The first warning came not from a missile launch or a naval clash, but from a procurement delay.
In February 2027, executives inside NVIDIA received notice that shipments of dysprosium and terbium magnets for next-generation AI server cooling systems would be delayed indefinitely. The minerals themselves were not rare in the earth’s crust. What was rare was the ability to refine them at industrial scale.
China controlled nearly all of that.
For years, American strategists had assumed software would determine the future balance of power. Silicon Valley produced frontier AI models that could autonomously write software, design molecular compounds, coordinate drone swarms, and optimize entire logistics systems. American firms dominated foundation models, hyperscale cloud infrastructure, and semiconductor architecture. OpenAI, Anthropic, and Google had become as strategically important as oil companies once were.
But China had spent twenty years preparing for a different battlefield.
Beijing’s leadership understood that intelligence required matter. AI was not magic. Every model depended on power transformers, gallium compounds, high-bandwidth memory, cooling systems, optical interconnects, and rare earth magnets. China had methodically secured mining concessions across Africa, Southeast Asia, and Latin America while building refining monopolies at home. By 2026, even American analysts admitted the uncomfortable truth: the United States still depended heavily on Chinese processing for critical minerals essential to advanced electronics and AI hardware.
Washington responded with tariffs, export controls, and emergency mineral stockpiles. Congress funded “Project Vault,” a strategic reserve intended to protect American manufacturers from Chinese supply disruptions.
But privately, senior officials inside the Pentagon understood the mathematics.
The United States could slow China’s access to advanced AI chips.
China could slow the entire world’s access to the materials required to build them.
The rivalry evolved into something stranger than a Cold War. It was not ideological in the twentieth-century sense. American venture capitalists still invested indirectly through Singapore and Dubai into Chinese supply-chain firms. Chinese sovereign funds quietly held stakes in Western semiconductor suppliers through offshore structures. Despite public hostility, both superpowers remained economically fused at the deepest industrial layers.
By 2028, analysts started using a new term.
The AI Major.
The phrase emerged from an obscure RAND workshop before spreading through think tanks and financial markets. A “major” was no longer simply a great power. It was a civilization-scale AI bloc capable of controlling the entire stack:
- frontier AI models,
- semiconductor design,
- energy generation,
- mineral refinement,
- satellite infrastructure,
- robotics manufacturing,
- and planetary-scale data collection.
Only two entities qualified.
America possessed the software ecosystem. China possessed the material ecosystem.
Separately, each had vulnerabilities.
Together, they could dominate the century.
That possibility terrified everyone else.
In Brussels, European officials realized the continent was becoming strategically irrelevant despite its regulatory influence. India accelerated its own semiconductor and rare-earth initiatives but remained dependent on foreign lithography and capital. Japan retained advanced materials expertise yet lacked sufficient domestic resources. Smaller nations increasingly faced a binary choice: attach themselves to one AI major or risk exclusion from the emerging machine economy.
The shift became visible in global trade patterns.
Shipping routes through the Strait of Malacca transformed into the most heavily monitored commercial corridor on earth. Autonomous cargo fleets escorted by AI-guided naval drones carried minerals westward and GPUs eastward. Insurance markets began pricing “compute sovereignty risk” into sovereign debt ratings.
Then came the Jakarta Incident.
An Indonesian nickel refinery operated jointly by Chinese state firms and American cloud companies suffered a catastrophic cyberattack. Automated safety systems failed simultaneously. Molten slag exploded through processing lines, killing eighty-three workers. Within hours, both Washington and Beijing accused each other of sabotage.
Financial markets panicked.
For forty-eight hours, military satellites tracked unusual submarine movements in the South China Sea while AI systems inside both countries generated escalation simulations faster than human diplomats could read them.
And then something unprecedented happened.
Instead of mobilizing for war, both governments opened a secure emergency AI hotline established only months earlier after private lobbying from tech executives terrified of accidental conflict.
The investigation eventually revealed the truth: neither side had launched the attack. A non-state ransomware collective had exploited vulnerabilities in industrial control software generated partly by open-source AI coding agents.
The realization shook both capitals.
The greatest threat was no longer merely each other.
It was uncontrolled machine proliferation itself.
Quiet negotiations followed.
Officially, the United States and China remained adversaries. Politicians still delivered speeches about technological sovereignty and national security. Export restrictions continued. Naval patrols intensified. Rare-earth disputes worsened.
But beneath the surface, a different architecture emerged.
American AI firms received limited access to Chinese mineral-processing consortia through intermediaries in the Gulf states. Chinese manufacturers gained indirect access to older generations of American AI accelerators. Joint standards bodies quietly established protocols for autonomous weapons containment, synthetic media authentication, and AI-driven cyber defense.
The public never saw the full agreements.
They only saw the results.
By 2032, most of the world’s cloud computation ran on systems whose software lineage traced back to California while their physical infrastructure depended on Chinese-controlled material networks. The two rivals had not merged politically. They had become structurally inseparable.
The AI Major had formed.
And nations outside it slowly realized the implications.
Countries that controlled neither advanced AI nor strategic minerals became economically dependent provinces of the new machine age. Their labor markets hollowed out first. Then their currencies weakened as AI-managed supply chains bypassed them entirely. Some attempted digital protectionism. Others nationalized data. Few succeeded.
The old map of developed and developing nations disappeared.
Now there were majors and non-majors.
Inside windowless strategy rooms in Washington and Beijing, officials studied the same terrifying projection generated by competing supercomputers:
Whoever controlled both cognition and matter would not merely lead the global economy.
They would define reality itself.
All names of people and organizations appearing in this story are pseudonyms
Comments