The Problem Nobody Is Framing Correctly

The United States is the global AI leader right now. That sentence has an expiration date attached to it, and Washington is running out the clock.

The mainstream coverage treats this as a simple debate: Nvidia says export controls hurt American business, Anthropic says tighten the screws, and somewhere in the middle Congress nods along. The debate obscures the actual stakes.

This isn't a trade dispute. It's a race for the infrastructure of the next era of warfare, economic power, and geopolitical leverage.

What the Testimony Actually Said

On December 2, 2025, Gregory C. Allen — Senior Advisor at the Wadhwani AI Center at the Center for Strategic and International Studies — testified before the Senate Foreign Relations Subcommittee on East Asia, the Pacific, and International Cybersecurity Policy.

His opening comparison was clarifying. Both the Biden and Trump administrations have called the U.S.-China AI race a new space race. Allen said that analogy undersells it.

The Apollo program cost the U.S. government $28 billion between 1960 and 1973 — roughly $326 billion in today's dollars, according to Allen's testimony. In 2026 alone, just five American companies — Meta, Alphabet, Microsoft, Amazon, and Oracle — are projected to spend more than $450 billion in AI-specific capital expenditures. OpenAI, Anthropic, and Elon Musk's xAI will add hundreds of billions more on top of that.

This race isn't primarily being run by governments. It's being run by corporations.

The Huawei Problem

Huawei is the piece most coverage glosses over.

Nvidia CEO Jensen Huang testified before the House Foreign Affairs Committee that U.S. chip export restrictions are inadvertently fueling Huawei's rise. By locking Chinese companies out of American chips, Washington is pushing them straight into Huawei's arms — and giving Huawei a captive domestic market of 1.4 billion people and thousands of Chinese enterprises to build against, according to reporting by Reuters cited via opentools.ai.

Huang's argument is straightforward: sanctions that seem tough on paper may actually be accelerating the competitor the U.S. is trying to slow down.

National Review has called for aggressive action specifically targeting Huawei's AI capabilities, arguing that U.S. advantages will evaporate without a coherent counter-strategy.

Badly designed export controls can harm American companies AND fail to stop Huawei. This is the policy outcome Washington is currently producing.

The Anthropic vs. Nvidia Fight

Anthropic and Nvidia are publicly at odds on this.

Anthropic has pushed for stricter export controls, raising alarms about AI chip smuggling and the danger of advanced American hardware reaching adversaries. Nvidia fired back, arguing that Anthropic's position undermines global competitive dynamics and shifts focus away from genuine innovation, according to opentools.ai.

Anthropic is an AI software company that benefits when hardware is harder for competitors to obtain. Nvidia sells hardware globally and loses revenue when exports are restricted. Both companies have financial incentives in this debate.

Congress should be smart enough to see that. Whether it is remains an open question.

What the Media Is Getting Wrong

Left-leaning outlets have largely framed this as a story about corporate lobbying — Nvidia protecting its bottom line. Right-leaning outlets have framed it as Biden-era regulatory overreach that Trump should dismantle.

The U.S. has NO coherent national AI strategy that integrates export controls, domestic investment, allied chip manufacturing, and Huawei-specific countermeasures into a single doctrine. Instead, there is a patchwork of reactive policies dressed up as a plan.

Allen's CSIS testimony is the most detailed attempt to lay out that bigger picture, and it received a fraction of the media attention given to the Nvidia-Anthropic spat.

The Numbers

$450 billion. That's American private sector AI investment in a single year. China's government is pouring state resources into the same race. Huawei's AI chip division is moving fast enough that Nvidia's own CEO is treating it as a serious threat.

The U.S. lead in AI is real. It is NOT permanent. The gap between leading and losing in this technology is measured in product cycles, not decades.

What Needs to Happen

First, export controls need to be redesigned with specific outcomes in mind — not reflexive bans that push customers toward Huawei. Huang is right about that.

Second, Congress needs to hear from people like Gregory Allen more and from corporate lobbyists less. The CSIS testimony laid out a framework for thinking about AI diffusion, economic security, and national security together.

Third, Huawei needs to be treated as the primary target. Not Chinese tech generically. Huawei specifically — because it is the instrument Beijing is using to wire the world's AI infrastructure to Chinese standards.

America is winning the AI race right now. Winning right now and winning in 2030 are two different things. Washington should stop acting like they're the same.