The artificial-intelligence buildout is driving earnings growth for the entire S&P 500, so projecting the pace of AI expansion has become a top priority for nervous investors.

There's nothing to worry about, though, if new numbers from PwC prove prescient. The consulting firm expects cumulative spending on AI to top $31 trillion by 2050. That's an incredible number, roughly amounting to the current size of the U.S. economy.

There are a couple of other observations, beyond the eye-popping headline number. For starters, it's a global number, with the U.S. representing roughly half of the total. Next, data-center capital spending will grow from about $800 billion in 2026 to $1.8 trillion a year by 2050. Power generation is a key bottleneck to growth. And, eventually, recurring semiconductor upgrades-not data-center construction-will drive "the majority of long-term capital investment."

The report offers several takeaways for investors. First, PwC calls the buildout "era-defining." Its numbers are bullish and suggest the AI trade has a long way to go.

Barron's recently suggested that the U.S. can absorb spending equivalent to about 25% of economic output before things truly become a bubble. That was the pattern in prior eras, including railroads, electrification, and the internet. It's a top-down rule of thumb, but it suggests $7.5 trillion in cumulative AI capital spending is doable, which implies another $5 trillion-plus is coming before debt and capital markets show signs of significant stress. PwC's projections of annual and total spending support the conclusion that AI spending is still in its early days.

Next, power generation is key. It's an issue SpaceX CEO Elon Musk has been harping about lately, and suggests, based on the projections, there is more business than investors imagined for companies such as GE Vernova, Siemens Energy, and just about any others that can make equipment that generates an electric current.

Third, it's a surprisingly bullish report for chip makers. PwC sees data centers needing to upgrade equipment from time to time. That's annual service revenue for industrial companies as well as annual sales for Nvidia and others. The maintenance of data centers isn't something investors are focused on right now, but equipment always needs to be kept in good working condition.

Bottom line, it is a bullish report that doesn't get into potential knowledge-worker job loss or AI's impact on communities and children. "The capex upside from faster AI adoption exceeds the downside from slower [AI] adoption," PwC argues. In other words, there is more reward than risk right now.

Investors don't typically trade stocks on 2050 outlooks, but maybe it's helping a little. Nvidia stock was up 4% Wednesday morning, while the S&P 500 was up 0.5%. Micron Technology stock was up 0.9%. GE Vernova stock, however, was flat.

Bond yields, corporate actions, and current valuations surely overwhelm 2050 projections, but projections still matter. AI growth is paramount for the stock market, and for now it looks as if demand growth is healthy.