Tootfinder

Many consumer electronics manufacturers
'will go bankrupt or exit product lines' -- by the end of 2026
due to the AI memory crisis,
Phison CEO reportedly says
https://www.

Many consumer electronics manufacturers 'will go bankrupt or exit product lines' by the end of 2026 due to the AI memory crisis, Phison CEO reportedly says
This could go on for another 10 years.

Global DRAM supply is likely to meet only 60% of demand through 2027; memory to hit ~40% of low-end smartphone manufacturing costs by mid-2026, up from 20% now (Nikkei Asia)
https://asia.nikkei.com/business/tech/semiconduct…

Memory shortage set to run until 2027 as chipmakers focus on AI
Samsung, SK Hynix and Micron invest in more output, but production hikes take time

heise | Memory Integrity Enforcement: Wie Apple künftig seine Hardware absichert
Eine neue Sicherheitstechnologie soll selbst komplexeste Angriffe abwehren – zunächst auf dem iPhone, später auch auf dem Mac. So funktioniert sie.

Is Memory Alpha rocking or what? ->
(This is a list of Jefferies tube appearances by series and episodes.)
Jefferies tube | Appearances
https://memory-alpha.fandom.com/wiki/Jefferies_tube#Appearances

🐀 Rats demonstrate ability to replay episodic memories in complex experimental settings
#psychology

Rats demonstrate ability to replay episodic memories in complex experimental settings
In a new study Indiana University researchers observed episodic memory in rats to a degree never documented before, suggesting that rats can serve as a model for complex cognitive processes often considered exclusively human. Unlike semantic memory, which involves isolated facts, episodic memory involves replaying events in the order and context in which they occurred.

I've got a patch that saves ~600MB peak memory on a long running memory hog. This would be great, if it weren't for the test hogging a total of over 5GB; if I can get it down to say 2.5GB then maybe it's useful.

I needed to build up our inventory of parts to build our network testing products.
So I discover:
1. The Chinese vendor who makes the motherboards has stopped shipping to the US.
2. The memory I need - pretty basic 16gb DDR 4 and 5 SODIMS - unavailable or more than quadrupled in price.
3. M.2 NVME SSDs - unavailable or much higher prices.
F**k those AI companies especially the one that bought all those memory wafers it can't use just to block its competition…

These collages rule…
https://theoo.dev/notes/19

from my link log —
The asymptotic cost of address translation on memory access time.
https://arxiv.org/abs/1212.0703
saved 2020-10-21 https://dotat.at/:…

The Cost of Address Translation
Modern computers are not random access machines (RAMs). They have a memory hierarchy, multiple cores, and virtual memory. In this paper, we address the computational cost of address translation in virtual memory. Starting point for our work is the observation that the analysis of some simple algorithms (random scan of an array, binary search, heapsort) in either the RAM model or the EM model (external memory model) does not correctly predict growth rates of actual running times. We propose the …

Memory and storage companies should be nationalized if they they don’t sell to consumers for fair prices

SK hynix says it has begun mass production of the 192GB SOCAMM2, a next-gen LPDDR5X low-power DRAM module designed particularly for Nvidia's Vera Rubin (The Korea Herald)
https://www.koreaherald.com/article/10720775

SK hynix begins mass production of next-gen AI server memory designed for Nvidia's Vera Rubin platform - The Korea Herald
SK hynix Inc. said Monday it has begun mass production of a next-generation memory module designed for artificial intelligence servers, as it seeks to strengthe

Linear Memory (2024)
A prototype version of my piece Linear Memory running inside a https://thi.ng/genart-api sandbox on a Motorola phone, connected to a custom early prototype Layer square screen...
The animation consists of dozens of layers of semi-random multi-scale pixel patterns, slowly scrolling in…

@… @… @… I would be honored that it lived as a memory for so long!!

Gensai Hasegawa – Memory Rebellion
https://mimirecords.bandcamp.com/album/memory-rebellion
https://www.…

After a lot of other poking and prodding at other stuff failed, the PAM edge detector got a 30% speedup and the sin(x)/x upsampling filter is about 4x faster just by restructuring some stuff to change memory access patterns.
As a reminder: a GPU is a device for turning compute bound problems into memory bandwidth bound problems.

After today's demonstation I'd gladly purge the memory card and never look at those pics again, but some sily off-topic shots happened and that's the only thing that prevents me from doing so.

This is my retirement plan. I'm rich!

[on the dog meme]
Security issues back in the day: by executing a specially prepared code you can place private data of some process in CPU's cache, and then predict its contents using a timing attack 💪.
Security issues today: I opened a big file and ran out of memory 😭.
"Back in my day, we solved these class of issues using a belt!"
#shitposting

"I have only to venture into the streets of my own neighborhood, the West Village, to see such Humean casualties by the thousand: younger people, for the most part, who have grown up in our social-media era, have no personal memory of how things were before, and no immunity to the seductions of digital life. What we are seeing—and bringing on ourselves—resembles a neurological catastrophe on a gigantic scale."
The Machine Stops
The neurologist on steam engines, smartphones, and fearing the future.
By Oliver Sacks
https://archive.ph/8FpeK

Sources: Google is in talks with Marvell Technology to develop a memory processing unit that works alongside TPUs, and a new TPU for running AI models (Qianer Liu/The Information)
https://www.theinformation.com/articles/google-talks-marvell-build-new…

Google in Talks With Marvell to Build New AI Chips for Inference
Google is in talks with Marvell Technology to develop two new chips aimed at running AI models more efficiently, according to two people with direct knowledge of the discussions. One is a memory processing unit designed to work alongside Google’s tensor processing unit. The other is a new TPU ...

@… Frankly the only reason I’m considering replacing my M1 Mac Mini with a MBP is the form factor and the FOMO.
I’ve not once run into “ugh, I wish this computer was faster” since ordering it on launch (I *did* get 16 GB of memory FWIW).

How psychedelics push your brain to dream while awake https://medicalxpress.com/news/2026-03-psychedelics-brain.html "psychedelics make the brain more likely to 'see' images from memory rather than what's actually in front of it"; mental imagery

How psychedelics push your brain to dream while awake: New study
A new study in mice suggests psychedelics make the brain more likely to "see" images from memory rather than what's actually in front of it.

I don't really care about memory tbh. But these funky quickshells use around up to 1.4Gb on clean startup. That's Gnome territory. No one complaining?

😈 Enzymes work as 'Maxwell's demon' by using memory stored as motion
#enzymes

Enzymes work as 'Maxwell's demon' by using memory stored as motion
Living cells are sustained by countless chemical reactions that must be carefully regulated to maintain internal order and function. Enzymes play a central role in this process, accelerating reactions that would otherwise proceed too slowly to support life.

A post from the archive 📫:
How much memory does this object use?
#diagnostics …

China's smartphone shipments declined 4% YoY in Q1 amid memory shortages; Huawei's shipments grew 2% YoY for a 20% market share, iPhone grew 20% for a 19% share (Ivan Lam/Counterpoint Research)
https://counterpointresearch.com/en/insights/china-smartphone…

from my link log —
The 2019 Linux storage, filesystem, and memory-management summit.
https://lwn.net/Articles/lsfmm2019/
saved 2019-05-17 https://do…

Oh I do love how you have your Kitten¹ app’s database in memory – thanks to Kitten’s built-in JavaScript Database (JSDB)² – and accessible via the Kitten Shell (REPL)³ during dev.
(If I do say so myself.) ;)
:kitten:💕
¹ https://kitten.small-web.org
²

The outcome of a bunch of shader tuning last night: the upsample filter (4x sin(x/x) from 20M to 80M points in this test) went from 6.55 ms to 1.5 ms.
Original: 8% of peak DRAM read BW, 31% write, 14% L2$ hit rate.
New (just changed memory access patterns to be more coalesce/cache friendly): 9% read, 37% write, 73% L2$ hit
A similar memory ordering optimization cut the PAM edge detector from about 14 to 10 ms but my SM occupancy is still crap (around 12% of warp slots used)…

Analysis: DRAM and NAND memory prices have jumped 600% over the past year for routers and set-top boxes, hitting telcos targeting aggressive broadband rollouts (Counterpoint Research)
https://counterpointresearch.com/en/in

🇺🇦 Auf radioeins läuft...
waterbaby:
🎵 Memory be a blade
#NowPlaying #waterbaby
https://waterbabyonline.bandcamp.com/album/memory-be-a-blade
https://open.spotify.com/track/6kM3Uqz3MRTVGTzu93EZIA

Volodymyr Zelenskyy honours disqualified skeleton racer with order of freedom | Winter Olympics 2026 | The Guardian
https://www.theguardian.com/sport/2026/feb/13/cas-dismisses-ukrainians-appeal-over-helmet-of-memory-disqualification

from my link log —
Running modern code on a 1960s UNIVAC computer via a RISC V emulator.
https://farlow.dev/2026/04/17/running-a-minecraft-server-and-more-on-a-1960s-univac-computer
saved 2026-04-18

tuning(7) begins:
"The swap partition should typically be approximately 2x the size of main memory for systems with less than 4GB of RAM, or approximately equal to the size of main memory if you have more. "
I can't believe that 64 GB swap should be a norm for a system with 64 GB RAM.
<

We issued an RFP for the same hardware that we ordered in September 2025. The quoted price is 3.5 times higher than the previous price for the exact same hardware. In addition, the offer validity is now limited to only three days.
#memory #aibubble

🇺🇦 #NowPlaying on #BBC6Music's #DreamTime
Total Blue:
🎵 The Path
#TotalBlue
https://music-from-memory.bandcamp.com/track/the-path
https://open.spotify.com/track/3AKpwoEOw7UE4WbckbMlFJ

It appears that you can resolve most of GrapheneOS's severe performance limitations and tendency to aggressively kill background apps (after like 15 seconds!) by globally disabling "secure app spawning" and per-app disabling the "hardened memory allocator." At that point, however, you're pretty much left with a gimped AOSP that's missing a bunch of core usability functionality yeeted in the name of security, and you might as well install LineageOS. Or maybe Ca…

"What Every Programmer Should Know About Memory"
https://www.akkadia.org/drepper/cpumemory.pdf

@… uptime only exists if you have long term memory

Valve says the Steam Deck OLED may be "intermittently" out of stock due to memory and storage shortages; it has been out of stock in the US for a few days (Jay Peters/The Verge)
https://www.theverge.com/games/879845/valve-steam-dec…

Valve’s Steam Deck OLED will be ‘intermittently’ out of stock because of the RAM crisis
Valve’s Steam Deck OLED website say that it may be out of stock “intermittently in some regions due to memory and storage shortages.” 

Between Security and Memory: Ukraine in the Political Rhetoric of Polish President Karol Nawrocki: https://benborges.xyz/2026/03/02/between-security-and-memory-ukraine.html

from my link log —
memstop: delay process start until memory is available.
https://github.com/surban/memstop
saved 2026-04-14 https://dotat.at/:/5DJRN…

GitHub - surban/memstop: Memstop monitors available memory and waits until a configurable percentage of memory becomes available before allowing a program to start.
Memstop monitors available memory and waits until a configurable percentage of memory becomes available before allowing a program to start. - surban/memstop

Memory chipmakers around the world shed close to $100bn in market value last week, after new research suggested AI systems may require far less hardware than previously expected.
https://www.computing.co.uk/news/2026/memory-chip-stocks-slide-af…

Memory chip stocks slide as AI efficiency breakthrough rattles investors
Memory chipmakers around the world shed close to $100bn in market value last week, after new research suggested AI systems may require far less hardware ...

Was haben Influenza, Influenzer und Künstliche Inteligenzia gemeinsam? (2004)
#sociolgy

Universal behavior in a generalized model of contagion
Models of contagion arise broadly both in the biological and social sciences, with applications ranging from the transmission of infectious diseases to the diffusion of innovations and the spread of cultural fads. In this Letter, we introduce a general model of contagion which, by explicitly incorporating memory of past exposures to, for example, an infectious agent, rumor, or new product, includes the main features of existing contagion models and interpolates between them. We obtain exact sol…

just dislodged a memory of attending the original 2002 auction of jerry garcia's guitars, at studio 54 for some reason, & guitar maker doug irwin (who garcia willed them to) literally cackling/laughing from the back of the room when tiger went for almost a million.

Equilibria: Fair Multi-Tenant CXL Memory Tiering At Scale
Kaiyang Zhao, Neha Gholkar, Hasan Maruf, Abhishek Dhanotia, Johannes Weiner, Gregory Price, Ning Sun, Bhavya Dwivedi, Stuart Clark, Dimitrios Skarlatos
https://arxiv.org/abs/2602.08800 https://arxiv.org/pdf/2602.08800 https://arxiv.org/html/2602.08800
arXiv:2602.08800v1 Announce Type: new
Abstract: Memory dominates datacenter system cost and power. Memory expansion via Compute Express Link (CXL) is an effective way to provide additional memory at lower cost and power, but its effective use requires software-level tiering for hyperscaler workloads. Existing tiering solutions, including current Linux support, face fundamental limitations in production deployments. First, they lack multi-tenancy support, failing to handle stacked homogeneous or heterogeneous workloads. Second, limited control-plane flexibility leads to fairness violations and performance variability. Finally, insufficient observability prevents operators from diagnosing performance pathologies at scale.
We present Equilibria, an OS framework enabling fair, multi-tenant CXL tiering at datacenter scale. Equilibria provides per-container controls for memory fair-share allocation and fine-grained observability of tiered-memory usage and operations. It further enforces flexible, user-specified fairness policies through regulated promotion and demotion, and mitigates noisy-neighbor interference by suppressing thrashing.
Evaluated in a large hyperscaler fleet using production workloads and benchmarks, Equilibria helps workloads meet service level objectives (SLOs) while avoiding performance interference. It improves performance over the state-of-the-art Linux solution, TPP, by up to 52% for production workloads and 1.7x for benchmarks. All Equilibria patches have been released to the Linux community.
toXiv_bot_toot

🌭 Fiber-free processed foods hit emotional memory fast, especially in older brains
#food

Fiber-free processed foods hit emotional memory fast, especially in older brains
Past studies in animals have shown that a highly processed diet is linked to memory problems and inflammation in the aged brain—and the effect can happen fast, after just three days of poor eating.

> Moving to a version of IOS XE that doesn’t write the log file is the obvious fix, but Cisco warns it’s not that simple because if your Wi-Fi box has filled its flash with junk, it may not have enough internal storage space to hold the updated OS.
https://www.theregister.com/2026/04/17/cis

Cisco Wi-Fi boxes are filling their disks with 5MB of undeletable data every day
: Fix for critical flaw is an OS update you may not be able to make because the junk data uses all memory

The only advantage of the AI bubble is that when someone brings the "memory is cheap" cannon again, you can ask: "excuse me, but have you *seen* the prices?!"
Also, there should be a law that whenever you say "memory is cheap", you are obliged to buy and install the memory for all the participants. To prevent from abuse, let's limit it to doubling the memory (both RAM and disks) in all PCs of every participant.

🇺🇦 #NowPlaying on KEXP's #SundaySoul
Donny Hathaway:
🎵 Memory of our Love
#DonnyHathaway
https://herrickg.bandcamp.com/track/donny-hathaway-memory-of-our-love-extended-version
https://open.spotify.com/track/0oCUs1tUJZTs7FwUbjt4eA

Memory prices are skyrocketing because of the AI delusion but I have taken precautions and stocked up

Sources: the glowing "26" in Apple's WWDC invite is teasing a revamped Siri, memory shortages may push Mac Studio and touch MacBook Pro launches by a few months (Mark Gurman/Bloomberg)
https://www.bloomberg.com/news/ne…

SK Group Chair Chey Tae-won says memory chip makers are expanding capacity but are unlikely to match demand until around 2030; basic wafer supply lags by 20% (Yoolim Lee/Bloomberg)
https://www.bloomberg.com/news/articles/2026-03-17…

Samsung says it is considering a shift toward multi-year contracts for memory chips that may help stabilize supply and ease concerns about a shortage (Yoolim Lee/Bloomberg)
https://www.bloomberg.com/news/articles/2026-03-18/…

Donald Trump, speaking at a press conference at his golf club in Doral, Florida,
because of course that is where you announce big decisions,
today said that the war against Iran will end “very soon.”
Talking to the New York Post about oil prices, Trump said there was no cause for worry,
because “I have a plan for everything. You’ll be very happy.”
He also said he doesn’t remember ever meeting Jeffrey Epstein, his poll numbers are higher than anyone has ever see…

Memory hole, meet Iran war.
Donald Trump, speaking at a press conference at his golf club in Doral, Florida, because of course that is where you announce big decisions, today said that the war against Iran will end “very soon.”

🇺🇦 #NowPlaying on #BBC6Music's #FreakZonePlaylist
Sega Bodega & Lucinda Chua:
🎵 The World Around Me Is Working Against Me
#SegaBodega #LucindaChua
https://segabodega.bandcamp.com/track/the-world-around-me-is-working-against-my-memory-with-lucinda-chua
https://open.spotify.com/track/0ENyzqpIERirFdIBu463Ij

TechInsights: chip manufacturing emissions will climb by ~33% to 247M metric tons of carbon dioxide equivalent by 2030, roughly equivalent to Algeria in 2024 (Aaron Clark/Bloomberg)
https://www.bloomberg.com/news/articles/2026-…

DOUBLE STANDARDS:Why did the IOC PUNISH Heraskevych for the HELMET OF MEMORY of the FALLEN athletes?: https://benborges.xyz/2026/02/14/double-standardswhy-did-the-ioc.html

adventures in attic spelunking: confirming blurry childhood memory that our first family VCR did, in fact, resemble a mondrian painting.

Samsung and AMD sign a preliminary deal for Samsung to supply its next-gen HBM4 for AMD's MI455X accelerators, used in data centers, and DDR5 for AMD's Helios (Yoolim Lee/Bloomberg)
https://www.bloomberg.com/news/articles/20

Micron reports Q2 revenue up 196% YoY to $23.9B, vs. $19.7B est., expects 2026 capex to exceed $25B, vs. $22.4B est., and forecasts Q3 revenue above estimates (Dina Bass/Bloomberg)
https://www.bloomberg.com/news/articles/202…

🇺🇦 #NowPlaying on KEXP's #AfternoonShow
Peter Cat Recording Co.:
🎵 Memory Box
#PeterCatRecordingCo
https://mesiebelmont.bandcamp.com/track/peter-cat-recording-co-memory-box-belmonts-disco-mix
https://open.spotify.com/track/6C5xm2roWdAIda9WJmu1jG

#Aphantasia and visual working memory: No direct evidence of impaired visual working memory in aphantasics, either in behavioral performance or the accuracy of a multivoxel pattern classifier https://www.sciencedi…

I love this premise:
"It has been 62 years since the last Federation contact with the planet."
It sounds real, gloomy, and plausible in the grander scale of the Star Trek universe compared to the usual warping about the Alpha Quadrant.
It's from TNG's 1x14 (Angel One) Memory Alpha:
https://

Angel One (episode)
While Riker leads an away team to a female-dominated planet, a mysterious virus spreads among the Enterprise crew. The USS Enterprise-D has discovered the wreckage of the Odin, a freighter that disappeared seven years earlier. It was disabled by a collision with an asteroid, and there are no life signs on board, but three escape pods have been discovered missing. The Enterprise travels to Angel I, the nearest class M planet, to search for the freighter's survivors. Angel I is a...

MEION – Drowned Memory
https://mahorka.bandcamp.com/album/drowned-memory
https://www.clongc…

@… it’s important to know the difference between the federation type 6 and type 7 shuttlecraft https://memory-beta.fandom.com/wiki/Fe

from my link log —
Linux internals: how /proc/self/mem writes to unwritable memory.
https://offlinemark.com/an-obscure-quirk-of-proc/
saved 2026-03-09

Linux Internals: How /proc/self/mem writes to unwritable memory - offlinemark
Introduction An obscure quirk of the /proc/*/mem pseudofile is its “punch through” semantics. Writes performed through this file will succeed even if the destination virtual memory is marked unwritable. In fact, this behavior is intentional and actively used by projects such as the Julia JIT compiler and rr debugger. This behavior raises some questions: Is […]

Cisco stock falls 12% as rising memory prices put pressure on the networking company's margins; it is the stock's worst day since 2022 (Lola Murti/CNBC)
https://www.cnbc.com/2026/02/12/cisco-stock-has-worst-day-since-2022-a…

Cisco stock has worst day since 2022 as memory prices pressure margins
Cisco shares plunged 12% Thursday after the company issued lukewarm guidance as memory prices weigh on profit margins

😎 Memory justifications remain surprisingly stable even as memories fade over time, study shows
https://medicalxpress.com/news/2026-01-memory-justifications-stable-memories.html

Memory justifications remain surprisingly stable even as memories fade over time, study shows
While memories may fade with time, the explanations people give for why they remember an event remain surprisingly stable and reliable, according to a new study by Ben-Gurion University of the Negev.

Meta is increasing the price of the Quest 3 by $100 to $599.99 and both Quest 3S models by $50 to $349.99 for 128GB and $449.99 for 256GB, starting April 19 (Jay Peters/The Verge)
https://www.theverge.com/tech/912921/meta-quest-3-3s-vr-price-hike-ram-me…

Meta blames RAM shortage for $100 Quest 3 price hike
Beginning April 19th, Meta’s 128GB Quest 3S will cost $349.99, the 256GB Quest 3S will cost $449.99, and the Quest 3 will cost $599.99.

Hypothesis: Every sufficiently large, complex, or performance sensitive application will eventually end up with a custom memory pool or allocator of some sort.

from my link log —
Mesh: a compacting memory allocator for C/C .
https://github.com/plasma-umass/Mesh
saved 2020-08-22 https://dotat.at/:/NQV9V.ht…

OpenAI updates its Codex desktop app with features like computer control, an in-app browser, image generation, automation memory, plugin support, and more (David Gewirtz/ZDNET)
https://www.zdnet.com/article/openai-codex-desktop-update/

OpenAI's Codex Desktop can run your computer now - and has its own browser
From coding tool to productivity powerhouse, Codex Desktop adds computer control, automation memory, and plugin support. But can it replace traditional software?

Sources: Sony is considering pushing back the debut of its next PlayStation console to 2028 or even 2029, as the AI boom fuels huge memory chip demand (Bloomberg)
https://www.bloomberg.com/news/articles/2026-02-15/rampant-ai-de…

Micron, which has undertaken a $200B US expansion, says it can currently meet about 50%-66% of demand for some key customers, as AI drives memory chip demand (Robbie Whelan/Wall Street Journal)
https://www.wsj.com/tech/micro…

Applied Materials partners with Micron and SK Hynix to develop next-gen memory chips for AI and HPC at its new EPIC center, part of a planned $5B R&D investment (Jaspreet Singh/Reuters)
https://www.reuters.com/technology/applied

Applied Materials forges partnerships with Micron and SK Hynix for AI memory chips
Applied Materials said on Tuesday it has partnered with memory chip companies Micron ​Technology and SK Hynix to develop next-generation chips that ‌are crucial for artificial intelligence and high-performance computing.

Analysis: surging memory and CPU costs could drive the retail price of a mainstream $900 notebook up by nearly 40% in 2026 to maintain current margin structures (TrendForce)
https://www.trendforce.com/presscenter/news/20260310-12959.html

Rising Memory and CPU Prices Could Push Mainstream Notebook Prices Up by Nearly 40%, Says TrendForce
For a mainstream notebook with an MSRP of US$900, surging memory prices alone could raise retail prices by more than 30%; combined with CPU price increases, the total price hike could approach 40%. If memory and CPU prices rise simultaneously, their combined share of notebook BOM costs could climb to 58%. CPU supply volatility is also emerging, with shortages beginning to affect entry-level platforms across multiple brands.

Global smartphone shipments fell 4.1% YoY in Q1, the first decline since 2023, amid a memory chip crunch; Samsung's shipments grew 3.6% and Apple's grew 3.3% (IDC)
https://www.idc.com/resource-center/press-releases/1q26-mpt-top5/

from my link log —
The state of Rust memory allocators in 2026.
https://cetra3.github.io/blog/state-of-allocators-2026/
saved 2026-03-12 https…

SMIC CEO Zhao Haijun says the industry is "a bit panicked" about memory shortages, and new supply may arrive in nine months; memory prices are up 80% in 2026 (Sherry Qin/Wall Street Journal)
https://www.wsj.com/tech/s…

Nvidia unveils a server rack with 256 Vera CPUs, with each CPU featuring 88 custom Olympus cores and LPDDR5X memory for up to 1.2 TB/s of bandwidth (Tobias Mann/The Register)
https://www.theregister.com/2026/03/16/nvidia_vera_cpu_rack/

Nvidia powers further into the CPU market with new rack systems packing 256 Vera processors
GTC: The cubicals of the agentic AI age are cores

Apple led global Q1 smartphone shipments for the first time with a 21% market share in Q1 2026; overall smartphone shipments fell due to memory chip shortages (Counterpoint Research)
https://counterpointresearch.com/en/insights/global-smartphone-shipments-q1-…

Sources: HP, Dell, Acer, and Asus are considering sourcing memory chips from Chinese chipmakers; HP and Dell are currently qualifying DRAM products from CXMT (Nikkei Asia)
https://asia.nikkei.com/spotlight/supply-chain/h…

HP, Dell, Acer and Asus mull using Chinese memory chips amid supply crunch
Shortage a golden opportunity for China's memory chipmakers and electronics manufacturers

China's smartphone market is seeing one of its widest price hikes as memory chip crunch deepens; Meizu announces it will suspend developing new mobile hardware (Nikkei Asia)
https://asia.nikkei.com/business/electronics…

Steep price hikes hit China's smartphone market as memory crunch deepens
Apple's new budget iPhone adds to pressure on Chinese brands

Anthropic brings Claude's memory feature to free users, after launching it for paid users in 2025 (Igor Bonifacic/Engadget)
https://www.engadget.com/ai/anthropic-brings-memory-to-claudes-free-plan-220729070.html

Anthropic brings memory to Claude's free plan
Anthropic is bringing another paid feature to Claude's free tier.

Sources: Nvidia delays the release of its incremental gaming GPU upgrade, codenamed Kicker, marking the first year in three decades without a new GPU for gaming (The Information)
https://www.theinformation.com/articles/nvidia-delay-new-gaming-chi…

Nvidia to Delay New Gaming Chip Due to Memory Chip Shortage
Nvidia won’t release a new graphics chip for gamers this year due to a deepening global shortage of memory chips, prompted by the AI boom, according to two people with direct knowledge of the matter. It would be the first year in three decades that Nvidia hasn’t released a new graphics ...

Sony Japan temporarily suspends fulfillment of orders for nearly all of its CFexpress and SD memory card product lines due to solid state memory shortages (Jaron Schneider/PetaPixel)
https://petapixel.com/2026/03/27/sony-shuts-down…

Sony Shuts Down Nearly Its Entire Memory Card Business Due to SSD Shortage
Sony has announced that it is suspending fulfillment of all orders for nearly its entire SD and CFexpress memory card product lines.

Lenovo reports Q3 revenue up 18% YoY to $22.2B on strong PC sales ahead of expected memory price hikes and rising AI server demand; net income fell 21% to $546M (Bloomberg)
https://www.bloomberg.com/news/articles/2026-02-12/l…

An interview with SemiAnalysis CEO Dylan Patel on logic, memory, and power bottlenecks in scaling AI compute, Nvidia securing TSMC N3 allocation early, and more (Dwarkesh Patel/Dwarkesh Podcast)
https://www.dwarkesh.com/p/dylan-patel

TSMC's N3 logic wafer capacity has become one of the AI industry's biggest constraints, which could push customers to explore greater foundry diversification (SemiAnalysis)
https://newsletter.semianalysis.com/p/the-great-ai-silicon-shortage

The Great AI Silicon Shortage
TSMC N3 Wafer Shortages, Memory Constraints, Datacenter Bottlenecks, Supply Chain Wars Winner

Analysis: despite investing $150B in 10 years, China is expected to produce 2% of global AI chips in 2026 and 70x less memory storage than foreign chipmakers (Meaghan Tobin/New York Times)
https://www.nytimes.com/2026/02/14/busin…

Samsung and SK Hynix executives say they expect global memory supply shortages will persist until 2027 due to AI demand, after the companies reported earnings (Kim Jaewon/Nikkei Asia)
https://asia.nikkei.com/business/tech/

Samsung and SK Hynix warn memory chip crunch to continue until 2027
South Korean makers acknowledge they can't keep up with 'extreme' AI-driven demand

Samsung reports Q4 revenue up 24% YoY to ~$65.6B, vs. ~$65.28B est., and operating profit up 200% to ~$14.06B, driven by memory chip price surge and HBM demand (Dylan Butts/CNBC)
https://www.cnbc.com/2026/01/29/samsung-record-q4-2025-profit-…

Samsung profit triples, beating estimates as AI chip demand fuels memory shortage
Samsung Electronics reported an over three-fold surge in fourth-quarter profits, hitting a new record and beating analysts' estimates.

Anthropic launches a tool to bring a user's preferences and context from other AI platforms to Claude via one copy-paste command, available on all paid plans (Claude)
https://claude.com/import-memory

Switch to Claude without starting over | Claude
Transfer your preferences, projects, and context from other AI providers into Claude. Switch without losing what makes your AI useful.

Micron plans to invest $24B over the next decade to build a 700,000 square foot memory chip facility in Singapore, with wafer output starting in H2 2028 (Reuters)
https://asia.nikkei.com/business/tech/semiconductors/micron-plans-24b…

Micron plans $24bn memory chipmaking plant in Singapore
US company says new wafer fabrication facility to meet growing demand

Microsoft raises prices for Surface PCs, with Laptop 7 and Pro 11 now $500 more expensive than at their 2024 launch, citing higher memory and component costs (Zac Bowden/Windows Central)
https://www.windowscentral.com/hardware/s…

Microsoft reveals major price increases for all Surface PCs as RAM crisis continues: Flagships now $500 more expensive than at launch
Microsoft is raising prices on all its current Surface PC offerings, with the midrange devices now starting at above $1,000, and flagships starting at $1,500.

Sandisk and Kioxia announce a five-year extension of their JV manufacturing contract from 2029 to 2034; the new deal involves SanDisk paying Kioxia $1.165B (Ryo Mukano/Nikkei Asia)
https://asia.nikkei.com/business/tech/sem…

Japan's Kioxia extends memory chip JV with Sandisk, receiving $1bn
US partner's access to Yokkaichi plant to last 5 more years through 2034

Samsung says it sent the first commercial HBM4 shipments to customers, as it seeks to supply Nvidia and compete with memory rivals like SK Hynix and Micron (Yoolim Lee/Bloomberg)
https://www.bloomberg.com/news/articles/20

SK Hynix reports Q4 revenue up 66% YoY to $23B, above ~$22.4B est., and operating profit up 137% YoY to $13.3B, both company records, driven by AI HBM demand (Dylan Butts/CNBC)
https://www.cnbc.com/2026/01/28/sk-hynix-smashes-…

SK Hynix smashes earnings estimates as AI memory demand drives record profit
South Korea's SK Hynix on Wednesday posted record quarterly revenue and profit, to beat estimates.