US-China 2026: The AI Arms Race That Is Rewriting Geopolitics
By Sanna the Weaver • Fri Mar 20 2026 • Geopolitics
In 2026, the US-China rivalry has converged on a single technology: artificial intelligence. The competition is no longer simply about economic dominance or military capability — it is about which nation's AI systems will power the world's infrastructure, set the norms for AI governance, and ultimately determine which model of society is most capable of adapting to the age of machine intelligence. America Builds the Models; China Wins the Market? American AI labs — OpenAI, Anthropic, Google DeepMind — are producing the most capable frontier models. GPT-5.4, Anthropic's Claude Mythos 5, and Google's Gemini 3.1 Pro are operating at or above human expert level on economically valuable tasks. But a growing share of the world's energy, mobility, and industrial systems are being built on Chinese foundations. Huawei's 950PR inference chip, unveiled in January 2026, has seen massive orders from ByteDance and Alibaba — domestically, China is building an AI ecosystem that doesn't need Nvidia. The strategic consequence is clear: the US may build the best AI, but if China can deploy AI at scale across the Global South — through its Belt and Road digital infrastructure, subsidized software, and willingness to operate in regulatory environments that Western companies won't enter — it may win the markets that will define 21st-century economic power. Taiwan and the Silicon Chokepoint Taiwan's TSMC still manufactures approximately 90% of the world's most advanced semiconductors. Every advanced AI chip — American or Chinese — depends on equipment that ultimately traces back to TSMC's fabs in Hsinchu and Tainan. This concentration of critical manufacturing in a geopolitically contested island remains the single greatest structural vulnerability in the global AI race. Beijing's military pressure on Taiwan has intensified in 2026 — not to the point of invasion, but in persistent gray-zone operations designed to demonstrate that Taiwan's security is contingent on Beijing's restraint. "We are in a race where the finish line keeps moving. But the consequences of falling behind are permanent." — US National Security Advisor, January 2026 The Rules Nobody Is Writing Perhaps most dangerous is the absence of any bilateral framework to manage AI competition. The US-China AI safety dialogue, launched tentatively in 2024, has collapsed. Neither country has agreed to any binding constraints on AI development for military applications. The use of AI-enabled cyberattacks — attributed by US intelligence to China in the 2025 period — continues, with each side accusing the other of crossing lines that have never been formally defined. The AI arms race has no arms control regime. That absence may be the defining risk of the decade.