There’s no escape. Artificial intelligence is creeping into Windows, and with it comes increased OS storage requirements.
For a long time, the minimum storage spec for Windows 11 was 64GB. With newer Copilot+ PCs, that’s been bumped up to 256GB. But it doesn’t end here. This is just for the OS itself. As more developers implement AI models into their apps, you’re going to need even more storage — to handle the system and the apps.
Let’s dive into what all this means and why it’s happening. If you want to use system AI features (like Windows Recall) and locally hosted generative AI models (like DeepSeek) and diffusion models (like Stable Diffusion), then you can’t just focus on NPUs. Pay attention to storage, too.
Current AI storage space requirements
Microsoft provides clear guidance on Windows system requirements for AI models. The company’s AI-focused Copilot+ PC push includes a minimum hardware specification, and Copilot+ PCs must have at least 256GB of storage to qualify for Copilot+ branding.
That may not seem like a lot, but it’s a huge leap up from Windows 11’s previous minimum of 64GB. A Copilot+ PC’s minimum storage requirement is 4x more than a vanilla Windows 11 PC.
Why the bump? It seems strange at first glance. Microsoft is shipping Windows with a few AI models bundled into a handful of apps, like Microsoft Paint and Photos, but they don’t consume gobs of storage yet. Windows 11 occupies about 38GB on my Microsoft Surface Laptop 7 — which is one of those Copilot+ PCs — while the same OS uses about 41GB on my five-year-old desktop PC.
All of that’s likely to change as AI models evolve.
AI models will need more storage soon
When people talk about the “size” of an AI model, they’re referring to the number of “parameters” it contains. A parameter is one variable in the AI model that determines how it generates output, and any given AI model can have billions of these parameters.
Also referred to as model weights, these parameters occupy storage space to operate properly — and when an AI model has billions of parameters, storage requirements can quickly balloon.
Here are a few current examples:
- Google Gemma 2 — 1.71GB (2 billion parameters)
- Meta Llama 3.1 — 4.92GB (8 billion parameters)
- Qwen 2.5 Coder — 8.52GB (14 billion parameters)
- Meta Llama 3.3 Instruct — 37.14GB (70 billion parameters)
- DeepSeek V3 — 221.25GB (671 billion parameters)
As you can see, the storage space consumed by an LLM increases with the size of its parameters. The same is true for other types of generative AI models, too. For example, Stable Diffusion v1.4 takes up about 4GB while Stable Diffusion XL consumes over 13GB. OpenAI’s Whisper, a speech-to-text model, consumes between 1GB and 6GB depending on whether you choose a lightweight variant or the full-fat model.
Matt Smith / Foundry
But there’s something else that also increases with the size of an AI model’s parameters: its intelligence. Llama 3.1 405B, the biggest version of Meta’s AI model, scored 88.6 on the MMLU benchmark, meaning it correctly answered 88.6 percent of questions in the benchmark. Meanwhile, the lesser Llama 3.1 70B model scored 86 and the even lesser Llama 3.1 8B scored 73.
To be clear, there are ways to improve the intelligence of AI models without increasing parameters. That was a key innovation of DeepSeek-V3, the Chinese AI model that emerged from nowhere in December 2024 with performance that rivaled OpenAI’s best.
Still, the very best AI models are massive. DeepSeek-V3 is presumed to be smaller than OpenAI’s models — I say “presumed” because OpenAI doesn’t disclose parameter counts for its cutting-edge models — and yet it’s still over 220GB. Most apps won’t require a local installation of DeepSeek-V3 to power AI features, but they might want to use mid-sized models like Qwen 2.5 14B (8.52GB) or Llama 3.3 70B (37.14GB).
And it gets worse for your hard drive, because different AI models have different strengths. Some are better for writing in natural language while others are optimized for logical reasoning or step-by-step planning. That means you can’t just get by with one locally installed AI model. You’ll likely want a few. Perhaps dozens.
Other data needs on top of AI models
Okay, let’s rewind a bit. AI models might take up a lot of storage space in the future, but forget about that for a second. Right now, at the beginning of 2025, those big AI models don’t play a big part in Windows 11 systems. Then why does Microsoft require at least 256GB of storage for Copilot+ PCs? Seems a bit aggressive, no?
It comes down to one thing: data. Specifically, the data needed by AI models to do their thing. Data that isn’t part of the model itself.
Windows Recall is the perfect practical example. Recall takes periodic screenshots of your Windows desktop while you use it, which are stored locally on your PC. These screenshots are a running record of your PC activity. Recall applies AI to analyze those screenshots and create a searchable “semantic index” of all your past PC activity, which you can then search and reference yourself.
Matt Smith / Foundry
For example, I’ve written several articles about DeepSeek. So when I search for “deepseek” in Windows Recall, I see results that span the web pages I had open while researching the model. Recall uses AI to understand all my past activity and show me those past moments in time that are relevant to whatever I’m searching for.
All that data must be stored somewhere! And it takes up a lot of space. (Microsoft won’t even let you turn on Windows Recall unless you have at least 50GB of free storage.) That space isn’t immediately used, of course. It takes a while for that data to accumulate. But it does accumulate.
For me, Windows 11 says Recall consumes 1.7GB. That’s with me using Recall for a couple months for a few hours per day. At this rate, I wouldn’t be surprised if it weighs in at 20GB by the end of 2025.
So, what size storage drive do you need?
If you’re interested in running AI features and models locally, I recommend you buy a Windows PC with at least 1TB of storage. That’s only considering the needs of AI, though. If you were going to use 1TB for other things — like games or videos — then bump it up to 2TB. If you want to be future-proof, maybe crank it up to 4TB, potentially by installing an SSD on your own.
That’s a lot of storage, I know. But that’ll easily handle most of the local AI models that are currently available today. As more apps begin to use locally hosted AI models and rely on locally stored data, we’re going to be shocked at how much RAM and storage we’ll need.
Further reading: Free AI tools that run locally on your PC