It's been over three years since OpenAI's ChatGPT introduced the world to the power of generative artificial intelligence (AI), and the megatrend shows no signs of stopping. Goldman Sachs predicts big tech companies could spend over $500 billion on data center equipment in 2026 alone, and a meaningful share of this will go toward Nvidia's (NVDA 1.53%) graphics processing units (GPUs) and other types of computing hardware.
That said, half a trillion dollars is a lot of money -- even for a group of the biggest hyperscalers in the world. They may be able to keep that level of spending up this year and possibly next year, but what about over the next five years? What might the future bring for Nvidia's business model?
NASDAQ: NVDA
Key Data Points
AI spending remains robust, but cracks are forming
Goldman's prediction about big tech's AI spending in 2026 was made in late 2025. And so far, the available data suggests that hyperscaler spending will meet or exceed that expectation. This month, in conjunction with its latest quarterly report, Alphabet revealed plans to almost double its capital expenditures year over year to between $175 billion and $185 billion in 2026. E-commerce and cloud computing giant Amazon announced that it will spend an eye-popping $200 billion on capex this year -- a sum $50 billion higher than Wall Street expected.
Big tech companies continue to pour money into AI to avoid falling behind their competition, even though the tangible returns from this spending remain unclear. The situation benefits Nvidia in the short term because its GPUs account for such a large fraction of the cost of building data centers. But long-term investors should focus more on the sustainability of this business model.
Like all companies, Nvidia's customers operate their businesses with the goal of earning profits and delivering value to their shareholders. And shareholders are getting nervous. In fact, Amazon's stock shed around a tenth of its value after it revealed its $200 billion spending plan. Capex represents a drain on cash flow that could be used for other things, like share buybacks or dividends. And over the next few years, tech company shareholders could pressure management to cut such spending and become more efficient.
Nvidia remains vulnerable to an AI spending slowdown
Trading at a forward price-to-earnings (P/E) multiple of 22, Nvidia's valuation based on its expected earnings is incredibly cheap. Fourth-quarter revenue grew by 62% year over year while earnings per share jumped 67%. A company with such explosive performance could realistically trade for a valuation two or three times higher than current levels without Wall Street batting an eye. That relative discount suggests the market is skeptical about Nvidia's ability to maintain its current business model over the long term.
Image source: Getty Images.
There are some good reasons for investors to be nervous. For starters, while the generative AI market has boosted Nvidia's growth, it has taken away the company's diversification. The data center segment accounted for a whopping 90% of fourth-quarter revenue, which means the company would be very vulnerable to a slowdown in AI GPU demand.
And while the company has built an economic moat around its chips through its CUDA programming platform, which allows developers to get the most out of Nvidia hardware (and which is not compatible with its rivals' chips), it can't prevent clients from leaving the ecosystem altogether by building their own AI chips. Nvidia's biggest customers are well capitalized and technologically advanced. And some have already begun to supplement the company's expensive GPUs with less expensive alternatives like Google's Tensor Processing Units.
Where will Nvidia be in five years?
AI data center demand remains massive, and that means Nvidia will probably continue to display strong growth over the next few years. That said, AI capital spending could eventually drop to more reasonable levels, and the chipmaker will face substantial competition from its own clients as they invest in custom chips.
The good news is that Nvidia has a rock-bottom valuation, which means the stock price is unlikely to crash. However, investors may want to wait for management to diversify the company's revenue streams before considering opening a position.
