Memory card makers cut storage speeds as flash shortage bites
Demand for NAND memory from AI infrastructure is pushing the flash market into shortage. Memory card makers are already cutting the speed and capacity of their models to hold prices down in the mass-market segment. Storage prices have risen by 124% on average, and some products by 261%. Buyers will need to look closely at a card's speed class, not just its brand.
AI-processed from CNews AI; edited by Hamidun News
The market for memory cards and other flash storage is shifting against buyers: manufacturers are increasingly releasing slower, lower-capacity models to keep prices within reasonable limits. The cause is a shortage of NAND chips, made worse by growing demand from artificial intelligence infrastructure.
Why the shortage happened
AI is not directly "eating" microSD cards. The problem is different: the same production capacity and the same types of NAND memory are increasingly going into servers, data-center SSDs and enterprise storage systems. Amid the generative AI boom, suppliers find it more profitable to serve large infrastructure contracts than the cheap mass-market segment. As a result, the consumer market gets fewer chips, and each shipment costs more.
The effect is already visible in prices. According to the source article, storage prices rose by 124% on average, and some products went up by 261%. This means some models have effectively approached a price that until recently would have been considered absurd for everyday accessories such as memory cards for cameras, smartphones, dash cams and consoles. If the shortage persists, manufacturers will have to choose between further price increases and cutting specifications to keep an affordable shelf.
How the market is making cards cheaper
To avoid scaring buyers off for good, brands are changing the products themselves. Instead of fast, high-capacity cards, they increasingly offer models with more modest write and read speeds and less storage. For the mass market this looks like a barely noticeable adjustment of the lineup, but in essence the bar is being lowered. It is easier for a manufacturer to ship a card that formally stays affordable than to explain why a familiar model now costs two or three times as much.
- Lower write speed — worse for 4K video and burst shooting
- Smaller capacity — more frequent offloading and clearing of files
- Less endurance headroom — less convenient for cameras, drones and dash cams
- Harder choice — a familiar brand no longer guarantees the same performance class
For users this is especially sensitive in categories where the memory card is not a secondary accessory but a critical component. Action cameras, drones, mirrorless cameras, handheld consoles and car dash cams depend on stable write speeds. If the card can't keep up, the device may limit shooting modes, interrupt recording or handle its buffer noticeably worse. In other words, a formally cheaper product is not a better deal: the buyer pays less at the register but gets a device with more compromises.
The paradox is that specs are degrading at the same time as prices rise. Buyers are not just overpaying; they are overpaying for a weaker configuration. It is a familiar scenario in electronics: when the cost of a key component jumps, vendors first optimize the lineup, then rebuild SKUs for the new budget, and then what used to be the baseline becomes the norm. If pressure from AI infrastructure does not ease, slow cards that the market would simply have rejected a few years ago may become that "new normal."
What this means
For the ordinary buyer the message is simple: look not only at price and brand, but at the real specifications of each card — speed class, capacity, memory type and intended use. For the market as a whole, it is a signal that the AI boom now affects not only GPUs and servers but also the consumer electronics around us. Cheap storage will stay on sale, but the very word "cheap" will increasingly mean "stripped down in capabilities." This is especially true where the storage is used for recording video, games or work files.
- Nov 2, 2026The Verge reviewer builds a $50 Meta Muse Charm alternative with Claude Code
- Nov 2, 20268 Skills for video in Claude Code: from 244 pixel effects to rough-cut editing
- Nov 1, 2026Pika Labs releases open video model H3 Flex: 6 steps instead of 50
- Nov 1, 2026Google's StreamVibe video agent beats other models by 10.7% on response timing and quality
Key Questions Answered
Why the shortage happened?
AI is not directly "eating" microSD cards. The problem is different: the same production capacity and the same types of NAND memory are increasingly going into servers, data-center SSDs and enterprise storage systems. Amid the generative AI boom, suppliers find it more profitable to serve large infrastructure contracts than the cheap mass-market segment. As a result, the consumer market gets fewer chips, and each shipment costs more. The effect is already visible in prices. According to the source article, storage prices rose by 124% on average, and some products went up by
How the market is making cards cheaper?
To avoid scaring buyers off for good, brands are changing the products themselves. Instead of fast, high-capacity cards, they increasingly offer models with more modest write and read speeds and less storage. For the mass market this looks like a barely noticeable adjustment of the lineup, but in essence the bar is being lowered. It is easier for a manufacturer to ship a card that formally stays affordable than to explain why a familiar model now costs two or three times as much. Lower write speed — worse for 4K video and burst shooting Smaller capacity — more frequent offl
What this means?
For the ordinary buyer the message is simple: look not only at price and brand, but at the real specifications of each card — speed class, capacity, memory type and intended use. For the market as a whole, it is a signal that the AI boom now affects not only GPUs and servers but also the consumer electronics around us. Cheap storage will stay on sale, but the very word "cheap" will increasingly mean "stripped down in capabilities." This is especially true where the storage is used for recording video, games or work files.
Test Frontier LLMs & Autonomous Agents in One Workspace
Run Claude Opus 5.5, GPT-6, Gemini 3.1, and DeepSeek side-by-side in Nomad AI (nomadnet.ai) — unified model routing, deep research, and multimodal workflows under one subscription.
The AI world, distilled — once a week
Seven stories that actually mattered, hand-picked. No noise, no reposts, no press releases.
Done! Check your inbox for a confirmation.