The safest place to be if the AI bubble bursts

The old saying goes that a picture is worth 1,000 words. But the chart Bank of America Global Research released this week might very well be worth a trillion.

The good folks at Yahoo Finance spiffed the image up with brighter colors to make it easier to read. But even in its original form, it perfectly sums up the 2026 market environment.

What this chart does is track cash flow from hyperscalers – the companies spending hundreds of billions of dollars annually to build out artificial intelligence (AI) infrastructure – and the biggest semiconductor companies that are filling those data centers with chips.

As you can see, the green line, the hyperscalers’ cash flow, has fallen off a cliff. Cash cows though they once were, they’ve willingly handed all their profits to semiconductor stocks (represented by the purple line).

This does beg the question of why. Though, fortunately for us, the answer is a pretty easy one.

Hyperscalers believe that AI is going to be the next big thing. That it’s going to totally revolutionize the world. And, ultimately, that the winners in this space will generate massive profits in the future.

In short, they’re trading cash flows in the present for the chance at even bigger ones later.

As I’ll explain shortly, I believe they have very good reason to think this bet will pay off. But the market isn’t convinced anymore, as evidenced by this next chart.

It shows hyperscalers’ year-to-date performance through the Roundhill Magnificent 7 ETF (MAGS) in blue versus semiconductors, which are represented by the iShares Semiconductor ETF (SOXX) in pink and VanEck Semiconductor ETF (SMH) in orange.

That might be the prevailing opinion on the subject. But here’s why I don’t believe betting against hyperscalers is in your best interest…

On the one hand, investment dollars are following the profits. And, really, that’s what they should be doing.

The fact is that earnings and cash flows from AI-specific semiconductor stocks are rising at unprecedented rates, often more than justifying their strong share price performance. From a fundamental perspective, this rally makes perfect sense.

If anything, Nvidia (NVDA), the world’s largest semiconductor stock, is cheaper today on a forward-looking price-to-earnings (P/E) ratio basis than it was back in 2019… even though the stock has climbed more than 1,000% since then!

That’s why I have to disagree with everyone who keeps chanting, “Bubble. Bubble. Bubble.” In fact, I’ll come right out and say it.

I think all the AI bubble talk is hogwash.

With that said, I’m not really looking at the semiconductor space for investment ideas right now. Sure, it’s been a great place to be, but I’m rarely one to chase momentum – even when the fundamentals support it.

You can make money that way, but not as much as when you look for out-of-favor opportunities elsewhere. That’s why I’m less interested in the purple line from that previous chart and more intrigued by the green one.

It’s not just that hyperscaler companies are trading on the cheap that attracts me. It’s the fact that they’re well-founded, well-managed, and well set up to keep profiting in the future.

These companies became as big as they did because of their growth capabilities and massive profits, which led to high-margin, asset-light business models. For a while there, they actually had more cash than they knew what to do with.

But then along came ChatGPT, shifting the game drastically. Just like that, hyperscalers went all in on the global AI arms race, directing just about every bit of that glorious cash flow into a single purpose.

That might sound irresponsible. It might look like it, too. But the chance that every single one of these intensely successful companies actually lost their better judgement overnight seems slim.

So I’m going to trust the men and women who created that wonderful cash-flow situation in the first place to get us to something bigger and better still.

This isn’t blind faith though. It’s not even only built on hyperscalers’ past results – which I’m very well aware are not a perfect indicator of the future.

I actually see two fairly straightforward ways for these companies to come out ahead.

First, let’s say I’m right that the verifiable geniuses who run them know what they’re doing. In that case, all of this AI-related spending will result in fantastic returns on investment (ROI), widening their competitive moats and resulting in trillions of dollars’ worth of high-margin sales over the next 5-10 years.

Cash flows will continue to rise despite astronomically high AI capital expenditure (capex) budgets. And hyperscalers will have set themselves up as primary financial beneficiaries of the AI revolution.

There’s that possibility.

Or, on the flipside, the AI bears are right.

Perhaps there’s a slowdown in innovation. Maybe technological breakthroughs decrease data center demand. Or perhaps China wins the AI race, leaving American hyperscalers in the profitless dust.

In any of those scenarios, yes. The hyperscalers would ultimately lose out on the eye-popping amounts of money they spent. But even then, they’d still be worth investing in.

Think about it…

The cash flows from their legacy businesses do remain in place, after all. So hyperscalers could simply pull back or entirely end their AI-related capex, sending their cash flow numbers through the roof again because of the strength of their other revenue streams.

Amazon (AMZN), Microsoft (MSFT), Alphabet (GOOG), Meta (META)… These companies are all great cloud providers that excel in the extremely profitable enterprise software and digital advertising industries.

Demand in those industries isn’t going anywhere anytime soon, no matter what happens with AI. In which case, hyperscalers should still ultimately come out ahead.

As far as I can see, this is about as much of a win-win situation as the market can give. So everyone who wants to stand on the sidelines, shaking their heads, and scoffing at the AI trade is more than welcome to continue doing so.

I just can’t join in.

Kind regards,

Nick Ward
Analyst, Wide Moat Research

We put together another great Wide Moat Show episode for you this week titled “7 Strong Buys”… two of which come with especially attractive price points.

There are always worthwhile bargains to be found in the stock market, no matter what. And we’re proving that with finds like:

Discover all this and more when you click right here.