Lockheed Martin is creating a digital testing ground to evaluate military AI algorithms through the AI Fight Club initiative, which is open to small businesses.
Lockheed Martin has announced the launch of the “AI Fight Club,” a simulation environment designed to test artificial intelligence algorithms for warfare. The project aims to accelerate the integration of AI into military operations—air, land, sea, and space—by pitting systems against complex scenarios validated by the US government. It also aims to open the door to small businesses, which are often at the forefront of innovation but excluded from large-scale testing due to a lack of infrastructure. The stated objective is clear: to identify viable solutions and integrate them into Pentagon programs. The program, scheduled to be operational by the end of 2025, is a strategic lever in the global technology race, particularly against China.
A private initiative to meet a public need
The AI Fight Club was designed to fill an identified gap in the military AI technology validation chain. Lockheed Martin, an American defense giant with annual revenues of more than $65 billion, initially conceived this platform to test its own internal algorithms. The company quickly realized that there was also potential in technologies developed outside the traditional military industry circles.
The strategic objective is twofold: to accelerate the evaluation of innovations and to provide a structured interface with the Pentagon. The Department of Defense (DoD) requires very specific validation standards, which are often beyond the reach of start-ups or SMEs. This lack of access slows down the flow of relevant technologies to military decision-makers. By providing a centralized infrastructure that complies with DoD standards, Lockheed is helping to bridge this gap.
This environment benefits hundreds of players who do not have servers, test benches, or simulators that comply with military standards. The cost of such an ecosystem often exceeds $2 million in technical infrastructure alone. Lockheed is therefore creating a pivotal interface between private innovation and public decision-making.
A technical response to geopolitical urgency
The launch of this program comes at a tense time internationally, with Chinese investment in algorithmic warfare raising serious concerns. Beijing is spending an increasing share of its military budget—estimated at more than €210 billion—on autonomous and digital technologies, including military artificial intelligence, collaborative drones, and automated command systems.
For Washington, the technological gap represents a direct strategic risk. The Pentagon has already launched initiatives such as Joint All-Domain Command and Control (JADC2) and the Replicator program, designed to mass-produce autonomous systems. However, algorithm evaluation remains a bottleneck due to the lack of a unified testing environment.
The model proposed by Lockheed Martin aims to respond to this urgent need by creating a realistic, large-scale simulation space that integrates hybrid threats, spatial data, and multi-theater conflict situations.
This type of digital test bed would, for example, make it possible to pit two AIs piloting air formations against each other in a simulated air defense theater, or to test a cyber defense algorithm embedded in a satellite against a simulated attack. These scenarios are modeled using US government simulation tools (such as AFSIM), which guarantees the compatibility and usability of the results by defense agencies.
A competitive and confidential space
Unlike a simple call for projects, the AI Fight Club introduces a direct competitive dimension. The algorithms are not only tested, they are compared in simulated battles. This confrontational approach makes it possible to move away from a declarative model (“our AI works”) to a demonstrative model: results are generated, compared, and interpreted in extreme situations.
Each session is supervised by DoD representatives who observe performance, evaluate responses to unforeseen events, and identify points of failure. This process strengthens the link between innovation and military doctrine. The aim is to put teams of engineers back in direct contact with the realities of modern combat.
Lockheed Martin also guarantees the protection of intellectual property for participants. The algorithms tested are not copied or analyzed outside the session, in accordance with a strict contractual commitment. This is an essential condition for convincing technology startups, which are often wary of large defense companies.
The best teams will be able to publish their work, access funding, or join Lockheed Martin’s circle of approved suppliers. The company points out that 60% of its revenue is redistributed to its suppliers, representing approximately $42 billion per year. The AI Fight Club is thus becoming a gateway to US institutional markets.
Consolidation of the defense AI sector in the United States
This program also reinforces the current trend toward structuring the defense AI sector in the United States. Faced with a fragmented offering, often made up of microstructures, there is a need for industrial unifiers. Through this program, Lockheed Martin is positioning itself as a technology aggregation platform, similar to an algorithmic solutions integrator.
This stance is not neutral. By centralizing evaluation, the company influences the Pentagon’s technological choices. Systems validated through the AI Fight Club are more likely to be integrated into the group’s industrial projects, such as Next Generation Air Dominance (NGAD) or Joint Strike Fighter.
In the medium term, this centralization could reinforce the dependence of small structures on large groups in exchange for access to public contracts. In short, Lockheed Martin is transforming military AI into a structured, standardized, competitive market filtered by actual performance, which excludes purely marketing-driven solutions.
Operational outlook and short-term deadlines
The first tests of the AI Fight Club are scheduled for the last quarter of 2025. The current phase, spread over three months, will allow the structure to be adjusted according to the level of participation. The environment is based on government-approved simulation tools, some of which, such as STK, AFSIM, and JSAF, are used in US military planning.
Lockheed anticipates strong demand, particularly from sectors such as space detection, naval air defense, and military cybersecurity. According to internal projections, several hundred organizations could participate in the first sessions.
This initiative synergizes with other federal programs, such as TRUST AI (DARPA’s algorithmic reliability program) and OFFSET (autonomous swarm systems). The AI Fight Club could become a key link in the US doctrine of algorithmic integration, particularly in contested theaters where system autonomy is crucial.
War Wings Daily is an independant magazine.