📰 AI News: Framework Hikes RAM And SSD Prices As AI Memory Crunch Bites

📝 TL;DR

Framework just raised its RAM and storage prices again and openly says more increases are likely in early 2026.The culprit is a global DRAM and NAND crunch driven by AI data centers, which is now spilling over into everyday laptops and desktops.

🧠 Overview

Framework published a detailed update explaining why memory for its laptops and desktop just got significantly more expensive, and why it may have to raise prices again soon. Behind the scenes, contract prices for DDR5 chips have exploded, with some indices showing 200 to 300 percent jumps since late summer, as manufacturers prioritize ultra profitable AI server memory over consumer parts.

This is one of the first times a PC maker has pulled back the curtain and said the quiet part out loud, memory is now a strategic bottleneck for the whole computing ecosystem.

📜 The Announcement

In mid December 2025, Framework increased the price of DDR5 memory on its DIY laptop configurations by around 50 percent, then updated the same post on December 17 and 24 to add further hikes and include SSDs. The company says it now prices RAM based on the weighted average cost it actually pays suppliers, landing at roughly 10 dollars per gigabyte for 8, 16, and 32 gigabyte modules and even higher for 48 gigabyte sticks.

That is still below what it sees on the open market and far below the per gigabyte markup some big brands charge, but Framework is blunt that supplier quotes keep rising and another increase within a month is very likely.

It also warns that storage pricing is now climbing in the same way and promises to keep adjusting both up and down as its own costs move.

⚙️ How It Works

• Weighted average pricing - Framework is constantly buying memory in small batches at different, often higher, prices, so it sets its RAM prices based on the weighted average cost of its current inventory rather than a flat margin.

• Clear per gigabyte math - The company has aligned its memory configurations at around 10 dollars per gigabyte for common sizes, with a premium on 48 gigabyte modules, so buyers can easily compare against third party memory kits.

• Bring your own RAM encouraged - Rather than locking you in, Framework is actively telling customers to buy barebones DIY systems and source cheaper compatible RAM and SSDs elsewhere if they can find better deals.

• Direct comparison tools - To make that easier, it is adding links from its configurator straight to price comparison and compatibility resources, so you can check market prices and tested modules before you order.

• Three pricing principles - Framework publicly commits to staying transparent, only raising prices to cover its own cost increases, and lowering prices again when the memory and storage markets cool off.

• Explaining the bottleneck - The blog walks through the bigger picture, AI racks that use tens of terabytes of HBM and LPDDR per cabinet, only three major DRAM die makers, and multi year lead times for new fabs, all of which mean server customers get priority while consumer PCs fight over what is left.

💡 Why This Matters

• The era of “RAM always gets cheaper” is over, for now - For years we have lived in a world where you expect more memory for less money each generation, this spike breaks that pattern and may last several years.

• AI is quietly taxing consumer hardware - Every new AI data center that goes live soaks up vast amounts of DRAM and HBM, which pushes up prices for the same commodity chips used in laptops, desktops, and consoles.

• Transparent pricing is becoming a competitive edge - Most manufacturers simply raise prices and blame “market conditions,” Framework is differentiating itself by showing its math and even telling you how to avoid paying more.

• DIY builders and power users feel it first - If you are planning a high RAM AI rig, dev machine, or editing workstation, memory has suddenly become one of the riskiest and most volatile line items in your budget.

• This is likely a multi year issue - Industry forecasts suggest meaningful relief only when new fabs come online around 2027 or later, and even then prices may stabilize above the historical norm instead of snapping back.

🏢 What This Means for Businesses

• Treat RAM and SSDs as strategic components - If you rely on AI workloads, video editing, data work, or VMs, start planning capacity and budgets around higher memory and storage costs rather than assuming bargain pricing will return soon.

• Be flexible about configurations - Consider buying systems with minimal or no pre installed memory and storage, then sourcing modules separately when you catch a good deal or can reuse parts from older machines.

• Prioritize efficiency, not just capacity - Look at ways to reduce memory pressure, for example lighter browsers, fewer background apps, smarter use of cloud resources, and tuning AI workflows so they are less memory hungry.

• Time your purchases deliberately - If you know a fleet refresh or a new AI workstation is coming, it may be smarter to pull plans forward or lock in pricing rather than hoping the market magically gets cheaper next quarter.

• Ask vendors for transparency - Framework has shown it is possible to explain price changes clearly, you can push other suppliers to share more about how they price memory and whether they will lower it again when costs fall.

🔚 The Bottom Line

Framework’s update is not just about one laptop brand, it is a clear signal that commodity memory and storage are now under the same kind of pressure we usually associate with GPUs. AI is soaking up the world’s DRAM and NAND, and everyone else is paying the premium.

For the AI Advantage community, the takeaway is simple, plan your hardware like a scarce resource, choose vendors who communicate honestly, and design your workflows so you get the most out of every gigabyte you pay for.

💬 Your Take

With RAM and SSD prices spiking, what is your next move, lock in the hardware you need now, stretch the life of your current machines, or rethink how much local compute you really need at all?

2 comments