June 15, 2026
BlogAzure MACConomics: Is a MACC Worth It at Any Discount?
There is a question Microsoft Azure customers are only now beginning to ask. It comes after consumption data has been gathered, forecasts assembled, growth plans documented. A commitment figure sits on the table with a Microsoft offer attached. Then someone on the customer’s side asks about the capacity behind that number. If the customer commits its full spend and Microsoft cannot deliver the infrastructure to consume it, then what?
Silence.
Someone promises to take the question under advisement. The response is not evasion. It is unfamiliarity. Microsoft designed the Microsoft Azure Consumption Commitment (MACC) in a world where capacity was abundant, and forecasting consumption was an art customers were mastering. The capacity question never needed asking. But now that world is no more, and soon this will be the first question every customer asks.
Capacity Was the Answer. Now It Is the Question
On paper, the MACC is a simple bargain. You commit to consume a fixed dollar amount of Azure over a term, typically three to five years. Microsoft discounts the meter in exchange for revenue certainty. Microsoft created the model in an era of effectively unlimited cloud supply, when the only meaningful risk in the deal was the accuracy of your own forecast. Capacity was the unstated premise of the entire arrangement.
Now that premise is cracking. Microsoft’s own leadership has told investors, quarter after quarter, that demand for its cloud exceeds the supply it can provide. AI workloads are absorbing datacenter capacity faster than concrete can be poured, and chips can be racked.
The public evidence of strain is everywhere: Quota requests sitting in backlogs with no estimated fulfillment date, deployments struggling in flagship regions, customers redesigning architectures around regions they never intended to use. Customers are living it, and many have already paid for workarounds out of their own budgets.
Capacity isn’t only a pillar of the MACC. It was the founding answer of cloud computing itself, with the promise that cloud supply was more accessible and scalable than anything you could build yourself. The MACC simply converted that answer into a financial instrument. You could commit with confidence, because the cloud always has room.
Now scarcity is running that conversion in reverse. The MACC agreements under construction right now are still built for the old answer, not the new question.
No Actuary Would Approve This
Strip the MACC down to its economics, and it is an insurance contract, with the roles reversed from what you might expect. The customer underwrites Microsoft’s revenue certainty, and the discount is the premium Microsoft pays for that protection.
Viewed through that lens, customers are being asked to absorb dramatically more risk while collecting dramatically less premium. The structure asks you to forecast your consumption with precision at the very moment Microsoft cannot forecast its capacity with confidence. If capacity is unavailable, the financial obligation survives anyway; the commitment is owed whether or not the cloud shows up. Workarounds carry real costs, from standby architectures in alternate regions to engineering time spent routing around constraint, and none of it appears anywhere in the deal.
And displacement itself carries a price tag that Microsoft’s own published pricing quantifies with uncomfortable precision. A D16as v5 virtual machine that runs $68.80 per 100 hours in East US 2 runs $82.60 in South Central US. Same machine, same meter, 20 percent more. Across thousands of meters, the regions customers overflow into when flagship capacity runs out may carry premiums of 10 to 20 percent over the regions they leave, and some of the available alternates stock barely six of every ten meters that a flagship region sells. Your only choices may be to pay more, to re-architect, or to pay more and re-architect.
All the risk sits on your side, all the protection sits on Microsoft’s, and the committed dollars arrive in Redmond either way. No actuary would approve that policy, and no MACC proposal on a table today has this risk priced into it. The MACC negotiation has lawyers, architects, procurement, and IT in the room. It has never had an actuary, and that absence is about to become expensive.
The remedy is not a new hire. Most large companies already employ people who price risk for a living, in treasury, in insurance, in enterprise risk management. They have simply never been invited to this table. But in the era of scarcity, they belong there.
Risk professionals read price tags differently. To them, a number is not a price until the risk traveling with it is counted. Pay-as-you-go (PAYG) is the perfect example. PAYG has long been framed as the mark of an immature cloud customer, the thing from which a MACC uplevels you. But reweigh the equation with risk in it and the hierarchy inverts. List price is the only price on the menu that carries no commitment risk, as it gives you the freedom to redirect spend, to rebalance across regions or providers, to wait out a shortage without owing anyone for capacity that never arrived. Once risk is priced, retail can become the best price in the room. Customers running production workloads at scale without a commitment are not necessarily behind. Under the new economics, they may simply be early.
The Pause Is the Power Move
So is a MACC worth it at any discount? The honest answer is that nobody knows yet, including Microsoft. The economic principles underneath every commitment-style agreement have shifted, and the new equilibrium has not been found. The discounts have not adjusted to the risk. The contracts have not adjusted to the question. The institution selling the commitment has not yet absorbed how much the ground has moved beneath it. The answers are forming, on both sides of the table, in real time.
All of this means that if a MACC proposal is in front of you, or a commitment renewal is on your horizon, the most valuable move available right now costs nothing. Take a pause. A pause is not refusal, and not paralysis, but deliberate diligence in a moment when speed serves only the old playbook.
Before anything gets signed, certain questions deserve real answers. What happens, contractually, if you cannot consume because the capacity is not there? What is Microsoft actually committing in return, beyond a number on a meter? What is your pay-as-you-go position truly worth, both as leverage and as insurance? How should a commitment even be sized in a world where supply is as uncertain as demand?
Those questions deserve answers. This MACC blog series will work through them in turn: How to size a commitment against capacity nobody guarantees; what to demand from Microsoft in exchange for a MACC; when to decide paying retail is the smartest price on the table; how to understand the hidden geography tax built into Azure’s regional price map; how to price capacity risk before Microsoft prices it for you; and how to take into account that a supply-constrained vendor might quietly prefer that you not commit at all.
The MACC was built for a world of abundance, and that world is gone. Until the agreement catches up, treat every commitment proposal as what it has quietly become — an open question wearing the costume of a routine deal.
Azure MACConomics Series
Part 1: Is a MACC Worth It at Any Discount? (this article)
