Okay, Darian, would you rather eat a shoe or sign up for the Indicator Newsletter?
Sign up for the Indicator Newsletter.
Great.
You can sign up at npr.org slash Indicator Newsletter.
NPR.
This is The Indicator from Planet Money.
I'm Waylon Wong.
And I'm Darian Woods.
Steve Burke got into building computers when he was a teenager.
I built my first system when I was 15.
So my dad took me to CompUSA, which is no longer in business.
RIP.
I remember going to CompUSA with my dad.
Yeah, it was a good time.
Today, Steve runs a website called Gamers Nexus that covers the computer industry.
Over the decades, he's seen the cost of processes and memory and storage come way down.
But in the last year, the AI industry's demand for RAM or memory has sent prices skyrocketing.
Some products have doubled in price or more.
It's never been like this.
Right now, it's such a huge spike that people are seeing it reflected in any consumer electronic that has memory, which is most of them now.
Like Nintendo is raising the price of its Switch 2 gaming system in the U.S. by $50.
Apple hiked prices on a bunch of products, including an entry-level MacBook and its cheapest iPad.
Today on the show, we look at how the rampocalypse is upending a long-held assumption in computer science called Moore's Law.
This sometimes contentious idea was interpreted to mean that computing power would keep getting better and cheaper.
But is it?
That's after the break.
Moore's Law is named for the late Gordon Moore, the co-founder of Intel and a tech industry pioneer.
Here he is talking back in 2006 when Moore's Law was going strong.
You know that next year you'll be able to buy something cheaper and more powerful, providing it's electronic.
Next year's computers will be faster.
Cheaper, next year's cameras will have more megapixels, more memory, cheaper.
Is there any other product area other than electronics that gives you this kind of history?
Gordon Moore said there's something special about semiconductors.
This is the material that goes into computer chips and the electronic devices that we all rely on.
And in 1965, he made a famous prediction about the semiconductor industry.
Gordon Moore said that the number of transistors on an integrated circuit would double every year.
Transistors are the microscopic on and off switches you find on computer chips.
So Gordon Moore was saying that the industry could fit exponentially more of these switches on a chip every year.
And therefore, electronics would get faster, smaller and cheaper.
Gordon Moore's assertion has been described as a prediction or an observation.
It was never meant as a law of nature.
Still, this idea became known as Moore's Law.
I love talking about Moore's Law.
Iris Bahar is a computer science professor and head of the computer science department at the Colorado School of Mines.
It's been so foundational to computer science, computer engineering.
It's really important to understand that history and how we've benefited throughout the years to understanding where we are today and what we're trying to do today.
Iris first learned about Moore's Law as an undergraduate student, and she saw it in action during her first job designing microprocessors for a now-defunct company called Digital Equipment Corporation.
It was a major player in the 80s.
And in fact, we were living that law, if you will.
We were designing these microprocessors and we saw that we could ride that law and we were able to put more and more functionality on the same chip.
Iris says the computing industry saw tremendous growth in performance through the 90s and early 2000s.
But as chips and transistors shrunk in size, achieving these exponential gains got harder.
When you're operating in the single nanometer scale, you really are
Looking at these single electrons, this just is a real physical limit.
To put a nanometer in perspective, a single human red blood cell is 7,000 nanometers wide.
So we're talking about scientists trying to fit ever more transistors onto tiny surfaces.
Iris says occasionally some people in the industry would start drafting the obituary for Moore's law, but then there would be some breakthrough.
It's dead, it's dead, it's coming.
Oh, wait, no, we figured out how to deal with this problem.
We really, really are hitting these walls.
Now, it's still not totally dead.
Yeah, not totally dead.
Just last month, IBM announced that it created the world's smallest transistor, measuring under one nanometer.
The company said it could cram around 100 billion of these transistors onto a chip the size of a fingernail.
According to IBM, this was the first time in human history anyone's been able to pack that many transistors into a space that small.
Long live Moore's Law!
But Iris says it is requiring more creativity and money to achieve these advancements.
And we can no longer take for granted that improvements in technology are associated with cheaper prices.
And that's because of our old friends supply and demand.
Right now, demand is huge for computing power and memory.
Yes, the rampocalypse.
Iris says computing memory has benefited from Moore's law over the decades too.
Like, remember how USB flash drives got so cheap that they'd just be given away as corporate swag at conferences?
Yeah, and I'll put them straight in the garbage.
But now memory prices are going through the roof thanks to the AI boom and the rapid build-out of data centers.
These facilities require a lot of computing power and a lot of memory.
Take OpenAI.
Last year, it announced a partnership where Samsung will be one of the main suppliers of memory for OpenAI's Stargate initiative.
This is a $500 billion AI infrastructure project.
Steve Burke of Gamers Nexus says the amount of memory OpenAI needs for Stargate amounts to 40% of the world's supply.
It's like, let's say you go to the grocery store to buy milk.
One guy takes 40% of the milk, right?
Like, that's going to have a shock on everyone else.
And it's just gone out of nowhere, right?
It's just like this guy woke up and decided he wanted all that milk today.
And that was really, for me, the one that stands out as the sort of the peak of, okay, this is going to be a problem.
Steve says there are really only three companies that supply the U.S. with computer memory, Samsung, Micron, and SK Hynix.
Last year, Micron announced it was pulling out of the direct-to-consumer market.
The company plans to focus on larger customers in fast-growing segments like AI data centers.
AI companies have a seemingly unending appetite for computing power and the deep pockets to match.
As a result, even big tech companies like Nintendo and Apple are feeling the supply squeeze and have to raise prices.
And people like Steve's audience of computer enthusiasts are left with a bad feeling.
If you feel like you've supported a particular company or industry for years or decades, and suddenly they are posting better profits and revenue than ever, but they no longer want to sell to you, you could see how that would feel like they've kind of turned on their roots.
Computer science professor Iris Bahar says the industry benefited from Moore's law for decades, and now it's time to get creative again.
The billion-dollar question is, can technological breakthroughs keep up with the incredible demand for more computing power?
What we're doing now, actually, I don't think is sustainable.
With this need for more and more computing and memory for running AI.
So we need to rethink how we should be doing computing in the future.
That's beyond Moore's Law.
Who knows, maybe in a few years, it will once again be time to dust off the Moore's Law death notice.
This episode was produced by Corey Bridges with engineering by Robert Rodriguez.
It was fact-checked by Sierra Juarez.
KK Cannon is our editor and The Indicator is a production of NPR.
