Soft ComputersA new study cited by ZDNet, commissioned by Seagate, found that most businesses are finally getting a return on their AI investments. That is the good news. The bad news is that 62% of those same businesses say they cannot handle the storage demands that AI workloads create.
This is not an abstract problem for large enterprises. If your team is using Microsoft 365 Copilot, an AI-powered CRM, automated reporting tools, or any cloud platform with AI features baked in, you are already generating more data than you probably realize.
Why AI Eats Storage So Fast
When AI tools run, they do not just process your existing files. They generate new data constantly. Conversation logs, model outputs, version histories, audit trails, cached results, and copies of documents used as context all pile up. A team of ten people using Copilot inside Microsoft 365 for a year can produce a surprisingly large volume of secondary data that most businesses never account for when they set up their cloud storage plans.
The Seagate study also points to a compounding issue. As businesses see positive AI results, they tend to expand usage. More users, more applications, more integrations. Each expansion adds another layer of storage demand on top of what was already there.
What This Looks Like in Practice for Canadian SMBs
We work with Canadian businesses running anywhere from 10 to 200 seats, and storage overruns are already showing up. Microsoft 365 Business Standard and Business Premium plans come with 1 TB of OneDrive storage per user and a shared SharePoint pool. That sounds like a lot until you factor in AI-generated content, meeting recordings through Teams, and automatic versioning on shared documents.
When you hit your storage ceiling, Microsoft does not quietly expand your capacity. Files stop syncing. Backups fail silently. Staff get confusing error messages and start working around the problem in ways that create security gaps, like saving files to personal Google Drive accounts or emailing large attachments instead of using SharePoint.
Canadian businesses also have data residency obligations to think about. If you are in healthcare, finance, or any regulated space, you cannot just let data spill into whatever storage region a vendor defaults to. Your AI-generated data needs to live in the same compliant environment as the rest of your records.
Three Things to Check Right Now
- Review your current Microsoft 365 or Google Workspace storage consumption. Both platforms show usage dashboards in the admin centre. If any user or shared drive is above 70% capacity, that is a warning sign worth acting on before something breaks.
- Find out what your AI tools are storing and where. Copilot interaction history, Power Automate logs, and AI Builder outputs all consume storage. These are often enabled by default and grow without any obvious notification to the admin.
- Match your backup coverage to your actual storage footprint. If your AI tools are generating data that is not covered by your backup policy, you have a recovery gap. A ransomware event or accidental deletion will expose it immediately.
What We Recommend
The first step is a storage audit. We do these for clients as part of our managed IT work, and the results are consistently surprising. Most businesses are using significantly more storage than they budgeted for when they originally set up their Microsoft 365 or cloud environment, and that gap widens every time a new AI feature gets switched on.
From there, the fix is usually a combination of tiered storage policies and a proper archiving setup. Not every file needs to live in your primary cloud environment at full cost. Older records, completed project folders, and AI output logs can move to a cheaper archive tier automatically, using policies you set once and do not have to manage manually.
For businesses that are scaling AI use more aggressively, we also look at whether the current Microsoft 365 plan still fits. Moving from Business Standard to a plan with add-on storage, or adding Microsoft 365 Archive for SharePoint, is often cheaper than the productivity disruption caused by storage limits that nobody planned for.
The Seagate study makes clear that the storage problem is widespread and growing. Getting ahead of it is straightforward with the right visibility into what your cloud environment is actually doing. If you are not sure where your storage stands today, that is the right place to start.