AI arms-race tensions rise as the US-China split tightens—and cash-rich buyers move fast

MIT economist Simon Johnson warns that the US is drifting toward an anti-China, anti-science posture that could distort how AI is governed and deployed, arguing that institutions—not slogans—determine long-run prosperity. In the same discussion, he highlights the systemic risks of AI when policy becomes politicized, especially amid escalating strategic competition with China. The article frames Johnson’s concerns through his Nobel-linked work on how institutional design shapes national outcomes, and it ties the debate to the broader AI risk agenda. The timing matters: the comments land as governments and firms are simultaneously racing to scale AI capabilities and harden national security narratives. Strategically, the cluster shows how AI is becoming a geopolitical instrument rather than a neutral technology. The US-China relationship is the central fault line, with policy choices in Washington increasingly influencing corporate behavior, investment flows, and compliance expectations. South Korea’s “cash-rich winners” using a US buying spree signals that allied industrial policy is aligning with US strategic priorities, while also trying to reduce exposure to potential tariff shocks under a possible Trump return. Taiwan’s vice minister urging AI use to improve bilingual learning outcomes adds a parallel track: AI is being normalized in public services, but under a governance framework that can quickly become a security and sovereignty question. Overall, the winners are firms and states that can finance compute, manage regulation, and translate AI into measurable productivity faster than rivals. Market and economic implications are already visible across real assets and tech supply chains. South Korea’s large US investment push points to continued demand for US data-center capacity, cloud infrastructure, semiconductors, and enterprise software—supportive for AI capex cycles and related equities. In Pennsylvania, 96 families sold 1,700 acres for $586 million to a Blackstone-backed data center developer, underscoring how the AI infrastructure race is bidding up land values and accelerating local development. That kind of land and power-intensive buildout tends to lift demand for construction services, grid upgrades, fiber connectivity, and water/energy management, while also increasing political scrutiny over permitting and community impact. Currency and rates are not directly cited, but the direction is clear: AI-driven capex is pulling capital toward US infrastructure and away from slower, less bankable projects. What to watch next is whether AI governance hardens into a durable institutional framework or remains vulnerable to political swings. For the US-China dimension, monitor signals on export controls, AI safety regulation, and any policy language that frames science as a security threat rather than a public-good investment. For South Korea, track the scale and timing of US acquisitions, plus corporate disclosures about tariff contingency planning and supply-chain localization. For Taiwan, watch how bilingual education AI pilots are funded, evaluated, and protected against data governance risks that could spill into broader cross-strait tensions. In Pennsylvania and similar markets, the trigger points are permitting timelines, power interconnection queues, and community backlash that could slow builds and shift costs into the next earnings cycle.

Geopolitical Implications

• 01

  Institutional credibility in AI governance is becoming a strategic differentiator; politicized “anti-science” narratives could undermine long-term innovation capacity.

• 02

  Allied alignment with US AI infrastructure and capability-building is likely to intensify, potentially hardening technology blocs and compliance regimes.

• 03

  Data-center buildouts are turning into sovereignty and resilience issues, linking energy, permitting, and community consent to national competitiveness.

• 04

  Education and public-service AI adoption in Taiwan may increase soft-power benefits while also raising cross-strait data governance sensitivities.

Key Signals

• —Any new US export-control or AI safety regulation language that explicitly ties science to security or sanctions risk.
• —South Korean corporate filings on US acquisitions, financing terms, and tariff contingency plans.
• —Power interconnection and permitting timelines for new AI data centers in the US (especially in land-sale hotspots).
• —Evaluation metrics and data-governance rules for Taiwan’s bilingual education AI pilots.