President Trump and Xi Jinping will address a new global challenge at their Beijing summit: the rise of agentic AI systems capable of devastating cyberattacks. Unlike Cold War-era rivals, the U.S. and China now face a shared foe in asymmetric AI tools that could empower small actors to cripple critical infrastructure. Anthropic and OpenAI have already limited access to their most powerful models, fearing misuse, but experts warn leaks are inevitable.

The Nixon-Mao summit of 1972 established a U.S.-China alliance against the Soviet Union, but today’s threats demand unprecedented cooperation. Craig Mundie, a former Microsoft strategist, argues that both nations must collaborate with tech giants like Google and Alibaba to regulate AI development. Failure to act risks a "mutually assured destruction" scenario, where either side’s cyber capabilities could destabilize the other.

The AI arms race mirrors nuclear proliferation: companies now hold power to create tools for good or harm. With Starlink-enabled hackers and AI-driven attacks looming, the U.S. and China must establish guardrails before it’s too late. Trump’s reported plan to preemptively oversee AI models signals growing urgency, but global coordination remains elusive.

Why this matters: Unchecked AI could render traditional defense strategies obsolete. The I7 coalition — including Meta and DeepSeek — must balance innovation with safeguards. As Dov Seidman noted, the world is now "fused," meaning no nation can escape interconnected risks alone.