At GTC 2026, ClearML announced the general availability of its Platform Management Center, an administrative dashboard purpose-built for IT administrators and AI platform leaders managing multi-tenant ClearML deployments at enterprise scale. Available under the ClearML Enterprise plan, it gives cluster admins a single place to monitor every tenant’s activity, resource usage, and costs while protecting the privacy of tenant workloads and data.
The Problem: AI Spend Without Accountability
Enterprises are spending billions on GPU infrastructure to fuel AI workloads, but hardware acquisition was the only the first step. Another well hidden challenge is making that investment measurable. Without unified visibility across teams and business units, finance and IT leaders face a costly paradox: compute sits idle in one department while AI teams in another queue for capacity, and nobody has a reliable handle on how much to bill each tenant.
This results in wasted GPU capacity, challenging FinOps, and AI ROI that’s nearly impossible to measure. ClearML’s Platform Management Center was built to solve exactly that.
What the Platform Management Center Does
The Platform Management Center is organized around a tenant-level administrative view. From a single interface, platform admins can see every tenant in the deployment and drill into per-tenant dashboards that surface four categories of information:
- Cost tracking and attribution: As a completely customizable dashboard, estimated costs are displayed for any metered telemetry events (compute, storage, user and service account activity) with breakdowns by category (e.g. cost per GPU type, cost per storage service). Admins can view costs over a selected reporting period alongside a reference period comparison, making it easy to spot trends and anomalies. This provides the foundation for chargeback and showback models, and forward-looking AI budget planning.
- Resource and activity metrics: Each tenant dashboard shows current totals for projects and tasks, task creation trends over time, and task runtime in hours, broken down by queue. This gives platform admins an accurate picture of how each team is actually using the platform without requiring access to the workloads themselves.
- User quota and status visibility: The dashboard breaks down each tenant’s users into active, pending (registered but never logged in), and total counts, displayed against the tenant’s configured user quota. Admins can drill into the full user list with details including email, ID, and role, making license and access management straightforward across a large deployment.
- Tenant isolation by design: All of the above visibility is achieved without platform admins having direct access to the workloads or data within each tenant’s environment. Infrastructure remains fully transparent to the administrator while tenant confidentiality is preserved, a critical requirement both for CSPs and Telcos offering their infrastructure to customers as well as enterprises operating across business units with different data governance obligations.
Why This Matters Now
There has been a broader shift in enterprise AI maturity. Pilot projects are giving way to organization-wide AI programs, and the operational model hasn’t kept pace. IT and FinOps teams are responsible for costly GPU infrastructure but lack the tooling to manage it with the rigor applied to other parts of the business (e.g., clear accountability, predictable costs, and measurable outcomes).
“The bottleneck has shifted from hardware capacity to controlled processes,” said Moses Guttmann, CEO and Co-Founder of ClearML. “The Platform Management Center gives IT the control and financial visibility they need, while ensuring AI teams can move fast without being blocked by infrastructure.”
Imagine a cloud service provider hosting AI infrastructure for a dozen enterprise clients. Tracking each client’s compute consumption, generating accurate invoices, and offboarding a churned client all require custom tooling or manual work. Now, the Platform Management Center handles all of that natively, making GPUaaS and AIaaS a viable business model, not just a technical capability.
Availability
The Platform Management Center is generally available now under the ClearML Enterprise plan and through ClearML’s partner ecosystem. Organizations can learn more through documentation here and through scheduling a demonstration at clear.ml/demo.