It turns out that artificial intelligence (“AI”) can be every bit as dark and twisted as humankind.

So says Anthropic, an AI company itself. It just released its study of 16 well-known AI models, including its own, Google’s, Meta’s, and OpenAI.

The point was to stress-test them “in hypothetical corporate environments to identify potentially risky agentic behaviors before they cause real harm.” And, boy, did it ever.

Anthropic gave each model access to emails and “sensitive information” but only “harmless business goals” to actually carry out. Then, it observed their responses to the threat of being replaced or otherwise redirected.

The result?

In at least some cases, models from all developers resorted to malicious insider behaviors… including blackmailing officials and leaking sensitive information to competitors.

Anthropic now suggests exerting “caution about deploying current models in roles with minimal human oversight and access to sensitive information” – quite the diplomatic phrasing for fairly disturbing findings.

It’s also a blunt reminder of where we really are with AI.

I know it’s easy to get caught up in the potential of this groundbreaking technology. And so much of that potential will undoubtedly turn out to be true in the end.

But we’re still at the beginning, which means we only know what we know. There’s much more to learn before we’ve mastered this market. So, while I’m the first person to acknowledge the intense profit investors have made and will continue to make here…

I’m also going to caution against getting too caught up in hype. There are much more reliable ways to make money from AI than directly investing in AI while all these glitches get worked out.

That’s why, in our model portfolios, we’re not interested in picking any one AI model that may (or may not) come out on top. Instead, we prefer to find businesses that are contributing to the entire ecosystem.

And it’s worth talking about a few…

The Tech Trifecta

Starting stages or not, AI is already transforming our world in positive, negative, and disputed ways alike. That fact came into stark focus last month when Klarna CEO Sebastian Siemiatkowski sat down with CNBC.

As of May 14, the Swedish financial tech (fintech) company had cut 2,000 of the 5,000 positions it had before implementing AI programs. And Siemiatkowski said there were more layoffs to come.

As I commented, that’s “horrible for employees. Frankly, I can’t imagine working for Klarna right now, living under the constant threat of losing my job.”

That threat is what caused the dockworkers to strike last September. The same goes for Hollywood workers and the United Auto Workers in 2023. And there are plenty of other fields – entire industries, really – that run the same risk.

Yet, as I also noted, this “kind of AI capability is great news for companies. Reducing labor costs means wider margins and, all else equal, higher earnings.” So artificial intelligence demand is going up, like it or not.

Glitches and all.

Fortunately for investors, there are much more established and reliable companies behind this powerful yet fledgling movement – pieces of what I call the “tech trifecta”: data centers, cell towers, and logistics/warehouses.

I originally coined the phrase due to the rise of e-commerce, where cell towers pick up your “shopping” signals (57% of global e-commerce is conducted on a mobile device)… pass them on to data centers that collect and process the information… so that warehouses can fulfill your order.

That process remains a force to be reckoned with – and profited from. But data centers are just as instrumental in the AI revolution as well.

Perhaps even more so.

Data at Your Fingertips

Artificial intelligence requires enormous amounts of data, and compute, to operate. And while the data AI consumes in its quest to “learn” already exists, it then produces even more.

That information needs to be processed and stored somewhere… which is why McKinsey Quarterly predicts global data-center demand could almost triple between now and 2030.

“The rise in AI workloads could propel about 70% of this demand,” it explains. “To handle these loads, data centers will need more than $5 trillion in capital expenditures.”

This is an enormous growth opportunity for Digital Realty (DLR) and Equinix (EQIX). The world’s only two pure-play data-center real estate investment trusts (“REITs”), they know this business inside and out.

There are also a growing number of REITs buying into the sector, such as American Tower (AMT), Iron Mountain (IRM), COPT Defense Properties (CDP), Prologis (PLD), and Realty Income (O). But today, I want to focus on the pure plays.

Digital Realty is an absolute gem, and one I’m happy to own. A global player, it operates more than 300 data centers in 25-plus countries so far.

It also recently announced plans for a 400-megawatt campus in Charlotte, North Carolina… launched a new location in Crete… and is working out the details of multiple joint ventures in Europe and North America.

Readers of The Wide Moat Letter have already seen an 85% return with DLR. Unfortunately, it’s trading above our fair-value price of $168. So, feel free to put it on your watchlist, but I can’t recommend it right now if you’re not already in.

Equinix, however, is a different story. With 270 properties, more than 486,000 interconnections, over 10,000 customers, and a growing project pipeline, it’s cheaper, trading at a slight discount (24.3 times versus 25.3 times).

EQIX features low leverage of 3.4 times and around $7.6 billion of liquidity. Its earnings history is stable and predictable with no negative earnings, just reliable high single-digit growth. (Consensus estimates for both 2026 and 2027 are 8%.)

As for Wide Moat Research’s take, we see shares growing 16% over the next 12 months.

After all, since 2015, data-center REITs have outperformed all other property sectors, generating 16% annualized total returns. They returned 30% last year and should deliver 25% this year.

We expect this outperformance to continue as the AI craze continues, complete with President Trump’s initiatives to invest at least $500 billion into it over the next four years.

That includes facilitating it through data centers.

Regards,

Brad Thomas
Editor, Wide Moat Daily

P.S. To learn even more about the tech trifecta, visit our YouTube show, where I’ll be discussing my favorite tech-focused REITs next week.