Nvidia management said it believes data center capital expenditures could reach $3 trillion to $4 trillion by 2030.

Nvidia’s stock is priced at a similar level to other big tech stocks despite superior growth.

10 stocks we like better than Nvidia ›

A look at the data would suggest that there hasn’t been a bad time to buy Nvidia (NASDAQ: NVDA) stock over the past few years.

If you bought at the start of the artificial intelligence (AI) arms race in 2023, you’re up over 1,180%. Even if you bought at its most recent peak (early November), you’re only down about 10% from your investment, which showcases the strength of the stock. Normally, when investors see a stock hovering around all-time highs, they get a bit concerned and assume that the stock is “expensive.”

However, that’s not how investors should look at it. Trying to buy the best companies means you’re likely buying into businesses that rarely trade at discounts. If it’s a great business, finding stocks trading near all-time highs isn’t necessarily a bad thing, as it shows that the business is executing at a high level and people know it.

Despite Nvidia being near its all-time high, I think it could be primed to skyrocket after Nov. 19, when it reports fiscal 2026 third-quarter results (for the period ending Oct. 31). If I’m right, investors should consider buying shares now, as Nvidia has plenty of room to run.

Image source: Nvidia.

Few have benefited as much from the AI arms race as Nvidia. It’s done so well because its graphics processing units (GPUs) are best-in-class, and the infrastructure that supports them is also top tier. Furthermore, because Nvidia gained a first-mover advantage in 2023 when the AI arms race began, many of the hyperscalers (cloud computing specialists that operate data centers) in need of GPUs developed workloads around the Nvidia ecosystem, which locked them and their clients into using Nvidia’s products for the foreseeable future. While this doesn’t prevent clients from switching to an alternative, it makes a lot more work (and possibly additional spending) to choose a different computing supplier.

The AI arms race is about to enter its fourth year, but it’s far from over. AI hyperscalers announced record-setting capital expenditures of hundreds of billions of dollars during 2025, and upped that significantly during 2026. This backs up what Nvidia told investors during its Q2 conference call, when CEO Jensen Huang noted that they expect global data center capital expenditures to reach $600 billion in 2025 and rise to $3 trillion to $4 trillion by 2030.