16GB of RAM has been the default recommendation for a long time. Buy a mid-range laptop, spec up from 8GB, and you'd be fine. That advice held up well through the early 2020s, but in 2026 the picture is more complicated. Operating systems consume more memory than they did three years ago, browser tabs are heavier, and AI-assisted tools have added a new layer of background overhead that wasn't there before. Whether 16GB is enough depends entirely on what you're doing with the machine.
What 16GB actually gets you today
For most everyday office tasks, 16GB is still adequate. Light document work, email, video calls, and basic spreadsheets don't put serious pressure on memory. Windows 11 itself sits at around 3–4GB at idle, and macOS Sequoia tends to run a little leaner. That leaves you 12GB or so to split across your open applications.
The problem is tabs. A Chrome session with 15 open tabs can consume 4–6GB on its own. Add a Teams or Zoom call, a local PDF viewer, and a background sync tool, and you're brushing the ceiling. The machine doesn't crash. It starts paging to disk, which on modern NVMe SSDs is fast enough that most users don't immediately notice. But it does slow things down. The slowdown is subtle until it isn't.
For users who keep 20-plus tabs open, run multiple collaboration tools simultaneously, or work with large local files, 16GB is a bottleneck in 2026. It's workable, but you're leaving headroom on the table.
When 16GB is not enough
The clearest cases where 16GB falls short are developers, creatives, and anyone running local AI tooling.
Software developers routinely run a local dev environment, a database container, an IDE, a browser with documentation tabs, and a terminal session at the same time. That stack can easily consume 12–14GB before the first line of code is written. Docker containers deserve special attention: a modest multi-service local environment can chew through 6–8GB on its own. 32GB is now the realistic baseline for a developer machine, not an upgrade.
Video editors and motion graphics designers working with 4K footage in tools like DaVinci Resolve or Adobe After Effects need RAM for frame buffering. 16GB forces constant cache flushing, which interrupts real-time playback. The practical minimum for creative production work is 32GB, with 64GB worth considering for complex timelines.
Local AI tooling is the newest pressure point. Running a 7-billion-parameter language model locally using tools like Ollama requires at least 8GB of RAM just for the model weights, often more. Pair that with a browser and your normal workflow and 16GB is gone. 32GB gives you room to run a local model without constantly swapping.
The case where 8GB is still fine
It's worth saying plainly: 8GB is not dead. A Chromebook or a lightweight laptop used primarily for web browsing, light document editing, and streaming video can function perfectly well on 8GB in 2026. The operative word is "light." If the machine never sees a Docker container, never runs video editing software, and never has more than ten browser tabs open at once, 8GB is not a problem.
That said, 8GB should not be bought new for a business laptop fleet in 2026. The three-year refresh cycle means those machines will still be in service in 2029. Software overhead trends one way. Buy at the floor and you'll be under it before the warranty expires.
How much RAM you actually need: a practical breakdown
- Light office use (email, documents, video calls): 16GB is sufficient; 8GB is marginal and not recommended for new purchases.
- Power users and heavy multitaskers: 32GB is the right starting point. It handles deep browser sessions, multiple collaboration tools, and local file work without paging.
- Software developers: 32GB minimum; 64GB if the codebase is large or the Docker stack is complex.
- Video production and motion graphics: 32GB to 64GB depending on resolution and project complexity.
- Local AI and ML workloads: 32GB for running smaller local models alongside a normal workflow; 64GB or more for serious inference work.
- Data scientists and analysts: 64GB if datasets are loaded into memory; 32GB is acceptable for work that leans on cloud compute.
The soldered RAM problem
Buying decisions in 2026 are complicated by the fact that most thin-and-light laptops use soldered LPDDR RAM. You can't upgrade it after purchase. This matters enormously. A laptop with 16GB soldered memory bought today cannot become a 32GB machine in two years. The difference between LPDDR and DDR memory and what it means for upgradeability is one of the most important questions to answer before signing a fleet purchase order.
If the RAM is soldered and the budget only allows 16GB now, that's the machine's ceiling for life. For most knowledge workers that's a calculated risk. For developers or creative professionals, it's a decision that will cause pain within 18 months.
Business laptop procurement: what this means for fleet buyers
IT buyers specifying business laptop fleets face a different version of the question. The right amount of RAM depends on the user tier, not a single number for the whole organisation. Specifying 16GB across the board undersupplies developers and power users; specifying 32GB across the board oversupplies staff whose heaviest task is a weekly video call.
A tiered approach is more defensible: 16GB for administrative and light office roles, 32GB as the standard for knowledge workers and analysts, 64GB for engineering and creative teams. This kind of segmentation is exactly what a solid RAM procurement framework for business laptops should address before orders go out.
The other variable is the operating system. Apple Silicon Macs handle memory differently from Windows machines. Unified memory in an M-series chip doesn't behave the same way as discrete DRAM, and Apple's own guidance acknowledges that 16GB on an M4 MacBook Pro performs closer to 24GB on an x86 machine for many workloads. That doesn't mean 16GB is always enough on Apple hardware, but it is a real architectural distinction worth factoring in.
The honest answer
16GB of RAM is enough in 2026 for a specific user: someone doing conventional office work with moderate browser use, who is not running containers, local AI tools, or creative production software. For that profile, 16GB is fine and won't cause daily frustration.
For everyone else, 32GB is now the sensible default. The gap in cost between 16GB and 32GB has narrowed considerably over the past two years, particularly on Windows laptops, while the gap in real-world performance has widened. Buying 16GB to save $100–$150 on a device that costs $1,500 and will be used for three years is rarely the right calculation. If the RAM is soldered, there's no second chance.
The sweet spot for most Australian knowledge workers buying a laptop in 2026 is 32GB. It comfortably handles today's software load and leaves reasonable headroom for whatever the next two years of AI-assisted tooling adds to that figure.