A credit is a one-off budget: it expires, it runs out, and it does not change your unit price. A discount is a long-term price change that applies for as long as you qualify. Confusing the two produces the conclusion that a few thousand in credits has solved your cost problem.
Search for AWS offers and you find two kinds of material: official pages listing credits and "80+ exclusive offers", and a stream of "complete discount guides" that flatten free tier, startup programmes and reseller discounts into one list.
Neither is wrong, but both omit the same layer: they put fundamentally different things side by side.
Credits and discounts are two different things - one answers "who pays this month bill", the other answers "what is my unit price from now on". Read together, they produce the conclusion that a few thousand in credits has solved your cost problem - which is exactly wrong.
The core difference: a credit is a budget, a discount is a price
| Credits | Discounts | |
|---|---|---|
| Nature | A budget denominated in money | A reduction in the unit price |
| Timing | Has an expiry, runs out | Has a term (a one or three year commitment) and is renewed |
| Does it change the price | No - it pays part of the bill on your behalf | Yes - same usage, genuinely lower charge |
| What determines how much | Eligibility (new account / startup / targeted issue) | Your usage shape and the commitment you make |
| When it ends | Back to on-demand pricing | Applies for the whole commitment term |
| Classic misunderstanding | "It is basically free money" | "Too complicated, I will wait for credits" |
One line to remember the difference: a credit pays the bill for you, the price has not moved; a discount means the price really is lower. So credits buy you time, while discounts are what change your cost structure.
That leads straight to a practical conclusion: once credits land, the priority is not to spend them, it is to build the long-term structure while they last. Because on the day they expire, the bill returns to whatever level you never optimised away from.
The official offers landscape: four entry points
The offers on the official site sit behind different entry points, and sorted by who qualifies there are four kinds:
| Entry point | Who it is for | Nature |
|---|---|---|
| AWS Free Tier | All new customers | Credits plus free allowances on selected services - they stack |
| AWS Activate | Eligible startups | Large credits, tiered by funding stage |
| Exclusive offers package | Depends on the specific offer | Third-party offers, membership benefits, education programmes - benefits, not cash |
| Commitment discounts | Anyone with steady usage | Discounts (not credits) - they live in pricing, not on the offers page |
Sorted this way, one thing becomes obvious: when most people search "AWS offers", what they actually want is in the third and fourth columns - and those two are obtained in completely different ways. The third depends on eligibility; the fourth depends on commitment.
- Free Tier (how to get the 200 USD, which account plan to pick) - see AWS Free Tier for new accounts.
- Activate (three tiers, eligibility, process) - see how to apply for AWS startup credits.
- Redeeming credits (where the code goes, the limits, the scams) - see how to redeem AWS promotional credits.
- Commitment discounts (Savings Plans versus reservations) - see how to buy AWS Savings Plans.
How to read "total value up to 800,000 USD"
On the startup column, the official wording promises 80+ exclusive offers - third-party offers, membership benefits and education programmes - with a total value of up to 800,000 USD.
That figure is easily misread as "you can get 800,000 USD". It is actually the sum of face values, with three important qualifications:
- It is face value, not cash: third-party software discounts and membership benefits add up to a large number at list price, but they can only be used for their designated purposes.
- It does not mean "useful to you": of those 80+ benefits, perhaps a handful relate to your business. Face value is worth nothing to you in the parts you cannot use.
- It is gated by the same eligibility as the credits: benefits of this kind usually arrive with Activate, so the first thing to satisfy is Activate eligibility - including the account plan requirement covered in the startup credits page.
⛔ When you see "total value of X hundred thousand", ask one question: how much of this is cash-equivalent? Credits are (they genuinely offset the bill); third-party benefits and membership discounts usually are not. Planning around face value as if it were budget overstates the resources you actually have.
Why the bill feels more expensive once the credits run out
This is the most common psychological gap, and the most practical section on this page: many teams run an on-demand-priced architecture for a whole year on credits. When the credits expire, the bill has not become more expensive - it has finally started showing the real price.
It "feels" more expensive because three things that should have happened got postponed:
- Right-sizing (bringing over-provisioned instances down) - no incentive during the credit period, because "it is not costing anything".
- Architecture work (moving to better price-performance silicon, keeping chattiest services in one availability zone) - postponed the same way.
- Buying commitment discounts to cover steady-state usage - some teams say "after the credits", and then start the next phase from full list price.
The correct order is the reverse: the credit period is the best window to do all three, because you have an environment where experimentation is free - you can shrink an instance and observe the effect, change architecture and validate stability, and size a commitment accurately with no bill pressure.
They stack - it is not either/or
This is often misread as "with credits I can ignore discounts". In fact they act on different layers and apply together:
| Layer | Determined by | Effect |
|---|---|---|
| Credits | Eligibility and issuance | Pay part of the bill on your behalf |
| Official discounts | The commitment you buy (Savings Plans / reservations) | Lower the unit price on the bill |
| Channel discount | How you settle | Lower the amount you actually pay |
The three act on different parts of the bill and do not conflict. So both of these statements are wrong: "I have credits, so I do not need to look at discounts" and "I already go through a channel, so official discounts do not matter". Each covers a different segment.
The channel layer is often misread as competing with official discounts. They lower different things: official discounts change the unit price, a channel changes what you pay. See AWS discounts and reseller pricing, and the cross-provider comparison in cloud discount mechanisms compared.
When to move from credits to discounts
A simple order of operations, by priority:
- Confirm the expiry date and the burn rate. If current usage leaves plenty of headroom, you have time; if it will be gone in months, start the next step immediately.
- Finish right-sizing and architecture work before the credits end. Neither requires a commitment and both are reversible, so they come first.
- Cover the part of usage that has already stabilised with commitment discounts. The already-stable part, not the hopefully-will-be-used part - over-committing wastes money the same way (see how to buy AWS Savings Plans).
- Then look at settling through a channel. It lowers what you actually pay and conflicts with none of the above, so it can be done at any point.
The whole page in one line: credits give you a time window; discounts give you a cost structure. Using the window to build the structure is the right use of a credit. Using it to postpone is why the cloud "suddenly gets expensive" on expiry day.
FAQ
I have credits - do I still need cost optimisation?
Yes, and this is the best time for it. Credits do not change your unit price; they pay the bill. They make your architecture no cheaper, they make the exploration free. When they run out, the bill reflects your architecture efficiency honestly.
Can credits and Savings Plans be used together?
Yes, they act on different layers: Savings Plans lower the unit price on the bill, credits offset the amount. They do not conflict and they stack. The only nuance is cash-flow planning - once a discount is active, the same credits cover more usage.
Are official discounts and channel discounts the same thing?
No. An official discount lowers the unit price on the provider bill; a channel discount lowers what you actually pay. Different positions on the same bill, and both can apply at once. Channel mechanics differ sharply by provider (on some, the channel is itself part of the pricing) - see cloud discount mechanisms compared.
Are "8 official AWS discounts" style lists reliable?
As a checklist, yes, with two caveats: such lists usually mix credits (an allowance) with discounts (a price change), which are different in nature; and the specific figures and criteria are on the official pages - third-party round-ups go stale. This page sorts by nature; check amounts and terms against the official source.
Should a new account take credits first or buy discounts first?
Take the credits you are entitled to first, but do not let them delay architecture decisions. The order is: secure available credits, use them to fund the exploration period, settle the architecture before they expire, then buy commitment discounts. Doing it the other way round (committing early, then discovering you had credits) loses at both ends.
What happens to unused credits at expiry?
The unused portion expires - it cannot be withdrawn, transferred, and generally cannot be extended (the official terms are authoritative). So leaving credits unused is not saving, it is waste. Treat "will I use this before it expires" as a question you have to answer.