The U.S. economy produced the world's most valuable companies for more than a century:

  • United States Steel became the first-ever $1 billion company in 1901.
  • General Motors rode the automotive boom to become the first $10 billion company in 1955.
  • General Electric built a conglomerate which became the first $100 billion enterprise in 1995.
  • Apple then crossed the most exclusive milestone of all in 2018, when it became the world's first $1 trillion company.

Eight other American companies have joined Apple in the trillion-dollar club since then: Nvidia, Microsoft, Amazon, Alphabet, Meta Platforms, Tesla, Broadcom, and Berkshire Hathaway.

I predict another company is likely to join them within the next few years. Oracle (ORCL -2.17%) operates some of the most powerful and most cost-efficient data center infrastructure in the world for artificial intelligence (AI) workloads. Demand is significantly exceeding available supply, as some of the industry's top developers are lining up around the block to use it.

Oracle had a market capitalization of $600 billion, so investors who buy its stock today could earn a whopping 65% return if it climbs into the $1 trillion club. Here's why I think it will.

A person looking down at a tablet device while standing in a data center.

Image source: Getty Images.

A leader in the AI data center race

Every new generation of AI model requires more computing capacity than the last. Nvidia CEO Jensen Huang says some of the latest "reasoning" models -- which spend more time before rendering a response to produce more accurate information -- use up to 1,000 times more computing power than traditional large language models (LLMs), which specialize in crafting one-shot responses.

That computing capacity is delivered by enormous data centers that are filled with graphics processing units (GPUs) from top chip suppliers like Nvidia and Advanced Micro Devices. Most businesses don't have the financial resources to build this infrastructure themselves, so they rent it from companies like Oracle that operate centralized data centers situated all over the world.

Oracle's Gen2 Cloud data centers allow developers to scale up to 131,072 of Nvidia's industry-leading Blackwell GB200 GPUs, paving the way for the most powerful AI models to date. Plus, Oracle's infrastructure uses a proprietary random direct memory access (RDMA) networking technology, which moves data from one point to another faster than traditional Ethernet networks. Developers typically pay for computing capacity by the minute, so speedy processing can translate into substantial cost savings.

Gen2's scalability and cost efficiency are the reasons top AI developers like OpenAI, Meta Platforms, and Elon Musk's xAI are lining up to use it. However, Oracle simply doesn't have enough capacity to meet demand right now, so it's planning to spend over $25 billion to build new data centers during its fiscal year 2026 (which began June 1).

Chairman Larry Ellison says that Oracle will eventually operate more data centers than every other player in the industry combined. He thinks the company can scale up to between 1,000 and 2,000 locations over the long term, which would be more than 10 times its current footprint.

Oracle's revenue growth is accelerating

Oracle generated $15.9 billion in total revenue during its fiscal 2025 fourth quarter (ended May 31). It was an 11% increase from the year-ago period, which represented an acceleration from the 6% growth the company delivered in the third quarter three months earlier.

But the real growth story lies beneath the surface of the headline number. The Oracle Cloud Infrastructure (OCI) segment -- which is where the company accounts for its AI data center revenue -- soared by 52% year over year to $3 billion. And it gets better, because Oracle CEO Safra Katz predicts that OCI's revenue growth will accelerate to 70% throughout fiscal 2026.

I mentioned earlier that demand for Oracle's data center capacity is far exceeding supply. That showed up in the company's remaining performance obligations (RPOs) in Q4, which surged by 41% to a record $138 billion. RPOs are like an order backlog that Oracle expects will convert into revenue over time, so they give investors an indication of how much future demand is currently in the pipeline.

Oracle's (mathematical) path to the $1 trillion club

Oracle generated $4.34 in earnings per share (EPS) during fiscal 2025, which places its stock at a price-to-earnings (P/E) ratio of 49.6. That isn't necessarily cheap, since the Nasdaq-100 technology index -- which hosts all of Oracle's big-tech peers – trades at a P/E ratio of 30.6.

However, Oracle's premium valuation might be justified for now considering its accelerating growth and its substantial RPOs, which will give investors confidence that the company can sustain its momentum.

Wall Street's consensus estimate (provided by Yahoo! Finance) suggests that Oracle's EPS could soar to $6.75 in fiscal 2026, and then to $8.17 in fiscal 2027. Based on those forecasts, Oracle stock trades at forward P/E ratios of 31.9 and 26.3, respectively.

That means that Oracle stock would have to soar by 88% over the next two years just to maintain its current P/E ratio of 47.2, which would catapult the company's valuation to over $1 trillion. To be clear, it won't be easy for Oracle to maintain its premium P/E ratio unless it can convince investors that OCI growth will continue to accelerate. This isn't out of the question, since the company plans to expand its data center footprint by more than tenfold in the coming years.

Oracle might still have a pathway to the $1 trillion club even if its P/E ratio shrinks, except it could take four or five years instead of two. In any case, the stock could be a great buy for investors right now.