To The Moon

There’s a fine line between charging a premium because you can and acting like a mobster to your customers.

$Apple(AAPL)$ can charge a premium because people value their products more highly than competitors.

On Shoes can charge a premium because it has a premium shoe brand among dozens of other choices.

$Ferrari NV(RACE)$ can charge whatever it wants because…it’s Ferrari.

That’s pricing power.

The POWER is in the consumer’s CHOICE to pay more.

Then there’s extortion pricing. Extortion is about power, but it’s about POWER over a customer that has NO CHOICE.

Extortion is the crime of obtaining money, property, or services through coercion, threats, or the misuse of official power.

Google

This is important today because pricing power is sustainable while extortion pricing isn’t.

What we’re seeing in memory markets isn’t pricing power; it’s extortion pricing. And that kind of pricing elicits a reaction.

I’m not arguing that this isn’t logical for memory-makers in the short term. I’m making the argument that good business practices result in a win-win, and this is a win-lose scenario.

But commodities aren’t always win-win, and $Micron Technology(MU)$ and others must be enjoying the fact that the shoe is on the other foot for once.

They just need to be careful what they wish for because it won’t be there for long.

Extortion in Memory

The explosion in memory prices has been incredible over the past twelve months. Below are two examples of price charts from PC Part Picker, but these prices track similarly across the product lineup.

First is a 5x increase in DDR4.

And here’s a 5x increase in DDR5.

The impact on Micron has been gross margins going from negative to 85%!

Each price increase goes straight to the bottom line, and Micron hasn’t increased production at all.

This is the reason Micron’s stock is up ~10x over the past 18 months. And some are arguing the momentum has just begun, but I’ll get to that in a moment.

When you have what customers need, and they’ll pay almost any price to get it, you can also get them to sign deals to keep the spice memory flowing. That’s why this week we got news of 5-year supply agreements with some customers.

We are pleased to announce that we have completed 16 SCAs with customers across the data center, consumer, and auto market segments. These SCAs accelerate the transformation of our business model, enhance partnership in technology and innovation, and provide customers with contracted supply assurance. Typically, these agreements have a five-year term from calendar 2026 through the end of calendar 2030. Automotive agreements generally have a three-year term.

The 16 signed agreements represent roughly 20% of our DRAM volume and a third of our NAND volume over this period. These SCAs include four very large customers and three medium-sized customers. The remaining agreements relate to smaller customers from the automotive industry and represent our commitment to the important sector. When completed, we expect approximately half or more of our company revenue to be under these SCAs with customers across end markets.

Our customer value, our U.S. supply plans, this is reflected in our SCAs. These SCAs are structured as take or pay agreements with binding commitments to purchase specific volumes over this multi-year term. The largest agreements generally have a ceiling price for existing products at the current CQ2 market price and a floor price through the term of this agreement.

Several SCAs, which account for a modest portion of the SCA related revenue, include either fixed prices or have no price bands associated with them, where pricing will be subject to market conditions. When all planned SCAs are executed, agreements with either fixed prices or price ceilings at or close to current CQ2 market prices are expected to be approximately 40% of our revenue. For SCAs which do contain such price bands, pricing is designed to stay within this floor to ceiling level through the course of the term.

Sanjay Mehrotra, Micron CEO

Notice that customers agreed to 5-year volume deals but are paying market price within a collar. Micron gives up potential upside if prices go higher, but gets a floor price that’s likely above the long-term trend.

When you can pick and choose who gets the supply, you can make the rules on pricing, too.

What I noticed in these deals is that Micron is seeing the writing on the wall. They know selling into the spot market isn’t in their best interest, so they’re trading volume certainty for potentially lower prices.

The Extorters Get Extorted

Here’s the part of the story I love!

Micron is taking a page out of Jensen Huang’s book.

For years, demand has been so high for NVIDIA chips that Huang himself has been able to decide who gets chips and how much they pay.

Now, he’s one of the customers looking for memory supply.

Micron has him over a barrel, just like he’s long had his customers over a barrel.

$Alphabet(GOOG)$ $Amazon.com(AMZN)$ $Microsoft(MSFT)$ $Meta Platforms, Inc.(META)$, and others have paid what they’ve had to short-term to get chips from Huang. But they’re showing the danger of extorting customers because they’re all building their own supply of custom chips long-term.

Extortion pricing doesn’t last forever. And when it ends, the results can be disastrous.

Why Extortion Pricing Doesn’t End Well

Extortion pricing in commodity markets is especially dangerous because there’s no differentiation, and there’s ultimately pressure on both the supply and the demand side.

1. Extortion pricing invites competition

2. Extorted customers begin to look for alternatives

A great example of this in the past decade is rare earths. In 2011, rare earth prices skyrocketed because China (where most rare earths are from) cut off exports. Molycorp owned the only mine close to production in the U.S. and reached a valuation of ~$4 billion in 2011/2012.

But as rare earths (like Neodymium) tend to do, prices plunged, and Molycorp went bankrupt by 2015.

The problem with rare earths was China's control of supply, a multi-decade issue rearing its ugly head. But when China opened exports back up, prices plunged.

Demand is an issue, too. When prices go up 10x, customers take action. They designed magnets with other materials or changed to designs that didn’t require rare earths.

Customers have agency, too.

Rare earths are the most acute example I can remember in my career because it was such a unique mix of supply, demand, and geopolitics that could cause the spikes and crashes you see above.

But memory is the same dynamics, just on a larger scale.

Demand for memory products is up because AI eats memory, and supply can’t easily be ramped. But eventually supply will come, and customers will adapt.

The problem has gotten so bad that Apple announced sweeping price changes yesterday to its product lineup.

Is Apple going to just raise prices and shrug?

No! Apple is already looking for alternatives.

High prices for memory mean high profits, and that means someone is going to expand production to try to exploit the opportunity.

That increased supply leads to more of a supply/demand balance, and eventually the industry overshoots and prices crash.

And who is investing in capex to expand?

Everyone!!!

And this is just the Big 3. YMTC and CXMT in China are expanding aggressively and sourcing equipment to make chips from Chinese sources.

A combination of increased supply and more discerning demand could crash prices for memory.

When customers are being extorted, they look for alternatives and often find them.

This Too Shall Pass

What am I doing about this current chip/memory situation?

I’m not going to run out and buy memory stocks now. They’re likely at/near a peak in profitability, and as customers look for alternatives or reduce demand and more supply comes online, their earnings will fall.

Owning stocks that are extorting their customers is fun…until it isn’t.

I am kicking myself for not making Micron a stock pick in 2025, when, in January 2025, I made this video, calling it one of my top stocks of the year.

I am also looking for opportunities where a return to “normal” in memory will be a tailwind.

Nintendo’s margins should expand when memory prices come down.

Hyperscalers will see margin pressure because of high memory prices, but does that subside in time, leaving an opportunity for Google, Amazon, and Microsoft?

Maybe there’s an opportunity where compute costs (driven by memory and NVIDIA prices) are too high today, but usage would be higher if token costs come down. Duolingo’s AI interactions in learning modules is an example here.

The memory boom is a cyclical one, and while it’s taking over the market today, I don’t think owning stocks with extortion pricing is a long-term, durable advantage.

I did miss the run-up in the meantime, though.