The budget line is empty. The cloud commitment is not.

A buyer who says "the line I'd pay you from is empty for the year" is not negotiating price. They are reading you a record. Nothing you take off the quote changes it, because the constraint is not the amount — it is which line the amount would leave from. The useful move is to stop asking about budget and ask about the other number in the same building: the cloud commitment they signed, which they must spend or forfeit, and which a purchase made the right way draws down.

That is the answer. The rest of this is why the objection has that shape this year, how the drawdown actually works, and the two limits you have to say out loud before the buyer discovers them on their own.

The person who would sign your contract is measured on having fewer vendors

This year they are being graded on the opposite of signing you. A vendor-consolidation report for 2026 finds 68% of surveyed tech leaders planning to actively reduce their vendor portfolio in the next twelve months, with an average target of a 20% reduction. Gartner's forward view in the same write-up: by 2027, about 70% of organizations will consolidate cloud-native application vendors down to a maximum of three strategic providers.

The AI stack is where we would expect that pressure to land, because that is where the count grew. The same write-up, citing IDC, puts the average enterprise AI software vendor count at 11.4 per organization in the 1,000+ employee bracket, up from 4.2.

And the cutting is passive. SaaS portfolios shrank, per the same reporting, "quietly through expiring licences, not through active migration." That is not a migration project aimed at you — your contract simply does not come up for renewal.

Our read, not the report's: this is why "add one more small supplier" is the wrong shape of ask, and why a discount does not move it. You are not losing on price. You are losing on a row in a list somebody has been told to shorten.

The second number: money already committed, not yet spent

Any prospect large enough to have signed a cloud commitment is carrying a figure they must consume or lose. A practitioner guide to AWS EDP/PPA is blunt about what happens if they underspend: "underconsumption of the committed spending doesn't reduce the amount owed... or rollover credits to the next commitment term."

All three hyperscalers have the same instrument, and each lets a marketplace purchase count against it:

COMMITMENT VEHICLES
  AWS      EDP / PPA
  Azure    MACC
  GCP      CUD

IF UNDERSPENT
  amount owed        unchanged
  unused balance     does not roll over to the next term

IF SPENT VIA THE CLOUD MARKETPLACE
  retires the commit at        full value
  capped at                    25% of the annual commitment
  marketplace fee              3-5% (AWS), 3% (Azure, GCP)

Our read, not the source's: the fee is a real cost, and buyers take it, because the alternative is a new budget request walked through procurement, legal, finance and IT, in the year everyone has been told to shorten the vendor list. Spending money that would otherwise evaporate is easier than opening a fresh line.

Two limits, said first, before anyone finds them later

The rules tightened, and the tightening is where a marketplace purchase quietly stops counting. Since 1 May 2025, a product only counts toward an AWS commit if it runs entirely on AWS — the "Deployed on AWS" designation. And per the same guide, AWS notified sellers about the change, not customers.

Read that consequence slowly, because it cuts both ways. A buyer can route a purchase through the marketplace, believe they retired the commit, and find out at true-up that they did not. And a vendor with any part of its stack on a second cloud is, for this specific purpose, worth nothing to them.

DOES THIS PURCHASE ACTUALLY RETIRE THEIR COMMIT?

  1. Is there a commitment at all?        EDP / PPA / MACC / CUD
  2. Are they on track to forfeit part?   the forfeit is the real budget
  3. Does it fit under the 25% cap?       cap is annual, not per deal
  4. Does the product run entirely
     on that one cloud?                   AWS rule since 2025-05-01
  5. Is it listed on that marketplace?    fee 3-5% (AWS) / 3% (Azure, GCP)

  Any "no" -> it is an ordinary budget request again,
              and the line really is closed for the year.

Our position, stated as ours: you name the cap and the single-cloud rule yourself, in the first conversation. A buyer who discovers a limit you skipped stops trusting every other number in the proposal, and the cap is not a flaw in the mechanism — it is the mechanism.

The question that replaces "what is your budget"

Budget is the wrong question because it asks the buyer to re-argue a decision that has already been made above them. This one does not:

Don't ask:  "What's the budget for this?"

Ask:        "Are you on a cloud commitment — and how much of it
             are you on track to forfeit this year?"

Then:       "If we delivered this deployed and billed inside
             that cloud, does it come out of the commit
             instead of your line?"

The first question moves the conversation from a closed line to an open one. The second turns an architectural decision into a commercial one, and it has to be asked early, because "runs entirely on one cloud" is not something you retrofit the week before signature.

What it changes about what you build

This part is ours, not the sources'. Single-cloud deployability used to be an engineering preference, usually argued on portability grounds and usually lost. Under the May 2025 rule it is a commercial feature with a price attached: it decides whether the buyer pays you with money they already spent or with money they have to go and ask for.

That does not mean every product should collapse onto one provider. It means the trade-off is now legible — multi-cloud costs you access to a specific pot of your buyer's money, and if you take that trade, take it knowingly and say so.

FAQ

Is this only an AWS thing? No. AWS EDP/PPA, Azure MACC and GCP committed use all allow marketplace purchases to draw down the commitment. The "runs entirely on this cloud" tightening dated 1 May 2025 is the AWS rule specifically, and it is the one that catches buyers out, because sellers were notified and customers were not.

Does the buyer save money doing it this way? Not on the sticker — there is a marketplace fee of 3–5% on AWS and 3% on Azure and GCP. What they save is the forfeit: unused commitment does not reduce what they owe and does not carry into the next term.

How much of a deal can go through this route? Eligible marketplace spend retires the commit at full value, capped at 25% of the annual commitment. Above that cap, it is ordinary budget again.

We're small. Does any of this apply to our prospects? Only to prospects large enough to have signed a cloud commitment in the first place. But those are exactly the buyers running a consolidation mandate, so the two facts tend to arrive together: fewer vendors allowed, and a pot of money that has to be spent regardless.

What would change our mind about this? A provider removing the marketplace drawdown, or lifting the single-cloud requirement again. The rule moved once already, on 1 May 2025, and sellers heard about it while customers did not — so check the current terms before you build a pitch on them.

Check it before you use it

Everything above is a procurement mechanism, not a product claim, and we would rather it be checked than believed: both write-ups behind it are linked in the text, with their numbers. The single-cloud rule is the one to verify yourself, because it is the one that changed most recently and the one your architecture has to answer to. If you have something worth building, we'd like to hear about it.