AI is making your gadget products more expensive — here’s why

Joseph Juanillas

Artificial intelligence tools continue to become faster and more efficient after years of development. 

But as companies scale data centers and build more advanced hardware, they need two components that most pieces of technology use to function — RAM sticks and SSDs.

Because of the very high demand AI tools and data centers need for these specific items, their prices have been three to five times more expensive than they were nine months ago, increasing prices of many tech products like phones, laptops, and consoles, up.

Although new infrastructure is built to cope with the large demands, it is still uncertain if the prices will return to normal.

Impact in the RAM and SSD market

Random Access Memory or RAM modules store data that are used for computations for a short while, similar to the human brain’s pre-frontal cortex that focuses on short-term memory.

Meanwhile, Solid State Drives or SSDs are responsible for storing huge amounts of files that AI models use to generate outputs based on user prompts, just like how the hippocampus in the brain’s temporal lobes are responsible for long-term memory.

These components allow AI companies to run very fast calculations and train models like ChatGPT, Gemini, and Claude.

However, in order to cater to more users, more chips and drives must be built in order to meet demands.

RAM and SSD manufacturers, like SK Hynix and Samsung, prioritize AI chip manufacturers like NVIDIA and AMD via partnership deals, limiting supply for consumer devices.

The shortages are estimated to last from 2027 up until 2028.

Curbing the shortage

Strategies are now being placed in order to curb the shortage, like SK Hynix that is building new buildings that cover more than 30 soccer fields in order to produce newer memory chips.

Micron, a SSD production company, is also starting to construct new facilities for producing storage drives, which is estimated to cost over $24 billion.

These buildings may have to wait until 2027 or early 2028 before it could start production.

Google, together with researchers from the University of California San Diego are now testing their so-called phone cluster computing, where they take the motherboards —  which hold the memory and storage chips of phones — of retired or damaged Pixel phones, aimed to build a low-cost data center that students can use for various AI tasks.

However, these responses may still be not enough for the demand AI data centers need in the future.

But for now, consumers may have to just stick with their current gadgets and hardware.