During the Cold War, the Soviet Union could threaten the United States militarily. It could never match us technologically. China can do both. That single difference rewrites the strategic calculus, the policy, and where capital flows for the next thirty years.

The USSR's parity was episodic. Sputnik shocked the country in 1957, but the breakthroughs were narrow and never diffused into productivity. The regime could build a bomb and launch a satellite. It could not connect research to production, or tolerate the experimentation innovation requires. By 1985, computer backwardness alone cost an estimated 10 percent of national income. Because free information threatened political control, Soviet researchers were cut off from the global ecosystem exactly when integration became the whole game.

When the United States answered with NASA, semiconductors, and the computing revolution, the contest was over, and the world drew a lesson: market-led innovation beats state-directed industry at the frontier. Venture capital ran on it. Decentralized bets on technical founders, unconstrained by government priorities, produced Google, Facebook, and the entire SaaS ecosystem. The contrarian insight, patiently held, was the winning move.

China broke the pattern. CSIS's James Andrew Lewis calls it market-Leninism^*, market mechanisms for dynamism, the Party for control.

*Market-Leninism: in simple terms, China lets companies compete and prices move like a real market, but the Party can step in whenever it wants to move money where it likes, seize a business, or bend an entire industry to its goals.

The numbers carry the argument. For more than a century, no US adversary or coalition reached 60 percent of US GDP. China is past that. It manufactures more than twice what the United States does and trails in almost nothing except semiconductors and aviation. It controls over 80 percent of every stage of solar manufacturing, and more than 95 percent of the polysilicon and wafers upstream. It holds 80 percent or more of battery cell production and refines about 91 percent of the world's rare earths.

When technological leadership is contestable, the old playbook stops working. With no peer competitor, markets optimize for efficiency, capital chases the highest private return, and software scales with almost no friction. That was 2008 to 2020. With a peer rival, efficiency turns fragile, supply chains become weapons, and private return diverges from national survival. Markets chase margin, not sovereignty.

So industrial policy came back, because the alternative was losing. The CHIPS Act^* put up $280 billion, the Inflation Reduction Act^* close to $400 billion, the first time since the Cold War that Congress aimed industrial policy at a named rival.

*CHIPS Act: a 2022 law that put roughly $280 billion toward semiconductor research and manufacturing in the US, including about $52 billion in direct subsidies for building chip factories, explicitly to counter China.

*Inflation Reduction Act: a 2022 law originally scored at $369 billion (later estimates run far higher) for clean energy and domestic manufacturing, mostly through tax credits, with energy security as the strategic aim rather than environmentalism.

A narrow set of sectors turns disproportionately important: defense, energy, transportation, robotics, space. Each has the same shape, a committed government buyer underneath and a physical product that cannot be cloned in software.

This should change how capital gets allocated. TAM matters less than mandate: a company tied to a national priority has a demand floor no consumer startup can touch. The lowest-cost producer is not the winner if it sits in an adversary nation and can be cut off. Winner-take-all weakens, because governments refuse single points of failure and cultivate second and third vendors. And software alone captures little. In hardware-intensive industries, the platform that does not own the underlying asset takes almost none of the economics. The winners look less like Salesforce and more like Tesla, one loop of hardware, software, and manufacturing.

The last cycle rewarded the investor who saw what everyone missed and waited. I think that this one rewards whoever builds what a government needs and owns the hardware underneath it.