A near-fourfold multiple on the winners says nothing until it is set against the share of sites that ever get their power.
The break-even hit rate for a powered land strategy is the cost per site attempted divided by the sale price of a site that gets its power. On the book's 180 MW campus that is $11.6 million against $45 million, a break-even hit rate of 25.8 per cent: a 3.88 times multiple on the winners loses money below roughly one site in four. Require a 15 per cent return over the three years to sale and the break-even rises to 39.2 per cent.
Worked in full in The Data Center Development Handbook by Julian R. Sterling, with every figure reproduced in a free workbook.See the book on Amazon →
Powered land is the lightest of the three data centre business models: buy or option land, spend on pre-development until the utility commits capacity and the entitlements are in hand, then sell the site to a builder or a hyperscaler at a price per megawatt. Nothing is built. The value is the power, and the risk is that the power never comes. The book prints a near-fourfold return on a successful site and frames pre-development spend as an option premium. It does not state the hit rate that makes the business work, which is the number that decides everything else.
| Input | Value |
|---|---|
| Land, acres | 240 |
| Land price, $ per acre | 15,000 |
| Pre-development spend, $ million | 8 |
| Sale price, $ per MW | 250,000 |
| Capacity secured, MW | 180 |
| Years to sale | 3 |
| Required return, illustrative | 15% |
| Land value recovered on a failed site, illustrative | 70% |
Every site attempted costs its full basis. Only the winners are sold. The expected value of one attempt is the hit rate times the sale price, minus the basis, and the break-even hit rate is where that equals zero.
Basis per site = 240 × $15,000 + $8 million = $3.6 million + $8 million = $11.6 million
Sale price of a winner = 180 MW × $250,000 = $45 million
Multiple on the winners = 45 / 11.6 = 3.88 times, about 57.1 per cent a year over three years
Break-even hit rate = basis / sale price = 11.6 / 45 = 25.8 per cent, one site in 3.9
In Excel: =(B1*B2+B3)/(B4*B5) with acres, price per acre, pre-development, price per MW and megawatts in B1 to B5.
The break-even is simply the reciprocal of the winners' multiple. That is the whole point: a headline multiple tells you nothing until it is set against the share of sites that reach it.
| Hit rate | Expected value, $ million | Multiple on basis | Annual return over 3 years |
|---|---|---|---|
| 10.0% | −7.1 | 0.39 | −27.1% |
| 20.0% | −2.6 | 0.78 | −8.1% |
| 25.8% | 0.0 | 1.00 | 0.0% |
| 33.3% | 3.4 | 1.29 | 8.9% |
| 50.0% | 10.9 | 1.94 | 24.7% |
| 75.0% | 22.1 | 2.91 | 42.8% |
At one site in three, a hit rate many developers would call respectable, the programme returns 8.9 per cent a year. That is below the return a land developer would accept for taking utility and permitting risk. The strategy only earns an equity-like return somewhere near one site in two.
Two refinements pull in opposite directions. Time raises the bar: the basis is spent up front and the winners sell three years later. At a 15 per cent required return the sale price must cover the basis compounded by 1.52, so the break-even hit rate becomes 11.6 × 1.52 / 45 = 39.2 per cent, one site in 2.6.
Salvage lowers it. A failed site still owns its land. If 70 per cent of the $3.6 million land cost is recovered, $2.52 million, a failure costs $9.08 million rather than $11.6 million, and the break-even falls to 21.4 per cent. Pre-development spend, the larger part of the basis, is not recoverable: studies, deposits and legal fees have no buyer.
| Convention | Break-even hit rate |
|---|---|
| No time value, no salvage | 25.8% |
| Land salvage at 70% | 21.4% |
| 15% required return | 39.2% |
| 15% required return with salvage | 35.6% |
The common mistake is to underwrite powered land on the winners: present the 3.88 times multiple and the 57.1 per cent annual return as the strategy's return, when they are the return on the sites that worked. The second mistake is to treat the expected value as what a small programme will earn. Expected value is a long-run average, and a developer with four sites does not get the long run.
Four sites at a one-in-three hit rate cost $46.4 million and are expected to produce 1.33 winners. The distribution is what matters. There is a 19.8 per cent chance that no site gets its power and the whole $46.4 million is lost, and a 39.5 per cent chance of exactly one winner, which returns $45 million against $46.4 million spent, a loss of $1.4 million. Taken together, the programme loses money 59.3 per cent of the time, even though its expected value is positive. Two winners make $43.6 million, and that outcome comes up 29.6 per cent of the time.
The hit rate cannot be read from a spreadsheet. It lives in the developer's own pipeline history: how many sites entered the utility queue, and how many came out with a capacity commitment on time. A strategy pitched without that history is pitched on the winners' multiple alone.
Divide the basis per site by the sale price of a winner. On the book's campus that is 25.8 per cent before time value and 39.2 per cent at a 15 per cent required return, and a small programme loses money more often than not even above break-even. The book compares powered land with the two build models on the same site, also covered in powered shell against turnkey returns. The hit rate is a single input on sheet 6 of the free workbook for this case: set it to your own history and the expected value recomputes.
Powered land is a site with secured utility capacity and entitlements, sold to a builder or hyperscaler before anything is built. In the book's illustrative case a 240-acre site costing $11.6 million including pre-development is sold at $250,000 per MW of capacity, $45 million for 180 MW, a 3.88 times multiple on the sites that succeed.
Each attempt costs the full basis and only winners are sold, so expected value is hit rate times sale price minus basis. Setting that to zero gives hit rate equals basis divided by sale price, which is one over the winners' multiple. A 3.88 times multiple therefore breaks even at 25.8 per cent, about one site in 3.9.
Yes, if it is small. Four sites at a one-in-three hit rate have a positive expected value, yet there is a 19.8 per cent chance of no winner and a 39.5 per cent chance of exactly one, which still loses $1.4 million. The programme loses money 59.3 per cent of the time.
Powered land is one of the three business models compared in Chapter 13 of The Data Center Development Handbook. The book takes the same case from first principles to the decision, chapter by chapter, and every figure it prints is a live formula in the free companion workbooks.
Get the book on Amazon →Free companion files
Also on Amazon UK · Amazon Germany · Amazon France · Amazon Canada
Reading guide: infrastructure, data centres and energy → · All 324 articles →
If this book helped, or didn’t, a few lines on Amazon are worth more than they look: they are what the next reader goes on. Write a review. The workbook stays free either way.