Microsoft is preparing a substantial upgrade to its Purview Information Protection service, with a new Microsoft 365 Roadmap entry revealing plans to boost the daily auto-labeling limit for files stored in SharePoint and OneDrive by five times. The change, recorded as roadmap item 567890 and published on July 17, 2026, would raise the tenant-wide processing ceiling from 100,000 to 500,000 documents per day, slashing the time required to classify large backlogs of sensitive data.

General availability is targeted for August 2026 in the worldwide multi-tenant cloud, though the same entry also lists a preview date of September 2026—an atypical sequence that suggests the schedule could shift. Administrators should treat both dates as provisional.

A Long-Awaited Capacity Boost

The core of the update is straightforward: the daily cap on how many files Purview can automatically examine and label across all SharePoint sites and OneDrive accounts in a tenant is jumping from 100,000 to 500,000. For organizations that have been painstakingly rolling out sensitivity labels to millions of legacy documents, this is a game-changer.

Currently, Microsoft’s own documentation still points to the 100,000-file limit. At that rate, a company with 5 million unclassified files spread across team sites and personal OneDrive folders could face a 50-day backlog—assuming every file qualifies for labeling and no other policies compete for capacity. With the new limit, that timeline shrinks to as few as 10 days, enabling compliance teams to complete sweeping data protection projects in weeks rather than months.

It’s important to note that the limit is tenant-wide and applies across all active auto-labeling policies, not per policy or per repository. If an administrator runs three separate policies targeting different types of sensitive information, they will collectively draw from the same 500,000-file pool each day.

What Stays the Same

The dramatic throughput increase doesn’t come with a relaxation of the safeguards Microsoft built into service-side labeling. Auto-labeling remains a two-step affair: every policy must first run in simulation mode before it can be switched on for production. Simulation allows security teams to measure false positives and fine-tune conditions—such as sensitive information types, trainable classifiers, and location scopes—without risking the accidental application of encryption or access restrictions to thousands of files.

Supported file types are unchanged. Word, Excel, PowerPoint, and PDF documents at rest are all eligible, but files that are currently open by a user or locked by another process won’t be labeled during that session. List-item attachments in SharePoint are excluded entirely from this automated path. Failures due to unsupported formats, file protection settings, or transient service hiccups are still possible, though transient failures are automatically retried.

One nuance that remains critical: organizations that enable PDF support within their auto-labeling policies should pay particular attention. Microsoft’s troubleshooter documentation explicitly warns that turning on PDF classification can significantly increase match counts, which makes the higher daily limit especially relevant—and potentially risky if policies haven’t been rigorously tested.

Who Stands to Benefit—and What to Watch Out For

For large enterprises and highly regulated industries, the new ceiling is a practical necessity. Data volumes in SharePoint and OneDrive continue to balloon, and many organizations are only now embracing mandatory sensitivity labeling driven by zero-trust architectures and compliance mandates like GDPR, CCPA, and executive orders on data security. Being able to process half a million files per day means that a broad rollout of protections across a sprawling SharePoint farm no longer requires months of throttled scanning.

Compliance administrators and data protection officers stand to gain the most. Tasks like identifying all documents containing PII, financial data, or trade secrets and applying appropriate encryption labels can be completed in more realistic timeframes. Meanwhile, security architects who have been restraining their auto-labeling ambitions because of the old limit can now revisit those cautionary decisions.

But speed is a double-edged sword. A botched policy that generates a huge number of false positives and applies restrictive labels can now do so at five times the rate, potentially locking users out of their own content before anyone notices. The simulation phase becomes even more essential. Teams should treat the increased capacity as a prompt to revalidate existing rules, especially those that had been narrowly scoped to avoid dragging a historical backlog for weeks.

The Road to 500K

Microsoft’s sensitivity labeling framework has evolved steadily since its introduction. Auto-labeling for data at rest in SharePoint and OneDrive arrived several years ago, giving organizations a way to retroactively classify stored documents without relying on user prompts. The initial 100,000-file daily limit was intended to protect back-end performance and prevent runaway policies from overwhelming the service. At the time, it largely sufficed for typical mid-size deployments.

However, as tenant sizes grew and more customers migrated entire on-premises file shares into Microsoft 365, the ceiling became a well-known pain point among admins. Forum threads and Microsoft Tech Community posts regularly surfaced the question: “Why can’t Purview label faster?” This roadmap update is a direct response to that feedback, and it arrives alongside ongoing investments in Microsoft’s data governance and security stack, including tighter integration with Microsoft 365 Copilot and adaptive protection.

The staggered GA-previews timeline likely reflects the complexity of scaling the underlying infrastructure. Delivering a fivefold increase without degrading tenant performance requires careful rollout across data centers, which may explain the unusual order of dates.

Getting Ready: Steps for Administrators

No configuration changes are required to receive the higher limit once it becomes available; it will simply show up in tenants as Microsoft completes the deployment. But administrators should use the interim to prepare.

First, audit existing auto-labeling policies. Identify any rules that were deliberately restricted to small scopes because the old 100,000-file ceiling would have caused multi-week processing delays. With the new capacity, those restrictions might no longer be necessary. Carefully expand scopes only after reviewing simulation runs and confirming that false positive rates are acceptable.

Second, put PDF support under the microscope. If your policies currently exclude PDFs or you’ve hesitated to enable them, now is the time to run a dedicated simulation that includes PDFs and observe the impact. The extra headroom doesn’t excuse an unexpected wave of mislabeled files.

Third, monitor labeling activity in the Purview portal. After the limit increases, keep an eye on the auto-labeling reports for spikes in failures or unexpected matches. A wider net can surface previously hidden issues like unsupported file types or protected documents that block labeling.

Finally, update your organization’s internal documentation and communication plans. Let stakeholders know that historical classification projects will complete faster, and that the compliance dashboard might show a sudden acceleration in labeled assets. Setting expectations prevents confusion.

Eyes on August

If Microsoft sticks to its current timeline, the 500,000-file daily ceiling will begin rolling out in August 2026. That’s just weeks away, though the atypical preview delay suggests some caution is warranted. Regardless of the exact date, the move signals Microsoft’s commitment to helping enterprises tighten data protection at scale. For administrators who have been waiting for a faster auto-labeling engine, that’s welcome news—provided they double-check their policies before the accelerator gets stepped on.