Chinese forces nearing Taiwan. And undersea, vehicles like the Speartooth could place  a torpedo’s-worth of explosives under a vessel’s hull to cause catastrophic damage. Although they may lack the survivability or destructiveness of traditional weapons, these vehicles will damage transports and make them easier pickings for long-range U.S. fires.  

Buying more of today’s ships, aircraft, or weapons even with better integration, will not solve the DoD’s fundamental geographic, strategic, and fiscal disadvantages in stopping a Chinese invasion of Taiwan. To turn the tables and regain the advantage, the Pentagon needs a hedge force that will ensure every shot counts.

Bryan Clark is a senior fellow and director of the Center for Defense Concepts and Technology at the Hudson Institute, and an expert in naval operations, electronic warfare, autonomous systems, military competitions and wargaming. Previously, he served as special assistant to the chief of naval operations and director of the CNO’s Commander’s Action Group, led studies on the Navy headquarters staff, and was an enlisted and officer submariner in the Navy.

Dan Patt is a senior fellow with the Hudson Institute’s Center for Defense Concepts and Technology, where he focuses on the role of information and innovation in national security. Patt also supports strategy at national security technology company STR and supports Thomas H. Lee Partners’ automation and technology investment practice. Previously, he co-founded and was CEO of Vecna Robotics and served as deputy director for the Defense Advanced Research Projects Agency’s (DARPA) Strategic Technology Office.

The post CJADC2 needs a hedge to succeed appeared first on DefenseScoop.

The Pentagon’s way of building and buying new gear was not designed for a military that gains an edge by recomposing its pieces like Lego blocks or mosaic tiles. Even recent reforms like the Middle Tier of Acquisition are still designed to create a single product that would later be integrated with other pieces of a kill chain or system-of-systems.

In a fight against China, U.S. commanders will need to synchronize changing combination of systems from multiple domains to succeed. Even if a new weapon or sensor brings extraordinary capabilities, it won’t matter if it is only carried by one type of aircraft or ship and cannot talk to anything else. A peer opponent like the People’s Liberation Army will exploit those limitations to keep the new system out of the fight.

More flexible integration among U.S. forces could give PLA commanders more kill chains to worry about, but is not one of DoD’s strong suits. Previous attempts like the Net-Ready Key Performance Parameter (KPP) and Joint All Domain Command and Control (JADC2) initiative have struggled to deliver tangible interoperability improvements and often become mired in bureaucracy or sidetracked by competing priorities.

However, there are promising signs of a solution. The Undersecretary of Defense for Acquisition and Sustainment’s establishment of a new office for Acquisition, Integration, and Interoperability, or AI2, suggests a recognition that instead of simply buying a new gadget to solve an operational problem, the DoD will increasingly need to buy the “glue” — or software, networks, and engineering — that combines existing systems in novel ways.

Longtime defense watchers know a new office or reorganization plan rarely solves a problem in DoD. For AI2 to succeed it will need to drive changes in how the U.S. military prioritizes interoperability and funds the glue for a more recomposable force. These capabilities often fall outside of a specific service’s lane.

Metrics will be essential to compel needed investments as well as define what constitutes “good enough.” Arguably one of the problems with JADC2 has been its overly-ambitious goals for interoperability, which lead to a top-down approach that has yielded few new effects chains.

Most defense professionals are well acquainted with Technology Readiness Levels, or TRLs, a metric system devised by NASA in the 1970s to assess technology maturity and later adopted by the DoD. TRLs work well for evaluating individual systems, but they do not help assess if a new capability is useful.

The introduction of an Integration Readiness Level (IRL) could address this shortcoming. Where TRLs assess a technology’s maturity in isolation, IRLs would evaluate a component’s readiness for integration into larger, dynamic systems-of-systems. An IRL would consider factors such as the exposure of interfaces and the component’s compatibility with Continuous Integration/Continuous Deployment pipelines used by DoD software factories to integrate new kill chains.

Like TRLs, IRLs could indicate greater degrees for interoperability using a scale from 1 to 9. An IRL of 1-3 suggests a nascent level of integration, where the system generally is used only by the original developer in a narrow set of systems of systems. IRL 4-6 represents a more mature stage, where integration or reuse by third parties is feasible with support from the original developer. The highest levels, IRL 7-9, depict a state where integration is more “plug-and-play,” and can be achieved without the direct involvement of the original developer.

Using the IRL metric could help manage risks associated with system integration and identify opportunities to enhance the modularity and agility of large-scale systems. For instance, the Electro-Optical Targeting System of the F-35 has a high TRL, but its IRL may be low because it cannot talk with other parts of the F-35. Conversely, a new F-35 plugin for the Android Tactical Assault Kit (ATAK) software might have a lower TRL but a high IRL, given its ease of integration across multiple contexts. The opportunity with ATAK would be highlighted by its IRL.

By adopting IRL as a measure, DoD could step away from a narrow focus on individual systems and lay the groundwork for a truly data-centric enterprise. In practice, a new metric will enable joint and OSD leaders to highlight shortfalls, direct appropriate investments by recalcitrant services, and define acceptable levels of interoperability to manage spending in a budget-constrained environment.  

IRL, in concert with other strategic initiatives such as the establishment of AI2 and the Chief Digital and AI Office, presents a pathway for the DoD to transform its approach towards technology development and acquisition, pivoting from a focus on standalone systems to a holistic view of integrated capabilities. Unless it makes this shift, the U.S. military could find itself like Russian troops in Ukraine — attempting to use a dwindling supply of silver bullets to stop a learning, adapting force that can incorporate almost anything into a kill chain.

Bryan Clark is a Senior Fellow at the Hudson Institute and Director of the Hudson Center for Defense Concepts and Technology. Dan Patt is a Senior Fellow at the Hudson Institute.

The post Establish integration metrics to prioritize ‘glue’ for all-domain operations appeared first on DefenseScoop.