WASHINGTON & SAN FRANCISCO — In the high-security boardrooms of Silicon Valley, the mood has shifted from unbridled technological optimism to a grim, defensive pragmatism. According to a recent, deeply unsettling report by Axios, top executives and safety researchers at leading artificial intelligence firms—including industry heavyweights Anthropic and OpenAI—are privately conducting high-stakes "war games" and…