In times of war, the market's direction is dictated more by geopolitical events and military strategy than by traditional financial metrics. Understanding a conflict's potential duration (e.g., a swift operation vs. a prolonged war) becomes the most critical forecasting tool for investors and risk managers.

Current military assessments focus on inputs like '6,000 targets struck,' creating a false sense of progress. This echoes the Vietnam War's body count metric, which measures activity but fails to assess actual strategic effects like achieving free navigation or eroding the enemy's power.

The strategy's focus on AI simulation acknowledges a key risk: AI systems can develop winning tactics by exploiting unrealistic aspects of a simulation. If simulation physics or capabilities don't perfectly match reality, these AI-derived strategies could fail catastrophically when deployed.

Strategic military planning, which looks decades into the future, is still based on a 2% inflation target. This is a critical flaw, as even slightly higher sustained inflation will drastically cut the real budget, severely limiting the military's ability to procure equipment and maintain readiness.

Standard AI evaluations use well-defined scenarios. Military operations are inherently dynamic and unpredictable. National security AI therefore requires a new evaluation paradigm focused on specific, tailored use cases and operational reliability under unforeseen circumstances.

When questioned on progress, the Pentagon defaults to 'input metrics' like money invested, avoiding 'output metrics' like the number of missiles produced by a specific date. This focus on process over results allows critical projects to languish for years.

Debates over military budgets are often implicitly debates about the discount rate of combat power. In an era of rapid AI-driven change, power delivered in 10 years is heavily discounted. This framework favors immediate software and autonomy investments over long-term hardware programs.

The future of military strategy involves merging planning software with wargaming simulations. AI agents will generate a course of action, play it out in a physics-based simulation against an AI adversary, analyze the results, and automatically improve the plan, creating "superhuman" strategies.

Smack Technologies argues that general-purpose LLMs fail in military strategy because they rely on historical labeled data. For novel, high-stakes conflicts, a different approach like deep reinforcement learning is required, training models within physics-grounded simulations of potential future battlefields.