Microsoft AI PC Requirements: 16GB RAM and 40 TOPS?

Updated 9 November 2026 · Standard formula, shown below

Microsoft’s Copilot+ PC category sets minimum requirements of 16 GB of RAM and an NPU capable of at least 40 trillion operations per second (TOPS). Those figures are a threshold for a specific class of Windows PCs—not a universal definition of every AI-capable computer, nor a guarantee that every AI feature will run well.

What the 16GB and 40 TOPS figures mean

Microsoft introduced Copilot+ PCs as a category of Windows computers designed to support a set of AI experiences that use a neural processing unit, or NPU. The published baseline includes 16 GB of RAM, 256 GB of storage, and an NPU rated at 40 TOPS or higher. A device must also meet Microsoft’s other hardware and software requirements to qualify; the two numbers alone do not establish eligibility.

RAM is the computer’s working memory. It holds data and instructions that active programs need, and it is distinct from storage such as an SSD. TOPS means “trillion operations per second.” In this context, it describes the NPU’s rated throughput for AI-related calculations. It is not a measure of the computer’s overall speed.

So the direct answer is: 16 GB of RAM and a 40-TOPS NPU are Microsoft’s baseline requirements for Copilot+ PCs, but they are not a standard that every AI PC must meet. Other manufacturers and software providers may use “AI PC” more broadly, and individual features can have separate requirements.

Why Microsoft set a separate PC category

Many computers can run AI software, including systems that use a CPU, a graphics processor (GPU), or remote cloud servers. Copilot+ PCs are a narrower designation focused on local, on-device AI workloads that can use a dedicated NPU. The category gives buyers a hardware floor for the features Microsoft associates with that experience.

That distinction matters because “AI PC” is used as a general marketing term as well as a description of particular hardware. A computer may include an NPU but fall short of Microsoft’s 40-TOPS requirement; it can still run some AI applications, depending on the app and workload. Conversely, a PC meeting the baseline does not necessarily support every AI tool or future feature.

Requirements can change as software evolves. Check the current requirements for the specific Windows feature or application you intend to use, rather than relying only on a product label or an older specification sheet.

How an NPU and TOPS work

An NPU is a processor designed to handle certain AI operations efficiently. A computer’s CPU remains the general-purpose processor, while a GPU is suited to parallel workloads such as graphics and some machine-learning tasks. Software can use one or more of these processors, or send work to cloud services, depending on its design.

TOPS is a peak-throughput figure, not a direct prediction of how quickly a particular app will respond. Vendors may measure or report performance under different conditions, and the number does not capture memory bandwidth, software optimization, power limits, or the size and complexity of a model. Two devices with the same stated TOPS can therefore feel different in actual use.

For example, an NPU may help with supported effects such as background blur during a video call, or with a compatible local AI feature. A large generative model may still depend on a GPU, substantial system memory, or a remote service. The application’s implementation determines where the work runs.

Is 16GB of RAM enough?

Sixteen gigabytes is a reasonable baseline for everyday computing and the Copilot+ PC specification, but “enough” depends on what else the PC is doing. A browser with many tabs, video editing software, a virtual machine, and a locally run AI model can all compete for memory. Integrated graphics may also use some system memory, leaving less available to applications.

Consider more than the minimum if you expect to keep a computer for years or routinely use demanding creative tools. Check whether the RAM is upgradeable; many thin laptops have memory soldered to the motherboard and cannot be expanded later. Also check the amount of storage: the 256 GB Copilot+ baseline is a separate requirement, and large apps, projects, and local model files can fill a drive quickly.

How to check a PC before buying

Use the exact model’s specifications, not just the product family name. Configurations within one laptop line can differ in processor, memory, storage, and NPU capability.

Confirm that the advertised device is identified as a Copilot+ PC if that category is important to you.
Check for at least 16 GB of RAM and 256 GB of storage, and verify that the NPU is rated for at least 40 TOPS.
Look up requirements for the specific feature or application you plan to use; some have additional eligibility, region, language, or account conditions.
Consider workload, battery expectations, ports, display, repairability, and upgrade options alongside AI specifications.
On an existing Windows PC, review its system information and the manufacturer’s specification page. Windows may show processor details, but a model-specific specification is often needed to confirm NPU throughput.

Common misconceptions and limitations

“40 TOPS means the whole PC is 40 times faster.” No. It is a measure of a particular processor’s peak AI throughput, not a multiplier for general computing or a benchmark of every application.

“Any AI feature needs a 40-TOPS NPU.” No. Some AI features run on CPUs or GPUs, while others rely on cloud computing. The 40-TOPS figure is relevant to Microsoft’s Copilot+ PC category and compatible experiences, not a universal prerequisite for AI software.

“16 GB guarantees smooth local AI.” No. Memory requirements vary with the model, application, and other programs running at the same time. A device meeting the baseline can still be constrained by a demanding workload, and a faster NPU cannot compensate for insufficient memory or unsupported software.

AI processing on the device can reduce dependence on an internet connection for supported tasks, but it does not mean every feature works offline or that data handling is identical across apps. Review each application’s documentation and privacy settings to understand whether a task is processed locally or sent to a service.

Frequently Asked Questions

Does every AI PC need 16GB RAM and 40 TOPS?
No. Those are Microsoft’s baseline hardware requirements for Copilot+ PCs. Other computers can run AI software with different hardware, depending on the application and where its processing happens.

Is a 40-TOPS NPU the same as a 40-TOPS GPU?
No. They are different processor types, and their workloads, architectures, and performance characteristics differ. A TOPS figure alone is not enough to compare overall performance across processors.

Can a PC with less than 40 TOPS run AI features?
Yes, it may run AI features that use its CPU, GPU, or cloud services. It may not meet Copilot+ PC eligibility or support particular on-device features that require the higher NPU threshold.

Is 16GB enough for running AI models locally?
It depends on the model and the rest of the system. Small or optimized workloads may fit, while larger models and multitasking can require substantially more memory. Check the software maker’s stated requirements.

Does 40 TOPS guarantee that Copilot+ features will be available?
No. The processor threshold is only one part of eligibility. The PC must meet the category’s other requirements, and specific features may have additional software, availability, language, or account conditions.

Method: the standard formula for this calculation, rounded as each result states. See how the calculators work or report an error.