Skip to content

Roadmap: managed compute

We are facilities-first today: build-to-suit and wholesale colocation for hyperscaler-grade inference. Over time, we intend to expand into managed compute where it is sensible for customers and operationally responsible.

This page describes the direction and the prerequisites—not a promise of current capacity.

  • Managed racks (or clusters) operated with production-grade runbooks and telemetry.
  • Clear tenant boundaries: dedicated vs shared environments explicitly defined.
  • Customer experience: onboarding, incident comms, change control, and measurable SLO targets.

Why we don’t lead with managed compute on day one

Section titled “Why we don’t lead with managed compute on day one”

Managed compute is credibility-intensive: it requires a mature operations posture, supply chain discipline, and security controls that must be proven over time. We prefer to earn it through repeatable commissioning and operations.

  • Operations maturity: SEV model, on-call coverage, post-incident process, preventative maintenance discipline.
  • Telemetry: power/thermal/network/security observability with actionable alerts (not dashboards-as-theater).
  • Security: access logging, tenant separation, auditability, and customer-approved workflows.
  • Supply chain reality: hardware availability and lead times that do not jeopardize customer delivery.
  • Customer demand signal: committed demand that matches a responsible rollout posture.

If you want managed compute later, we can design your facility engagement to make that path easier (controls, telemetry, access workflows, suite boundaries). Start with a facilities engagement and make the future options explicit in the milestone plan.