The IBM Institute for Business Value has published a new report examining defense in the era of artificial intelligence.
IBM's findings indicate that as AI becomes more integrated into defense, leaders identify cybersecurity, resilient networks, sovereign computing, and domestic talent development as critical priorities for establishing AI as national infrastructure.
Talent remains a significant challenge, with many organizations relying on vendors for initial projects.
Leaders emphasize the need for personnel who understand both mission objectives and technology to ensure reliability, resilience, and long-term scalability.
Insights from NATO and Five Eyes leaders highlight that AI is transforming defense and intelligence operations.
NATO is an intergovernmental military alliance of 32 member states in Europe and North America, established in 1949.
The Five Eyes is an intelligence-sharing alliance among Australia, Canada, New Zealand, the United Kingdom, and the United States, based on mutual trust and formalized through the UKUSA Agreement, which originated from World War II-era intelligence collaborations.
Talent remains a significant challenge, with many organizations relying on vendors for initial projects.
The IBM Institute for Business Value survey of 600 chief information and technology officers in NATO and Five Eyes countries found that leaders expect AI to impact areas such as intelligence analysis, cyber defense, logistics resilience, and multi-domain coordination.
However, only 10% of respondents report high success with AI today. Most defense organizations are unable to deploy AI within systems that drive daily decisions.
While AI investment is increasing, tangible results are lagging.
By 2028, AI budgets are expected to shift from traditional and data-focused AI to generative and agentic AI, supporting tasks like automated reporting, threat detection, and mission planning.
Spending on agentic AI is projected to more than double, reflecting its value in coordinating workflows with minimal input.
According to a CIO at a combat arms unit quoted in the report, modernizing legacy systems is crucial, as many current platforms restrict effective, large-scale AI use.
Only about half of the available data is suitable for AI, and just one-third is properly prepared.
Even organizations with advanced data readiness face challenges in achieving operational impact.
Less than one-third of core systems are AI-ready, as most command-and-control, logistics, and intelligence platforms were not designed to process machine-generated outputs.
In contrast, IT operations, cybersecurity monitoring, maintenance, procurement, and supply chain forecasting are progressing more consistently, due to structured data and ongoing modernization.