Nvidia sees AI spend at $3 trillion to $4 trillion by 2030

The article builds its case on a single projection: Nvidia's belief that AI spending will grow to $3 trillion to $4 trillion annually by 2030. It notes that hyperscalers (large cloud providers) are already spending billions on data centers to train AI models and run existing workloads, and that computing capacity is not yet sufficient to meet demand from an AI-first economy. That gap is the runway the article points to.

From there, the argument splits into two tiers. Nvidia is presented as the market leader in AI computing units, with $96 billion in second-quarter revenue, a $108 billion guide for next quarter, and a stated expectation of 70% revenue growth during 2027, yet trading below 15 times next year's earnings. Taiwan Semiconductor and Micron are framed as the neutral way to play the same trend, because they are foundries that supply Nvidia's competitors too. Micron is said to have more upside now because memory prices are skyrocketing amid a supply crunch, though the article says that will not last forever.

The stakes hinge on whether data center capital expenditures really keep climbing through 2030. If they do, the article's thesis is that all three names cash in; if the build-out cools, the projection that anchors the whole argument weakens. The valuation gap Nvidia shows relative to its own growth expectation is the detail likely to drive how investors read the opportunity.