From Gotham City to Tehran via the Silent Holy City: Artificial Intelligence at War

Antônio Dasmortes, 15/3/2026

The USraeli attack launched on February 28 against Iran – named Epic Fury on the US side and Silent Holy City on the Israeli side – can be considered the second major attack in history planned and partly executed by artificial intelligence, the first being the kidnapping of the Venezuelan president and his wife on January 3 in Caracas. In this operation, the Gotham and AIP (Artificial Intelligence Platform) systems from the company Palantir Technologies played a central role, acting as the indispensable nervous system that allowed the AI “Claude” to analyze massive data and contribute to the planning of the strikes.

The table below summarizes the distinct but complementary roles of Gotham and AIP in this operation.

System	Primary Role	Key Function in the Attack
Gotham	Data Integration Platform	Act as an “intelligence hub” to fuse and analyze disparate raw data (satellites, communications, etc.) on a single platform.
AIP	Secure Operating System	Serve as a single, authorized access gateway, allowing the AI “Claude” to operate and securely analyze classified data within a sensitive military environment.

The Architecture of the Attack: The “Brain” and the “Nervous System”

To better understand their interaction, one can compare the entire system to a human body:

• Palantir (Gotham and AIP) constituted the nervous system. It collected and integrated massive and disparate data flows, estimated at 2.3 petabytes (2,300 terabytes, approximately 120 million documents), from satellites, communication intercepts, human sources, and social networks.
• Claude (Anthropic’s AI) acted as the brain. Connected to Palantir’s nervous system via the AIP, Claude analyzed this data, executed tens of thousands of attack scenario simulations, and proposed optimized action plans to human commanders.

The specific contributions of Gotham and AIP

Their roles were decisive in several key phases of Operation “Epic Fury”/”Silent Holy City”:

• Lightning-Fast Intelligence Processing: Gotham enabled the cleaning, modeling, and extraction of crucial information from raw data in just 90 minutes, a task that would have taken a human team 100 days.
• Precision Targeting: By integrating location data, lifestyle habits, and even bodyguard shift schedules, the system was able to identify precise “windows of opportunity,” such as a 3-minute gap caused by an overlap between a radar blind spot and a guard change. This helped reduce the location error margin for high-value targets to less than 500 meters.
• Anomaly Detection: The AI, via the Palantir platform, was able to analyze minute details, like a 1.2-second stop of a vehicle convoy at an intersection – a signal a human analyst might have missed – which was crucial for locating targets.

In summary, Gotham served as the foundation by transforming an ocean of raw data into actionable intelligence, while AIP provided the secure environment and necessary protocols for the AI Claude to analyze this classified information and actively contribute to the strategic planning of the 11-minute and 23-second lightning strike against Iran.

Below are 3 infographics summarizing the characteristics of the new USraeli war machine.