Nvidia's $5 Trillion Problem: Too Big for Markets to Price

Nvidia hit $5.5 trillion in market capitalisation this week, a milestone that ought to be celebrated in the company's Santa Clara headquarters and instead should probably trigger a quiet conversation in rooms where people worry about systemic risk.

To contextualise the number: Nvidia is now worth more than the entire annual economic output of every nation on Earth except the United States and China. It is the world's second-largest asset by value behind gold at $32.6 trillion, but ahead of silver at $4.9 trillion. The stock has gained more than 50 per cent this year and more than 1,500 per cent over the last five years. These are not the metrics of a cyclical semiconductor business. These are the metrics of infrastructure.

The transformation is real enough. In 2021, Nvidia's market cap hovered around $345 billion. Gaming revenue was collapsing. The company looked like every other chip maker riding the end of a cycle. Then generative AI arrived, and Nvidia became the foundational hardware layer upon which nearly every enterprise AI system is built. Revenue reached $216 billion in its most recent fiscal year. Free cash flow stood at $96.6 billion. These numbers do not leave much room for disappointment.

But here is where the story stops being simply about a successful business and starts being about something else entirely: the question of whether a semiconductor company can become too strategically important for governments to let fail.

Consider the geopolitical architecture Nvidia now inhabits. The company has achieved its $5.5 trillion valuation without full access to China, constrained by U.S. export restrictions that prevent it from selling its most advanced AI chips into the world's second-largest economy. That is not incidental. That is deliberate policy, and it reflects how thoroughly Nvidia has become embedded in American strategic thinking. CEO Jensen Huang is now accompanying President Donald Trump on China-related trips alongside other technology executives. Nvidia's stake in Intel serves as what one might charitably call political insurance, aligning the company with the White House's stated goal of preserving American semiconductor manufacturing on American soil.

Meanwhile, companies like Nvidia depend almost entirely on Taiwan's chip manufacturing infrastructure. TSMC produces Nvidia's most advanced processors. Should geopolitical tensions with China escalate to the point of military conflict, Taiwan's semiconductor industry becomes not merely at risk but potentially subject to a deliberate 'kill switch'—a scenario where the transfer of trade secrets is prevented and production halts entirely. Nvidia's valuation assumes this does not happen. It assumes Taiwan remains stable, accessible, and committed to building the processors that power American AI dominance.

The uncomfortable math is simple: if Nvidia becomes systemically important enough to American economic and military advantage, the calculus around whether governments can afford to let it stumble changes entirely. Too big to fail is a term we associate with banks. It is a term we should probably start thinking about for semiconductor infrastructure.

None of this is to say Nvidia is in trouble. The company generates extraordinary cash, faces no immediate competitive threat at the high end of the market, and has built genuine moats around its software ecosystem. But valuation multiples that assume years of continued AI dominance contain embedded assumptions about competition, about geopolitical stability, and about the absence of the kind of semiconductor cycles that have historically cooled after periods of rapid expansion. Hyperscalers like Amazon, Microsoft, and Alphabet are building custom AI chips designed to reduce their dependence on Nvidia. Competition is increasing. Cycles always cool eventually.

The question worth asking is not whether Nvidia is overvalued by traditional semiconductor metrics. It probably is. The question is whether traditional valuation metrics still apply to a company that has become as strategically central to American interests as Nvidia now is. At $5.5 trillion, the market may have already priced in an answer: governments cannot afford to let this one fail, so it will not. That is not investing. That is insurance. The fact that it looks like both is precisely what makes the current moment worth watching.