Meridian Technology Infrastructure - Cost Transparency: Executive Summary
Prepared for: CFO / Head of Technology Infrastructure
Period: Q1 2025 · Total infrastructure cost: $40.2M
The headline
For the first time, Meridian’s $40.2M infrastructure run-rate has been fully attributed from the General Ledger down to the products, applications, and programs that consume it. The cost base is highly concentrated, which is good news: managing a small number of large items moves the number.
What drives the cost
- Compute is half of everything. A single product, Compute, accounts for $20.1M, almost exactly 50% of fully-loaded cost. The next largest, Database, is $4.8M.
- Three cost types drive 70%. Labour ($12.4M), Public Cloud ($9.3M), and hardware depreciation ($6.3M) together account for roughly two thirds of the base.
- Shared overhead is 8%. $3.2M cannot be traced to any one product. We spread it fairly across products but report it as a separate, visible line so nothing is hidden.
Budget picture
In total, actual spend lands within about 1% of budget, but that near-match hides large offsetting movements. Compute and Database are consistently and increasingly under budget (Compute by 10% rising to 21% across the quarter), which points to over-provisioning or work running behind plan. The opportunity is a reforecast that releases or redeploys the unused budget on the largest product line.
One caveat to read correctly: End-User and Security appear enormously over budget. This is not real overspend. It reflects how shared labour is attributed to headcount-heavy products against budgets that were set at token levels. The action is to reset those budgets, not to cut those teams.
Scenario findings
- Vendor rate increase (+8%): +$204k, certain. A key managed-service vendor raising rates flows straight through to Database and Cloud Platform, with no natural offset. This is the larger and more predictable of the two.
- Cloud migration (15% of Data Centre): a bet, not a saving. The impact ranges from about +$50k to +$555k, and its sign and size hinge on two unknowns. First, the true cloud unit rate: at a 10% premium the move costs about $50k, at a 10% discount it saves about $50k. Second, how much data-centre cost is actually avoidable: CP04 is a fixed cost pool, so if little of it is shed when workload leaves, the move keeps most of the data-centre bill and adds the cloud spend on top, closer to +$555k. The outcome is decided by those two numbers, neither of which is in the data.
Risks and opportunities
- Opportunity: reforecast the over-provisioned Compute and Database budgets and redeploy the headroom.
- Opportunity: pre-empt the vendor increase commercially through renegotiation or scope reduction before it lands.
- Risk: the cloud migration is not a default win. Committing before the per-unit economics are proven could quietly add cost, and fixed data-centre costs may not fall as workload leaves.
Recommendation
Cloud migration is a bet, not a saving. Its impact ranges from about +$50k to +$555k, driven by the true per-unit cloud rate and how much fixed data-centre cost can actually be shed. Because the data centre (CP04) is a fixed cost pool, little sheds in the short term, which pushes the likely outcome toward the high end. Prove both numbers before committing a dollar. In parallel, act on the two certainties: reforecast the under-spending Compute and Database lines to release budget, and open the vendor conversation now to blunt the certain +$204k increase. Focus management attention on Compute, since at half the cost base it is where transparency creates the most value.