Skip to content

Annual Commitment Break-Even

Model a committed-spend agreement against on-demand pricing month by month, including the months you pay the floor without using it.

Net effect of committing, over 12 months
$3,246

Committing is cheaper by this much across the term. You realise 5.7% off list overall — less than the 15% headline, because of the months you pay the floor without using it.

On-demand total, at list
$56,931
Committed total, floor honoured
$53,685
Commitment fully consumed in
month 7
Paid for and not used
$5,293
Final month at list
$6,995
Final month, discounted
$5,946
Discount actually realised
5.7%
What this assumes: that unused commitment is forfeited at the end of each month rather than rolled forward, that the discount applies to all usage for the whole term, and that growth compounds monthly from today’s figure. Contracts vary on every one of those: rollover, tiered rates above the commit, and true-up at the end of the year all change the answer. Read the term sheet, then set the fields to match it.

What a commitment really trades

A committed-spend agreement is not a discount. It is a discount sold in exchange for taking on the risk that your usage falls — and usage falls for reasons that have nothing to do with the contract. A model gets cheaper. A prompt gets shorter. A cache lands. A feature is retired. Every one of those is a good outcome that the commitment converts into money you pay for nothing.

The month the commitment is fully consumed is the number worth screenshotting. Before it, your effective discount is worse than the headline and can easily be negative; after it, you are getting the rate you signed for. If that month lands late in the term, the deal is being priced off a growth curve you have not delivered yet.

The asymmetry to notice

The upside of committing is capped at the discount rate. The downside is capped at the commitment. That asymmetry is why the standard advice — commit to a level you are already exceeding, not to the level you project — survives every negotiation: it makes the bad case impossible rather than merely unlikely. Set the growth field to zero and see whether the deal still works. If it only works on the forecast, it is a bet on the forecast.

Annual Commitment Break-Even · Multigrid