20 Jul 20th July 2026

Markets Year to Date

SpaceX is now trading below its IPO price of $135 at $125. Well done to all those who agreed with me and stayed away. But even those that played the short-term idea I published just before the listing did even better. They got in around $190 and out around $220. Some of the few people who have made any money out of this joke. Barring, of course, the banks who took massive fees for the listing, the pre-IPO investors who are selling their stock to dumb retail money, and Elon Musk who, through very clever financial restructuring made himself the first ever trillionaire. Although that did not last long and he is down over $200billion since he earned that title.

The tech sector has come under some recent pressure along with SpaceX.

Those semiconductor chip stocks that were flying are now falling. Not by much and I see this as more a correction of an overextension than anything more permanent.

However, I don’t think these memory stocks, or even the GPU chip guys are really the long-term play on AI. I see them more like a commodity company like BHP than a true tech company.

That might be a strange thing to say, but hear me out.

We all understand the commodity cycle. For guys like BHP, their profits all depend on the price of the commodity they dig out of the ground. Their cost base is fairly flat, so their profit margins go up, along with profits, when demand for their commodity increases.

Think of Fortescue back in 2006 to 2008. It went from 50 cents to 12 dollars because China went on a building spree and needed Iron. All that Chinese buying pushed up the price of iron and Fortescue just happened to be in the right place at the right time. Profit margins expanded dramatically and the share price went with it.

When demand for something increases fast, it draws in other players who also want to profit from the phenomenon. That creates extra supply, which drives prices down, profit margins collapse and the stock price goes with it.

Memory stocks are no different. They are the commodity producers of AI. At present memory is the bottleneck for AI advancement. At first it was chips that could compute what was needed fast enough.

Back in 2021, before the release of ChatGPT, NVidia’s profit margin was 26%. It just so happened that its Graphics Chips worked the best for AI compute. Everyone suddenly wanted them. NVidia just happened to be in the right place at the right time. All that demand enabled NVidia to increase its prices for the same goods. Cost base to produce them remained fairly stable so profit margins increased to 56% in 2026.

The profit margin doubled and the stock price increased 5 times because they were able to increase prices.

Because these chips are now so expensive it has drawn in other players who want a slice of the action

AMD has always been an NVidia competitor as has Intel and both have made new processors as an alternative to NVidia.

And now the Hyperscalers are also getting in on the action. Google has produced what it calls a Tensor Processing Unit (TPU) which is designed specifically for AI, whilst NVidia still focuses on GPUs. Amazon has two different chips,. one for AI training and one for AI inference. Even META now has a family of chips.

Demand for chips is still high because of the worldwide data centre build out, but pricing is now competitive. NVidia has lost its advantage as the only player.

NVidia profit margins were the same in 2026 as 2025. It has lost its ability to demand whatever it wants for its chips.

And the NVidia stock price is the same as it was in October 2025. It has gone nowhere for 8 months and is unlikely to do much unless the demand and supply curve changes again.

In fact, by one measure, NVidia is cheaper today than it was in 2019.

PE is a measure of how confident the market is in the ability of a stock to increase its profits above what is expected.

The average PE for the whole market is 19

Ford, for example, trades on a PE of 10. That says the market expects very little from Ford and that is because of the maturity of the sector it works in. No one expects Ford to be able to dramatically expand profit margins and profits.

In 2019, the PE of NVidia was 21. About the same as the average stock in the market. It meant investors expected average results from the company. AI changed that.

In April 2023 NVidia hit a PE of 138.

The market expected profits to grow exponentially.

And they did. In Q2 of 2023 profits were $6billion. In Q2 of 2026 profits were $26billion.

But now that PE ratio is back to average for the market, at 19. It means, from here, investors don’t think Nvidia can grow any faster than the average stock in the market.

And that is because of the commodity cycle.

There is now more supply for chips, the demand supply curve is changing. And Nvidia no longer has an advantage.

GPUs were the first bottle neck for AI. That bottle neck was solved within 2-3 years. That is when AI hit its next bottleneck. Memory.

The same thing that happened to NVidia happened to memory stocks in the last 12 months.

Now, you know how this story will play out. Consider Micron. Or Samsung, which, incidentally, just took the record for the most profitable company ever away from Apple. Basically all from its memory division.

It won’t last. More memory supply will come on and these companies will no longer have the advantage of charging whatever they like. Profits might well remain high, but they will no longer be able to surprise the market with large jumps.

No surprise means no change in stock price.

Micron stock, MU, presently trades on a PE of 19. The market does not expect much from it. The stock has tripled in the last 12 months because of the surprise memory bottleneck in AI. The stock price has tripled because profits have gone from a loss of $6billion in 2023 to a profit of $90billion in the last 12 months.

But, from this point, the market is saying future profits will not surprise much. They might increase a little bit but no more than the average company in the whole market.

Will memory continue to surprise? I doubt it, so I don’t expect much from these stocks either, from this point in time.

All of which leads to the obvious big question

What is going to be the next Big Surprise?

I have no idea, because, by definition it will be a surprise.

But I will be reading a lot to see if I can catch it early enough to profit from the next sector to really move.

Warning

Stock values can go down as well as up. It is possible to lose 100% of your investment in a stock. Any advice given by Capital 19 is general advice only and does not take your personal circumstances into account and might not be suitable for you.