Eurasia Group identifies the business strategies of leading Artificial Intelligence firms as a top geopolitical risk for 2026, arguing that under intense pressure to generate revenue and operating without sufficient guardrails, several major companies are likely to adopt models that threaten social and political stability. The analysis contends that these firms are following what it calls social media’s destructive playbook, but they are doing so more quickly and at greater scale, raising the stakes for governments, societies, and markets.
The report emphasizes that Eurasia Group remains optimistic about the transformative potential of Artificial Intelligence, highlighting that current frontier models can reason through complex problems, show their work, and are already embedded in coding, research, and knowledge workflows. It notes that the hyperscalers are offloading large chunks of software development to Artificial Intelligence, accelerating their own R&D cycles. In areas such as biotech and materials science, Artificial Intelligence is opening new research pathways, although commercial breakthroughs remain mostly ahead. The authors add that hundreds of millions of people now use chatbots daily for tasks ranging from drafting emails to debugging code and learning new skills, illustrating that adoption by individuals is real and still in its early stages.
At the same time, the report argues that Artificial Intelligence cannot meet investors’ short-term expectations. It states that even after hundreds of billions of dollars of investment, the most advanced models still hallucinate and display jagged capabilities, excelling at some tasks while being unreliable at others in ways that are often unpredictable, which makes them difficult to deploy in high-stakes contexts where mistakes are costly. Business uptake is described as uneven, with only about 10% of US firms using Artificial Intelligence to produce goods and services, according to the Census Bureau. While many companies report significant productivity gains, surveys suggest that most have not yet seen meaningful bottom-line impact, and the authors stress that genuine productivity improvements will depend on widespread diffusion of the technology across the economy, a process that will take time, even as markets have priced in revolution rather than evolution.
