The baseline has shifted from 8 to 16 GB, but the right answer depends on the role, not on “more is better.” This guide is for anyone who needs to equip someone next week and needs to get it right the first time.
Up until a year and a half ago, the configuration most frequently purchased by the companies we work with was 8 GB of RAM. Today, with AI integrated into operating systems and everyday tools, that amount falls short.
But AI doesn't explain everything. The shift didn't happen because work became more technical, but because the way we work changed: everything happens at the same time, and what runs in parallel lives in memory.
The question always arises when it's time to equip a new team member: “How many GB of RAM do you recommend?” Below, we break down how much RAM each role needs based on their equipment usage, what to consider before buying, and what options exist for computers that have fallen behind.
What is RAM and why does it matter more today than before?
RAM is a computer's immediate workspace: that's where the operating system, open applications, and data currently in use are loaded. The more RAM a machine has, the more tasks it can handle simultaneously without slowing down. Unlike storage, RAM clears when the computer is turned off. It doesn't save files: it holds what is in use.
For years, 8 GB was enough because work was more sequential: you opened your email, then a spreadsheet, then the browser. Today, a typical workday is simultaneous. There's a video call with screen sharing, a browser with twenty tabs open, a messaging tool, a CRM or project manager, and an AI assistant running on the side. Each of these things consumes memory at the same time, and when memory runs out, the computer starts using the disk as backup. That's when the slowdown hits.
Why is 8 GB of RAM no longer enough for work?
An 8 GB laptop today boots up with a large portion of its memory already occupied before you even open your first work tool. Added to this are the multitasking demands of remote work and a clear signal from the industry: the two most widely used enterprise operating system manufacturers set 16 GB as the minimum for their AI features.
The operating system takes a large share
Windows 11 technically runs on 4 GB, but that number says little about real-world usage. In its review of the 2026 Dell XPS 13 with 8 GB, Engadget found that 80% of the memory was already committed to background processes as soon as Windows booted up. With the browser open and several tabs running, usage climbed to 95%, and scrolling or switching windows caused minor stuttering. The 16 GB version of the same machine performed about 30% better in PCMark 10, and its battery lasted 1 hour and 44 minutes longer in the same test. EngadgetEngadget
Manufacturers set 16 GB as the baseline for AI
For a computer to be considered a Copilot+ PC (the category with AI features built into Windows), Microsoft requires a processor with a neural processing unit capable of over 40 TOPS and 16 GB of DDR5 or LPDDR5 RAM, according to its official requirements page. microsoft
Apple made an equivalent move. In October 2024, when introducing its new MacBook Pros, it announced that the M2 and M3 MacBook Airs were doubling their base memory to 16 GB without changing the entry price, linking the change to Apple Intelligence. Since then, the Mac lineup starts at 16 GB, with one noteworthy exception discussed below. idownloadblog
AI assistants add overhead even when running in the cloud
Tools like ChatGPT, Copilot, Gemini, or Claude process most of their workload on external servers, so they don’t "run" locally on the laptop. But they do take up an additional tab or application that remains open alongside everything else. The AI that genuinely demands local memory is the kind that runs on the device itself: built-in operating system features and the local models that some technical users run on their own computers.

How much RAM does each profile need?
The right amount of RAM depends on how many tools each role uses at the same time and how demanding they are. For most business profiles, 16 GB is the baseline. Analysts and managers with a heavy concurrent workload benefit from 24 GB. Design and development require 32 GB or more.
The table summarizes the criteria we use for equipment intended to last several years without falling short. It is not a universal rule: it is the starting point for establishing a standard by role.
Profile | Typical use | Minimum | Recommended |
|---|---|---|---|
Sales, customer success, HR, administration | Browser with many tabs, CRM, messaging, video calls, cloud AI assistant | 16 GB | 16 GB |
Managers, analysts, finance, product | The above + heavy spreadsheets, BI tools, frequent screen sharing | 16 GB | 24 GB |
Design | Figma, Adobe suite, prototypes, image editing; video in some cases | 24 GB | 32 GB (more if editing video) |
Development and data | IDE, containers, virtual machines, local databases, local AI models | 32 GB | 32 to 64 GB depending on the stack |
Two takeaways from the table. First: the difference between “minimum” and “recommended” appears precisely in roles where the workload grows over time. An analyst who handles medium-sized spreadsheets today will likely handle more data and more tools in two years. Second: in development, the number heavily depends on the stack. A frontend developer working primarily in the browser is not the same as someone running multiple containers or testing language models locally on their machine. For those profiles, it is best to check with the tech lead before requesting a quote.

Can a laptop's RAM be upgraded after purchase?
On Macs with Apple silicon, no: the memory is integrated into the chip, and the capacity you choose at purchase is what the machine will have for its entire lifespan. The same is true for many ultralight Windows laptops, because the memory is soldered or integrated into the processor. That's why RAM is the hardest component to rectify if you buy too little.
This is the least discussed factor when buying, yet the one with the biggest consequences. Apple explicitly states in its support documentation: memory is integrated into the M-chip package and cannot be upgraded, with the only option being to configure it at the time of purchase. In the case of the MacBook Neo, the limit is even stricter: according to Cult of Mac, due to its chip design, it cannot even be factory-configured with 16 GB. stereonetcultofmac
On Windows, the situation depends on the model. Some business laptops still feature slots to add memory, but the trend in lightweight machines is heading in the opposite direction. With Intel Lunar Lake processors, for instance, RAM comes integrated into the package in 16 or 32 GB variants and cannot be upgraded later. Before ordering a Windows model, it's worth checking the spec sheet to see if the memory is soldered or if it has an open slot. gsmarena
How to tell if a laptop has run out of RAM?
A laptop without enough memory shows in the daily routine: it takes time to switch between applications, video calls freeze when sharing the screen, tabs reload on their own, and the fan turns on for no apparent reason. The confirmation is in the system's activity monitor, which shows if the memory is maxed out.
The most common signs in day-to-day use:
Delays when switching applications, especially during an active video call.
Browser tabs that reload upon returning to them, because the system unloaded them to free up memory.
Video calls that freeze when sharing the screen or opening another tool during the call.
Recurring "the computer is slow" tickets on relatively new devices.
To confirm it, on macOS the Activity Monitor has a Memory tab with a "memory pressure" graph: if it stays in yellow or red during a normal workday, the computer has fallen short. On Windows, the Task Manager shows the percentage of memory in use under the Performance tab. If it hovers around 90% with standard tools open, it is the same sign.
It must also be said: not all slowness is the fault of RAM. A nearly full drive, an outdated operating system, or misconfigured security software produce similar symptoms. That is why it is best to check the monitor before concluding that another device is needed.
What to do with laptops that have run out of RAM?
A laptop that has run out of memory has three options: reassign it to a role with a lighter workload, replace it and retrieve the previous device, or dispose of it permanently. Since RAM cannot be upgraded in most modern laptops, the decision is about the device's lifecycle rather than a repair.
Reassignment works when the difference is one step down. A 16 GB laptop that is no longer enough for an analyst can still be useful in a sales role.
When replacement is inevitable, the previous laptop cannot be left forgotten at the employee's home. It must be retrieved, securely wiped, and recorded in the inventory. Then it can be stored for reassignment or disposed of. At First Plug, we also have buyback options for devices that the company will no longer use: you can check the terms with our team.
When replacement is inevitable, the previous laptop cannot be left forgotten at the employee's home. It must be retrieved and securely wiped, and recorded in the inventory. Then it can be stored in a warehouse until someone else needs it, transferred to another employee or another country, or disposed of.
For devices the company will no longer use, at First Plug we have a buyback program: we evaluate the equipment, make an offer, and its value can be recovered in cash or as credit toward purchasing new devices. You can check the terms with our team.
How do we define the RAM for each device at First Plug?
The simplest way to avoid mistakes with RAM is to dedicate time to understanding specifically what each role or area of the company needs. At First Plug, we advise you on establishing a standard configuration per role—including processor, memory, and storage—so that each new hire is handled with a “sales profile” or “developer profile” rather than a debate over specs during onboarding week.
A standard also helps at the moment when the temptation to cut corners is strongest: when an 8 GB model appears at a great price. If the role's configuration is already agreed upon, the comparison is between devices that meet the standard, not between the cheapest one and the rest.
The equipment arrives configured and ready to work, thanks to us taking care of the entire process.
Additionally, we have our own operations in LATAM, which is a major plus in the region. To learn more about how to choose the right model for each profile, check out our guide to choosing laptops for home office .
Frequently asked questions
Is 8 GB of RAM good for anything today?
For office work with video calls, multiple tools open, and AI assistants, 8 GB falls short from day one. It may suffice for very limited uses, such as a shared computer, a training station, or a task handled by a single application. For a remote employee working eight hours on that computer, it is not a recommended configuration.
Is 16 GB of RAM enough to use AI?
To use cloud-based AI assistants like ChatGPT, Copilot, Gemini, or Claude, yes: the processing happens on external servers. 16 GB is also the minimum required by Microsoft and Apple for their system-integrated AI features. To run language models locally on the machine—common for development and data profiles—it is best to start from 32 GB.
Does 16 GB on a Mac perform the same as 16 GB on Windows?
There is no exact equivalence, and it is better not to assume that a Mac “performs twice as well.” Apple's unified memory is very efficient, but it is shared between the processor and graphics, just like in Windows laptops with integrated graphics. When sizing a work computer, the prudent approach is to use the same numbers across both systems.
Is it advisable to buy 32 GB of RAM for all employees?
No. For sales or administrative profiles, 32 GB is capacity that won't be utilized, and in 2026, memory is one of the components that makes laptops most expensive. A role-based standard works better: 16 GB as a general baseline, 24 GB for profiles with heavy multitasking, and 32 GB or more for design and development. The exception is a company that reassigns devices across very different roles: in that case, it is worth stepping up a tier.
Can the RAM on a MacBook be upgraded?
No. On all Macs with Apple silicon, the memory is integrated into the chip and cannot be upgraded after purchase. The capacity is chosen only once, when configuring the device. If the RAM falls short, the only alternative is to replace the computer.
Key takeaways
RAM is the one specification on a work laptop that almost never allows for correction. A small drive can be compensated for with cloud storage, and an accessory can be swapped out in a day. The memory chosen at the time of purchase is what the employee will have for the entire lifespan of the device. In 2026, with rising memory prices and 8 GB models returning to the market, that decision carries more weight than ever.
The solution is not to buy the maximum for everyone. It is to define a standard by role, with 16 GB as the baseline, and apply it every time a new hire joins, even if a cheaper model appears.
If you want to review the configuration for your team's upcoming hires, we can help you define that standard and provide a quote.
You might also be interested in…
















