Mark Carney told middle powers to grab a seat at the geopolitical table. The private sector got there first

The commentary argues that private-sector firms—not middle powers—are increasingly shaping geopolitics, particularly through AI, while the U.S. and China dominate model development. Stanford’s 2026 AI Index cited 59 notable U.S. models and 35 Chinese models through 2025; South Korea was the only other country with multiple, at eight. The piece also describes AI misuse cases reported by Anthropic and says private firms are beginning to set rules amid the lack of global consensus on AI governance.

Analysis

The investable implication is not that AI companies gain unconstrained power, but that government demand and government control rise together. Over 1–3 months, watch for defense or intelligence procurement, export-control changes, and model-use restrictions: these can create revenue opportunities for AI infrastructure providers while abruptly narrowing addressable markets. Over 6–18 months, sovereign compute programs may support demand for NVIDIA and TSMC, but also encourage local capacity and supplier diversification—potentially diluting incumbent pricing power. TSMC carries an additional tail risk: geopolitical leverage increases the strategic value of its output, but does not reduce the consequences of a Taiwan disruption. SpaceX’s strategic utility could support public-sector demand, while licensing and government access decisions remain material constraints.

Contrarian view: the commentary overstates corporate autonomy. States still control procurement, export licenses, market access, and national-security designations; firms can be both beneficiaries of geopolitical competition and its instruments or targets. The piece supplies no evidence of incremental contracts, capex, or earnings revisions, so it is not a standalone catalyst for repricing AAPL, AMZN, NVDA, TSM, or SPCX. Near term, treat this as a catalyst map, not a directional signal.

Key Decisions for Investors

• No trade on the commentary alone. Avoid adding broad AI exposure until company disclosures or procurement data show geopolitical demand converting into orders, revenue, or guidance.
• For existing NVDA/TSM exposure, track hyperscaler capex and AI-related revenue conversion alongside export-license actions; reduce risk if restrictions materially impair China-addressable sales or guidance, rather than assuming national-security importance protects earnings.
• Monitor TSM-specific Taiwan shipping, insurance, and customer inventory indicators as a downside alert. A deterioration would falsify the view that strategic importance is a net support and could overwhelm AI demand.
• Reassess SPCX exposure only against verifiable contract awards, licensing decisions, or changes in government access; strategic relevance alone does not establish durable commercial returns.