Calls to “slow down” American artificial intelligence are often packaged as prudence. In practice, they function as a unilateral pause in a race the Chinese Communist Party has no intention of pausing. Beijing treats frontier AI as a pillar of national power. Washington treating it as a luxury we can throttle is not caution. It is strategic self-harm. Anyone in the United States who treats a domestic slowdown as the default safety plan is reciting a talking point that serves the CCP far better than it serves American workers, labs, or allies.
The Heritage Foundation has documented the pattern. Data centers and compute are the industrial base of the AI era. CCP-linked actors have amplified opposition to U.S. data-center construction even as Beijing subsidizes its own buildout. Slowing American infrastructure while China races is not a safety policy. It is a transfer of time to an adversary.¹ Heritage analysts have also warned that export-control loopholes and chip smuggling exist precisely because Beijing treats American compute as a prize to steal, not a system to match on open terms.² The correct response is to keep the U.S. lead and deny China the stack—not to idle American labs so Beijing can close a months-long gap.
That gap is real and shrinking. Testimony assembled at the American Enterprise Institute stresses that leadership will not be decided by a single model score. It will be decided by compute, talent, electricity, and whether companies actually deploy AI across the economy.³ Hal Brands has argued that a pause among democracies is meaningful only if China also restrains itself—an outcome Beijing’s own media has dismissed as a “containment” plot from a Cold War playbook.⁴ A one-sided American slowdown does not produce a safer world. It produces a world in which the next generation of tools is trained under CCP rules.
The Hoover Institution’s work on China’s technology playbook makes the same point from the other direction. Xi Jinping treats technology as “the cornerstone of a strong nation,” and the playbook—investment, indigenization, insulation, integration, internationalization—is already in motion.⁵ If the United States answers that campaign with moratoria, permitting gridlock, and moralizing about “too much progress,” it is not regulating risk. It is volunteering to lose the inputs that decide who writes the next industrial standard.
Nowhere is the cost of that choice more concrete than in the Pacific Northwest. Greater Seattle supports nearly 188,000 technology jobs and a regional tech output measured in the mid-hundreds of billions of dollars. Average earnings in the sector dwarf national norms. The region is home to AWS and Microsoft Azure, two of the world’s core cloud and AI platforms, plus the Allen Institute for AI and a deep University of Washington pipeline. It holds one of the country’s largest concentrations of AI specialists—on the order of tens of thousands of workers—and hundreds of AI firms and startups.⁶ Seattle accounts for a large share of the nation’s AI engineering talent. That is not a coastal abstraction. It is payroll, property tax, and supplier networks from Redmond to the Columbia Basin.
The buildout is already visible beyond the waterfront campuses. Central Washington’s data-center corridor has pulled in electricians, apprentices, and local tax base because hyperscalers need power, land, and skilled trades.⁷ Washington’s recent GDP spikes have been driven in large part by information and AI-related capital spending from Amazon and Microsoft.⁸ Punitive slowdowns, moratoria on data centers, and a national mood that treats compute as a pollutant would not “protect” the region. They would freeze the one sector still generating outsized output in a state whose economy is tightly coupled to a handful of technology firms.
Safety, export controls against the PLA, and honest debate about energy and land use are compatible with speed. A campaign to stall American frontier development is not. The CCP does not need Americans to love Beijing. It only needs enough of them to treat U.S. acceleration as the problem. The Pacific Northwest’s prosperity, and the country’s lead, depend on refusing that script.
Notes
- Wilson, C.-M. (2026, September 28). Is China working against U.S. data centers? The Heritage Foundation. https://www.heritage.org/technology/commentary/china-working-against-us-data-centers
- Dorsey, A. (2026, May 26). The AI arms race has a tracking problem. The Heritage Foundation. https://www.heritage.org/china/commentary/the-ai-arms-race-has-tracking-problem
- Miller, C. (2025, December 2). China’s challenge to American AI leadership. American Enterprise Institute. https://www.aei.org/research-products/testimony/chinas-challenge-to-american-ai-leadership/
- Brands, H. (2026, September 16). The US can slow down AI and stay ahead of China. American Enterprise Institute. https://www.aei.org/commentary/the-us-can-slow-down-ai-and-stay-ahead-of-china/
- Economy, E. (2026). A close look at China’s tech playbook. Hoover Institution. https://www.hoover.org/research/close-look-chinas-tech-playbook
- Greater Seattle Partners. (2025). Greater Seattle technology & AI industry report 2025. https://greater-seattle.com/greater-seattle-technology-ai-industry-report-2025/
- An A.I. boom makes electricians flock to Central Washington. (2024, December 25). The New York Times.
- Regional reporting on Washington GDP and information-sector growth; see also Axios Seattle coverage of Seattle’s AI workforce concentration (2026).
