America Wants to Control Every AI Chip on Earth. The Rest of the World Heard Something Else.

The Trump administration is drafting rules that would let the US government approve or block nearly every major AI chip sale on the planet. Washington says it's protecting national security. Beijing says it's weaponizing technology to contain China's rise.

They're describing the same policy document.

What the Rules Actually Say

On March 5, Bloomberg reported that the US Commerce Department is building a worldwide licensing system for advanced AI chips. Here's how it breaks down:

Shipments under 1,000 Nvidia GB300 GPUs get a simplified review. Bigger installations need pre-authorization — with conditions that could include on-site inspections by US officials. Clusters above 200,000 chips (the kind run by Amazon, Microsoft, and OpenAI) would require direct government-to-government agreements.

That last part is the one that's rattling capitals from London to Beijing. It means a UK company wanting to build a major AI data center would need Washington's blessing. So would France, Japan, and South Korea.

This isn't a ban. It's a lever. And the question of who gets to pull it is where the world splits.

The American Frame: Security First

Inside the US, the debate isn't about whether to control AI chips — it's about how tightly.

The new rules replace Biden's AI Diffusion Rule, which Trump rescinded. But the replacement turns out to be stricter, not looser. Where Biden sorted countries into tiers, Trump's approach gives the administration case-by-case authority over every large sale worldwide.

At a House Foreign Affairs hearing in January, former Deputy National Security Adviser Matt Pottinger called even the limited H200 chip sales to China a "wrong path" that would "supercharge" Beijing's military. Lawmakers from both parties pushed back on the administration's approval of Nvidia sales to China, with Ranking Member Gregory Meeks saying the US was "ceding our advantage" and "actively undermining our national security."

The framing is consistent: AI chips are strategic weapons. Controlling them is self-defense.

The Asia-Pacific Frame: Containment by Another Name

From Beijing's side, the picture is inverted.

When Commerce Secretary Howard Lutnick explained the chip strategy by saying the US wanted Chinese developers to "get addicted to the American technology stack," Chinese commentators heard echoes of the Opium Wars — a loaded historical reference in a country where that period represents imperial exploitation at its worst.

China's response was swift and practical. The country's top internet regulator ordered tech companies not to acquire Nvidia's H20 chips, citing "serious security issues." Beijing issued guidance telling firms to double down on domestic alternatives instead.

The message: we won't be dependent on a supply that can be cut at someone else's whim.

And China isn't just declining to buy. It's building its own export control regime. Beijing has restricted rare earth exports — materials essential for chip manufacturing — creating a mirror-image pressure point. As Reuters reported in late February, "Beijing's export controls come of age" as a tool of its own.

The Part That Gets Lost

Here's what neither frame captures well: this policy doesn't just affect the US and China.

A UK cloud provider wanting 200,000 GPUs now needs to negotiate with Washington. A French AI startup scaling up its compute needs US approval. Japan and South Korea — both close allies — would operate under the same licensing requirements.

Some American allies are uncomfortable. When the Biden-era diffusion rule first dropped, countries from Europe to Southeast Asia pushed back, saying they were "being shut out from crucial technology needed to develop artificial intelligence."

Trump's version addresses some of those complaints by creating faster tracks for smaller orders. But the big projects — the ones that actually define who leads in AI — all run through Washington.

The Numbers Behind the Gap

The Albis Perception Gap Index scored this story at 6.80 — Different Realities tier.

The dimensional breakdown tells the story:

• Factual divergence: 4.0 — Both sides agree on what the rules say.
• Causal attribution: 6.5 — Washington says "security." Beijing says "economic warfare."
• Framing: 7.0 — American outlets lead with protecting AI leadership. Asian outlets lead with technological sovereignty.
• Actor portrayal: 7.5 — Trump is either a pragmatic protector of American interests or an aggressor using chips as leverage.
• Cui bono: 8.0 — The widest gap. American coverage serves the security establishment's interests. Asia-Pacific coverage serves the sovereignty narrative.

The US-Asia Pacific region pair divergence sits at 6.8 — meaning two audiences reading about the same policy walk away with nearly opposite understandings of what it means and who it serves.

Why It Matters

AI chips aren't smartphones or sneakers. They're the raw material for the systems that will run military planning, financial markets, drug discovery, and infrastructure management for the next decade.

Whoever controls chip access shapes what gets built, where it gets built, and by whom.

The US position is that this power should sit in Washington. China's position is that no single country should hold that switch. Both positions have internal logic. Both serve specific interests.

What's missing from most coverage — in any region — is what this means for the 190 countries that don't make advanced AI chips. They're not at the table. But they'll live with whatever gets decided.

---

This story was scored by the Albis Perception Gap Index — measuring how differently the world frames the same events. See today's most divided stories →