Compare the routes to usable capacity.

A data centre operator needs an additional 12 megawatts of usable IT capacity within two years. It can expand an existing site or lease capacity from a partner. Both routes must meet the customer's availability and data-location requirements.

The client wants the better five-year cash outcome and can commit no more than £30 million upfront. Customer IT hardware is outside this comparison and is funded separately under either route.

The important words are usable IT capacity. A site's total electricity supply also supports cooling and other facility equipment. A 12 MW electricity connection does not automatically mean 12 MW is available for customer IT.

Confirm the capacity definition, when the customer needs it and whether 12 MW must be available together or can arrive in phases. Those answers determine which options are feasible.

I would present the opening in this order:

“First, I would check which routes can provide the full 12 MW by the deadline and meet the service requirements. Second, I would compare their cash flows over the same five-year period, including upfront investment and ongoing costs. Third, I would test the risks most likely to change that answer, particularly power availability, construction timing and the partner's committed capacity.”

This order prevents detailed financial work on an option that fails a non-negotiable requirement.

Check grid connections and electrical equipment as specific sources of delay. A named dependency is more useful than a vague bullet labelled “implementation risk.”

The interviewer supplies a simplified cash exhibit. These are case figures, not construction-cost benchmarks.

Both options are initially assumed to deliver the full 12 MW on the promised date. The comparison covers the five years beginning on that date. Assume full customer occupancy throughout, no price changes and no residual value at the end. Ongoing costs include the facility costs in scope; customer IT costs are excluded from both options.

Over the stated window:

• Build: 5 × £11m − £28m = £27m.
• Lease: 5 × £5m − £2m = £23m.

Build leads by £4 million and fits the £30 million upfront limit. It consumes £26 million more upfront than leasing, then produces £6 million more operating cash each year.

Those differences explain the trade-off. The client pays more initially to keep more of the operating economics later.

This exhibit gives an undiscounted cash comparison. If the client asks for an investment valuation, request the discount rate, actual payment dates and end-of-period asset or contractual value. Do not quietly relabel £27 million as net present value.

Now the interviewer reveals that the build could open six months late. The lease remains available on time.

The five-year comparison window stays fixed. For this sensitivity, assume building costs remain £28 million, operating receipts and costs start together when the site opens, and there are no delay penalties. The delayed build has only 4.5 operating years in the window:

4.5 × £11m − £28m = £21.5m.

Lease still produces £23 million. The financial preference reverses even before any additional construction costs or customer penalties.

The original £4 million advantage buys only £4m ÷ £11m = 0.364 operating years, roughly 4.4 months, of delay under these assumptions.

That is a useful sensitivity to say aloud. It turns “construction timing matters” into a threshold the client can investigate.

But do not let the calculation obscure the deadline. If the customer cannot accept a late opening, the delayed build fails the service requirement regardless of its eventual cash return. You would need an on-time alternative or a credible bridging arrangement.

Suppose the partner then confirms that only 8 MW is firmly available by the deadline. The remaining 4 MW is an option without a committed delivery date.

You now have a new feasibility problem. An 8 MW lease cannot inherit the £15 million receipts assumed for a fully occupied 12 MW facility. Nor can you compare its lower price with the full build price and declare leasing cheaper.

Ask for a firm 12 MW offer or construct a complete alternative: 8 MW with this partner plus 4 MW elsewhere, including any extra connectivity, coordination and service costs. Check whether splitting the workload across locations is acceptable to the customer.

The correct response is to update the comparison. You do not need to defend the earlier answer after an assumption changes.

Power efficiency can affect operating costs, but it does not replace a capacity check.

PUE is the ratio of total facility power to IT power. It describes facility overhead, but does not by itself establish the usable IT capacity the site can guarantee under its operating conditions.

For the interview, ask what capacity the engineering team will commit to, what connection supports it and what electricity cost assumptions are in the exhibit. You can follow the commercial implications without pretending to size the facility yourself.

On the initial numbers, the answer is:

“Build provides £4 million more cash over the five-year window and stays within the upfront limit. I would choose it if the full 12 MW can be delivered on time. The advantage disappears after roughly four months of delay, so the next priority is validating the power connection and construction schedule. If the deadline cannot be met, we need a firm lease or bridging option for the full requirement.”

That recommendation explains what you prefer, why, and what could change it.

Practise the same comparison in the manufacturing capacity structure drill. For the broader entry decision, read the market-entry framework.