☕️ Micron says the AI trade is far from over

Good morning! 👋 

It’s Hayley here while Keith is making his way back from his trip... he’s currently somewhere between Japan and consciousness. 

All three major indices are green in premarket trading this morning, which already tells you something important: despite the noise we keep seeing, money isn’t running for the exits. 

Actually, quite the opposite is taking place regardless of all this talk about how the AI trade is supposedly ‘over.’ 

Keith made this point yesterday, saying he remains genuinely puzzled by anyone who believes that’s the case. 

In fact, he summed it up neatly in a recent post on X:

Now, let’s talk about why. 

AI demand isn’t slowing. It’s becoming visible, contractual, supply-constrained, and importantly, profitable. 

Take Micron as just the latest example and proof point. The company just told us that its profits are about to be far higher than analysts expected. (Read) 

The reason is simple. 

Memory chips are in short supply. Prices are rising. And AI demand is exploding.  

Management didn’t mince words either: 

• Second-quarter profits are forecast at nearly double Wall Street expectations. 

• Memory markets are expected to stay tight beyond 2026. 

• Some major customers will only get half to two-thirds of the chips they want. 

• All of Micron’s 2026 HBM supply, including next-gen HBM4, is already locked in. 

It’s also worth being very clear about something here. 

This isn’t really a Micron story at all. 

Micron just happens to sit at one of the tightest choke points in the entire AI ecosystem, which makes it a useful window into what’s really going on underneath the headlines. 

When a company like this says demand is overwhelming supply, and will remain that way for years, it’s telling us something important about the broader AI market, not just its own business. 

And the data underneath supports it. 

• At Micron, teams using agentic AI are already seeing productivity gains of 30% or more in real-world workflows. 

• More than 80% of Micron’s professional workforce now actively uses generative AI in their day-to-day work. 
• And the high-bandwidth memory market alone, just one critical input for AI, is expected to grow from roughly $35 billion in 2025 to around $100 billion by 2028, two years earlier than previously expected. 

We’re seeing this show up well beyond traditional tech roles, too. 

Palantir’s Chief Technology Officer, Shyam Sankar, said this week that one of the most underreported shifts in AI is how quickly it’s boosting productivity on the front lines of the real economy, from factory floors to healthcare. (Read) 

By streamlining planning, training and scheduling, AI is making it profitable to hire more people, not fewer. In some cases, roles that once took years to train for are now being filled in months. 

And this is exactly where the disconnect shows up in markets. 

Even as AI fundamentals strengthen, prices don’t move in a straight line. Capital rotates, risk gets trimmed, and positions get reset.

That’s not money leaving the market but instead money moving within it. 

We’re seeing that rotation clearly right now. Some of the most crowded AI trades have cooled, while previously ignored or “boring” names have quietly caught a bid. 

A big part of this comes down to who is doing the selling. 

Institutions aren’t paid to maximise upside. They’re paid to reduce risk early. 

They have committees to answer to, models to obey, and careers to protect. So when uncertainty rises, even headline-driven uncertainty, they de-risk first and explain later. 

Retail investors don’t have those constraints. 

That can be a real advantage, if it’s used deliberately. 

It's the difference between the very thing Keith regularly mentions to you, 'playing to win vs playing not to lose'.

This week alone, a single report about a $10bn Oracle-linked data-centre financing delay helped knock AI stocks lower in the US, and even dragged Japanese tech names down overnight, despite no change in underlying demand. 

That’s not fundamentals breaking. That’s sentiment ricocheting across markets. 

So, channelling my best Keith voice: 

Oh, and before I wrap it up, there’s one data point that caught my eye: 

According to a recent Finimize survey, roughly two-thirds of retail investors plan to maintain or increase their AI exposure, even as enthusiasm around the rally cools. (Read) 

Many other studies are mimicking the same narrative.  

The real risk right now isn’t volatility. It’s missing the next leg while waiting for perfect clarity. 

Just something to think about today.  

With that, you’ve got this — I promise. 💯 

As always, let’s MAKE it a great day. 

Hayley E 😀