From our other sites
The story
According to Reuters, Micron’s CEO said memory and storage supply-demand conditions are expected to tighten even further in fiscal years 2027 and 2028 compared to fiscal 2026. On 5ch’s Hardware/Industry (Geha) board, users discussed why manufacturers aren’t ramping up production despite continued AI-driven demand, with some pointing out that new fabs likely won’t come online until 2029-2030 at the earliest, while others brought up past examples of makers who nearly went under after overexpanding production. Meanwhile, on the question of when the AI bubble will burst, the thread split between those arguing it “can’t possibly burst” given the US-China development race, and reactions to Sony’s already-pricier new smartphone — with the conversation branching off in different directions.
“We expect the supply-demand environment for memory and storage to tighten further in fiscal years 2027 and 2028, compared to fiscal 2026.”
Source: reuters.com / Original article here
What people said
That'd make sense if it were smooth sailing for at least the next 10 years, but nobody knows when the bubble will pop. 2028-2030 is basically going to be the chicken-race period.
Building a fab is a multi-year commitment. How long ago did we first hear about Rapidus's semiconductor plant? Are they even mass-producing yet?
They're already running full tilt — the construction for capacity expansion won't come online for any of them until around 2030.
China isn't exactly selling cheap either, though.
These days they're building power plants right alongside the fabs.
Even with the power shortages, OpenAI just keeps buying up and hoarding memory chips.
It's basically a war between the US and China over AI development, so they're not going to let the bubble burst.
It won't burst. There's no way it can burst.
Corporate demand for AI isn't slowing down at all. If anything, companies are doubling down on AI like 'fall behind and you're dead,' so even if they wanted to stop, they can't — stopping means death, so nothing's going to change. Even companies with zero obvious connection to AI are starting to use it.
Some have pointed out that memory makers keep stoking panic because they want the bubble to drag on longer.
China did have a huge crash, though.
The reason other companies don't jump straight into memory production is that the know-how is hard to master, plus memory is inherently cheap — apparently if it ever became widespread and prices fell, it wouldn't be profitable anymore anyway.
The expanded capacity won't come online until 2029.
There aren't, actually. The capital investment required is so massive that most makers got weeded out. A memory maker CEO even called out Apple by name, saying buyers hammered prices down so hard during the cheap years that it held back capital investment.
There aren't 'tons' of them. It's an industry with a handful of dominant players where you don't even need a cartel — you can just read everyone else's pricing. It's also the end result of getting squeezed on price until the weeding-out went too far, so honestly both the manufacturers who use memory and ordinary consumers bullied the memory industry too much.
Yeah, this. Unless you're an idiot, you'd be cautious. They're already raking in more than enough profit as it is.
Just stop obsessing over graphics — a high-resolution GameCube-equivalent is all we need.
If consoles went that route, regular people would just say 'fine, I'll use my smartphone or PC instead.'
Controllers already connect easily over Bluetooth too.
Smartphones prioritize social media and web browsing, so they'll never end up being used in a way that gets in the way of that. As a result they stay built around touch controls, which leaves an opening for consoles to exploit.
Controller connection quality varies wildly by device model, so it won't catch on that way. Nothing changes until Apple or Google sets a standard reference model — that's also exactly why USB pass-through/dongle-type controllers have stubbornly survived.
＼Xperia 10 VIII pricing is here 🥳🙌／
✅6.1-inch display (OLED, up to 120Hz)
✅Snapdragon 6 Gen 3
✅RAM: 8GB
✅Storage: 128GB
✅Expandable storage: microSD support (up to 2TB)
✅Dual camera
50MP wide + 13MP ultra-wide
✅Body size: approx. 153×72×8.3mm
✅5,000mAh battery / 168g weight
✅Stereo speakers / Osaifu-Keitai (Japan's mobile wallet/NFC payment system)
✅Fingerprint sensor (side-mounted) / 3.5mm headphone jack
✅Water/dust resistance: IP65・IP68 / up to 4 OS updates / 6 years of security updates
✅Price: ¥104,400 (SoftBank version)
And this costs over ¥100,000…? (´・ω・｀)
That's about double the price of a Chinese phone with the same specs.
Classic 'behold our company's technology' smugness, they're just thrilled with themselves lol
It wouldn't turn out that way on the Wii's path, but that path already got rejected.
The Wii took a thorny road by abandoning the traditional approach entirely, and it fizzled out — but the Switch picked up everything it left behind and made it work. Still, you have to keep giving people a reason to upgrade, so 'never raise the specs' just isn't realistic.
I left out the subject — it was the users who rejected it, not Nintendo. The late-stage sales slump was brutal…
If they raise the specs, the price goes up too and users just walk away, no? Both the Switch 3 and the PS6 look like they're shaping up to be seriously expensive.
Who knows what'll happen. Even the current situation wasn't something anyone could've predicted a few years ago.
I guess it just means demand grew even bigger than expected.
That was Acer talking — a company that doesn't even make memory.
Background and Key Points of This Discussion
The DRAM and NAND semiconductor memory industry is dominated by a handful of players — Samsung, SK Hynix, Micron, and Kioxia (formerly Toshiba) — with capital investment running into the hundreds of billions to trillions of yen, which makes new entrants rare. There really have been cases in the past where a production race ended in a price collapse that pushed a manufacturer to the brink of financial crisis, and the “company that nearly collapsed from overproduction” mentioned in the thread is likely referring to this historical boom-bust cycle. Also, new fabs take several years from groundbreaking to full operation, so even if a decision to expand production were made today, it wouldn’t show up in the market until around 2029-2030. One more thing worth noting: this “tightening forecast” comes from Micron’s own CEO, not from an objective third-party supply-demand analysis. Since the remark also touches on the company’s own earnings outlook and pricing leverage, it’s no guarantee of future supply-demand conditions — and some in the thread floated the view that it’s really just a “warning shot to keep prices from dropping.”
※This article is excerpted and summarized from the 5ch (Hardware/Industry (Geha)) thread “Micron CEO: “Memory and storage supply-demand conditions will tighten further in FY27 and FY28, compared to FY26”.”
Leave a Reply