The AI Trough: A Perfectly Predictable Dip

by

The markets, as they are wont to do, have recently decided to be…challenging. For artificial intelligence stocks, it’s been less a challenge and more a rather insistent dunking. Microsoft, for example, is currently experiencing a significant gravitational disagreement with its previous peak, down over 20%. Broadcom isn’t faring much better, and Oracle…well, Oracle is contemplating the wisdom of spending a lot of money on something that may or may not actually do anything. (It’s a feeling many companies are intimately familiar with, if we’re being honest.)

What’s going on? Simply put, investors have discovered that AI, despite all the breathless pronouncements, isn’t free. Nor is it a magic wand. It’s…complicated. And, shockingly, it doesn’t solve every problem instantly. (The sheer audacity of expecting it to is almost…endearing.) The market, in a fit of belated rationality, is now re-evaluating things. Which, historically, is usually a good sign…eventually.

Don’t, however, abandon ship just yet. Consider this not a catastrophe, but a perfectly predictable lull in the storm. A momentary pause before the inevitable, and usually slightly chaotic, resumption of progress. This isn’t a failure of the technology, it’s a demonstration of the inherent human tendency to wildly overestimate things, then be mildly surprised when reality intrudes. (It’s a pattern stretching back at least to the invention of the wheel, probably further.)

We are, as it happens, currently residing in what Gartner, a company dedicated to charting the unpredictable whims of technology adoption, calls the “Trough of Disillusionment.” It’s a rather dramatic name, suggesting a desolate, windswept valley of despair. In reality, it’s just a temporary dip in enthusiasm, reliably followed by a recovery. (Think of it as the technological equivalent of needing a lie-down after a particularly enthusiastic parade.)

Gartner’s Hype Cycle: A Map of Collective Delusion

Gartner, bless their data-driven souls, have codified this cycle of hype and disappointment. It consists of five stages, a sort of emotional rollercoaster for emerging technologies. Let’s break it down, shall we?

  1. Innovation Trigger: Someone builds something. It works. No one is quite sure what it’s for, but it’s shiny.
  2. Peak of Inflated Expectations: Everyone suddenly realizes what it’s for! It will solve all our problems! Funding pours in! The champagne flows! (The champagne is, historically, a poor indicator of long-term success.)
  3. Trough of Disillusionment: It turns out it can’t actually do all those things. Some companies fail. Investors get grumpy. The champagne supply dwindles.
  4. Slope of Enlightenment: Costs come down. Functionality improves. People start figuring out what it’s actually good for. Practical applications emerge.
  5. Plateau of Productivity: The technology becomes commonplace. Unprofitable companies have vanished. The remaining ones make money. (And the cycle begins anew.)

Experienced investors have seen this play out countless times. Virtual reality, solar panels, VoIP, 3D printing, speech recognition…all went through the same phases. Initial excitement, followed by a sobering dose of reality, then, eventually, quiet, profitable existence. (It’s a bit like watching a particularly energetic puppy eventually settle down for a nap.)

The dot-com boom and bust of the late 90s and early 2000s remains the definitive example. Many companies vanished, leaving behind a trail of broken dreams and overpriced office furniture. But the survivors – Amazon, Google, eBay – became the cornerstones of the modern internet. (A rather dramatic transformation, when you think about it.)

The Best Bets, Presently Submerged

So, here we are, firmly in the AI Trough of Disillusionment. It’s almost certain that AI will play a significant role in the future. However, it’s also clear that it hasn’t lived up to the hype in every area. A recent survey indicated that over 80% of CFOs and CEOs reported no positive impact on employee productivity. (Which, frankly, is a relief for some employees.)

But that doesn’t mean it’s worthless. It means we need to focus on what AI is actually good at. It’s not about replacing office workers with digital assistants (most of whom would probably just order more stationery). It’s about applying it to tasks like cybersecurity, forecasting, and image editing. (Things that don’t require nuanced judgment or a sense of humor.)

With that in mind, which AI stocks are worth a look? Oracle, despite its origins as a database provider, is making a promising shift towards AI. Management expects AI revenue to swell from $18 billion this year to $144 billion by 2030. (A rather ambitious projection, but then again, ambition is rarely in short supply.)

Alphabet (Google) is another intriguing option. It’s been relatively resilient during this downturn, which is saying something. AI isn’t even its biggest business – that’s still search. But its cloud computing arm, where much of the AI work is happening, is growing rapidly. (And, crucially, it’s better positioned to capitalize on future growth.)

Loading widget...

Google’s Gemini is gaining ground on ChatGPT, and Google Docs is steadily displacing Microsoft Office. (A quiet revolution, unfolding one spreadsheet at a time.) Google Cloud is also outperforming its competitors. (Which, admittedly, isn’t always a high bar.)

Alphabet seems particularly well-positioned to serve institutional and enterprise clients – the ones who actually have problems that AI can solve. (And, crucially, the ones who can afford to pay for those solutions.)

And, if you’re looking for something a little more off the radar, Recursion Pharmaceuticals (using AI for drug discovery) and UiPath (specializing in workflow automation) are worth a look.

The one AI name that’s arguably not worth scooping up right now? Microsoft. Despite its overall strength, it hasn’t demonstrated clear superiority in any aspect of the AI business. (Dominance, as any historian will tell you, is a key ingredient for success.)

Read More

2026-03-22 20:32