The AI Rally's Dirty Secret: It's Built on Geopolitical Sand

But the AI trade keeps chugging along, apparently immune to the fact that the companies driving these gains depend on:

• Global supply chains that run through regions currently experiencing, you know, active military conflict
• Massive energy consumption at a time when oil prices are climbing
• Consumer and enterprise spending that tends to contract when people start worrying about broader wars
• Regulatory environments that get unpredictable when governments are focused on security rather than innovation policy

None of this means AI isn't real or important. It absolutely is! I've been covering robotics and autonomy long enough to know the technology is genuinely transformative. But there's a difference between "this technology matters" and "this stock price makes sense right now." The market is pricing in a future where everything goes right, while the present is actively going wrong in multiple theaters.

And then there's the tariff situation, which everyone seems to be ignoring because it's less exciting than chatbots. The Trump administration is proposing new tariffs of at least 10% on imports from 60 trading partners. Sixty! That's not a targeted action, that's a wholesale rebuilding of the protectionist wall that the Supreme Court struck down in his earlier term. For an industry that depends on chips from Taiwan, rare earths from various places, and manufacturing capacity spread across the globe, this should be setting off alarm bells.

It remains unclear how these tariffs will actually be implemented, or whether they'll survive legal challenges this time around. But the uncertainty alone is enough to make long-term planning basically impossible for hardware companies. And guess what AI runs on? Hardware.

I talked to a few portfolio managers this week (off the record, because nobody wants to be the guy calling the top). The consensus, sort of, is that the AI rally has its own momentum now, divorced from fundamentals. One of them compared it to a train that's gone too fast to stop safely, you just have to hope the track doesn't run out. Not exactly a ringing endorsement.

The thing that gets me, and maybe I'm just old and cranky at this point, is how short everyone's memory is. We just went through a period where supply chain disruptions sent the whole economy into chaos. We just watched a pandemic expose how fragile our interconnected systems actually are. And now we're back to pretending that AI companies exist in some frictionless void where wars and tariffs and energy prices don't matter.

They matter! They've always mattered! The fact that we're in a "risk-on" mood doesn't change the underlying reality that the Middle East is one miscalculation away from a broader conflict, that US-China relations remain tense, and that the policy environment for tech companies is about to get a lot more complicated.

I'm not saying sell everything and hide in a bunker. I'm saying maybe, just maybe, the longest winning streak in over a year is a moment to take some profits rather than double down. The AI revolution will still be there in six months. The question is whether your portfolio will be.

Look, I've been wrong before, plenty of times. The market can stay irrational longer than I can stay solvent, as the saying goes. But I've also watched enough cycles to know that the moment everyone agrees something can only go up is usually right before it goes down. And right now, the AI trade has achieved that kind of unanimity.

The fundamentals will reassert themselves eventually. They always do. Whether that's next week or next year, I genuinely don't know, it's too early to say with any confidence. But when US forces are actively striking Iranian command centers and the administration is slapping tariffs on 60 countries simultaneously, maybe "longest winning streak in a year" isn't something to celebrate.

Maybe it's something to worry about.

But what do I know. If you want to argue, my email's on the about page.

Mark Kowalski · 4 days ago · 6 min