RAM

RAM is a computer’s working memory. It holds everything the computer is currently occupied with: the running program, the open document, the twenty browser tabs. The processing chip can access this memory extremely quickly, much faster than the hard drive. In exchange, RAM is volatile: as soon as the power is gone, the contents are gone. That’s why you have to save files before shutting down the computer. The abbreviation stands for Random Access Memory — the chip can address any location directly, without having to work through everything before it.

Why eight gigabytes eventually isn’t enough anymore

A good comparison is a workbench. The hard drive is the cabinet holding all the tools you own. The RAM is the surface in front of you, where the things you currently need are laid out. A large surface doesn’t mean you work faster. But a surface that’s too small forces you to constantly put things away and pull them back out.

That’s exactly what happens when RAM is full. The operating system moves parts out to the hard drive, into what’s called a swap file. But the hard drive is orders of magnitude slower than RAM. The computer becomes sluggish, programs take seconds to switch. Anyone with many applications open at once notices immediately.

In the AI world, memory has become the decisive bottleneck. A large language model must sit completely in memory before it can respond. The same applies to the graphics cards in data centers, where their memory is called VRAM. A model that doesn’t fit either doesn’t run at all or runs with severe speed penalties. That’s why the price of memory chips has become a topic for investors.

Capacitors refreshed a thousand times per second

The common type is called DRAM, dynamic RAM. Each individual bit is stored as a tiny electrical charge in a capacitor, essentially a kind of mini-battery. Charged means one, uncharged means zero. However, this charge drains away on its own within milliseconds.

That’s why the memory chip permanently reads out its own contents and writes them freshly back in. This process is called refresh and happens thousands of times per second. It’s also the reason RAM constantly needs power and why the data disappears when it’s switched off. The structure of just one capacitor and one transistor per bit is very space-efficient for this purpose. That’s how billions of memory cells fit on a single chip.

Besides size, bandwidth matters too — that is, how much data flows between memory and processor per second. Modern standards are called DDR5 in PCs and HBM in AI accelerators. HBM stacks several memory chips on top of each other and connects them with a huge number of lines. This delivers a multiple of the bandwidth of normal modules — and costs accordingly.

RAM in product specs and in stock market news

When buying a laptop or phone, RAM always appears in the spec sheet, usually as a number in gigabytes. For smartphones, eight to twelve gigabytes is common, for notebooks sixteen or more. A common mistake is confusing RAM with hard drive storage. A phone with 256 gigabytes of storage only has a small fraction of that as working memory.

A second mistake: more RAM doesn’t automatically make a computer faster. It only helps if the existing memory is actually running short. If twelve gigabytes is enough for your usage, thirty-two brings practically no benefit.

RAM manufacturers regularly appear in business news. Samsung, SK Hynix, and Micron largely divide the market among themselves. Prices fluctuate strongly because factories take years to build while demand changes quickly. The AI data center boom has recently pushed demand for fast memory noticeably upward.