Meta Cuts More Jobs for AI, and Wall Street Starts Looking at Micron

Two stories in the last 24 hours point to the same question: where is AI money actually going now.

One is about Meta pushing even harder into AI while cutting deeper elsewhere.
The other is about investors trimming Nvidia and moving some of that money into Micron and HBM.

One question is whether AI is really improving productivity inside big tech.
The other is whether memory becomes the next big AI trade.

1. Meta is reportedly planning another major round of layoffs, potentially affecting 20 percent or more of its workforce

Meta is reportedly preparing for another large round of layoffs that could affect 20 percent or more of the company, or roughly 16,000 jobs.

Commentary:

If this is accurate, the direction is pretty obvious. The cuts are not hitting the core AI teams first. They are hitting Reality Labs, non AI priority groups, and a lot of normal admin and operations roles. Meta is still doing the same thing, just more aggressively: moving money, attention, and headcount toward AI.

The company has already spent heavily on this strategy. It has paid up for top AI talent, kept pushing compute spending higher, and started using AI coding and project tools to compress layers inside the company. So the picture now looks pretty blunt: hire the most expensive people at the top, cut a much larger number of people elsewhere, and call it efficiency.

The awkward part is that the output still does not look great. Llama 4 did not really land the way Meta probably wanted, and its flagship efforts still have not put the company in the same product position as ChatGPT, Claude, Gemini, or Grok.

That is why this story matters beyond the layoffs themselves. It gets at a more uncomfortable question. If AI spending keeps rising, headcount keeps falling, and the company still does not have a clear top tier product, then a lot of what gets labeled as productivity starts to look more like cost shifting.

In plain terms, Meta is forcing the company toward AI whether the rest of the organization likes it or not. Whether that works will not be judged by how many people leave. It will be judged by whether Meta finally ships something people actually see as first tier.

2. David Tepper and Michael Platt sold part of their Nvidia stakes and bought into Micron

Billionaire investors David Tepper and Michael Platt recently reduced some of their Nvidia holdings while adding to Micron.

Commentary:

This is exactly the kind of move people overread.

The fast take is always the dramatic one: money is leaving Nvidia, Micron is next. That is too neat.

The trade itself makes sense. The first phase of the AI boom was mostly about training. Bigger models meant huge demand for Nvidia GPUs. Now the market is moving into a heavier inference phase, and that puts more attention on memory, especially high bandwidth memory. Micron is one of the key names there.

So Tepper’s shift is not hard to understand. He is also adding to companies like Meta and Alphabet, which makes this look more like a rebalance across the AI stack than some big call against Nvidia. It says capital is spreading out a bit. It does not say the Nvidia story is broken.

The part worth slowing down on is the Micron comparison. Micron is not Nvidia.

Nvidia is not just selling chips. It has CUDA, software leverage, system level control, and serious pricing power. That is why it can sustain margins above 70 percent. Micron matters, especially in HBM, but it is still a memory supplier. Its place in the stack is important, but it does not come with the same control.

So the better read here is simpler. Yes, storage and memory are becoming more central to AI. Yes, money is starting to move in that direction. But that is not the same thing as saying Micron is about to become the next Nvidia.

Most important AI events from the past 72 hours

Meta delays Avocado while Musk moves into AI chip manufacturing

Microsoft locks in $174B of GPU capacity as Adobe pays $150M over subscription practices