How to get $100K in AWS credits in 2026 — the honest, step-by-step guide.

IThe question behind the question

Almost everyone who searches "how to get $100K in AWS credits" is really asking one of two things: "Is the $100K number real, or is it marketing?" and "Why can't I find the form?" Both answers are the foundation for everything else, so start here.

The $100K number is real. It is the typical award inside AWS Activate Portfolio — the credit tier for venture-backed startups. What makes it feel mythical is that AWS deliberately does not publish a dollar figure next to it. The public Activate page shows a tidy ladder: $1,000 for Builders, $5,000 for Founders, and then "Portfolio — for startups backed by an eligible investor or accelerator" with no number at all. That blank space is not an oversight. Portfolio is a discretionary range, and AWS does not want founders anchoring on a ceiling or treating it as an entitlement.

Inside that range, the floor is roughly $25,000 and the headline award is $100,000. The reason so many startups land on exactly $100,000 is mundane: it is the round number AWS reviewers reach for when a company clears the bar and the projected spend supports it. It is not a negotiated figure and it is not random — it is the default for a credible, well-scoped, venture-backed application.

The reason you cannot find the form is that there is no self-serve form for Portfolio. The $5,000 Founders tier has a public application you can fill out in five minutes. Portfolio does not. It is gated to two submission channels, and neither of them is a page you can reach by browsing aws.amazon.com. That single fact explains most of the confusion online — people apply for the $5K self-serve tier, receive $5K, and conclude that "$100K AWS credits" was a myth. It was not. They simply used the wrong door.

So the practical question is not "how do I qualify for a bigger number" so much as "how do I get into the channel that issues the bigger number." The rest of this guide is about exactly that: the channels, the application they feed into, and what makes a reviewer approve $100,000 rather than downgrade to $25,000.

IIWho actually qualifies — and who quietly does not

The eligibility bar for $100K is lower than founder folklore suggests, but it is not nonexistent. Here is the real test, stripped of the "must be YC-backed" mythology that circulates in founder Slacks.

The single hard requirement is an institutional vouch. AWS is underwriting a five-figure bet that your company will still be on AWS, paying full price, after the credits run out. To take that bet on a startup with little or no billing history, they need someone credible to vouch that you are a real, funded, going concern. There are exactly two acceptable forms of that vouch, and you need only one of them.

• An eligible investor or accelerator — A VC fund, micro-fund, syndicate, or accelerator that AWS recognizes as a Portfolio-eligible org. If you raised an institutional round — even a seed round — from a fund on AWS's eligible list, you likely already qualify. The list is broad: it is not limited to tier-1 brand-name funds.
• A partner attestation via ACE — An AWS partner with Advanced or Premier tier files an opportunity record on your behalf and attests to the engagement. The partner's track record substitutes for the VC's vouch. This is the route open to founders whose investor is unresponsive, whose fund is not on the list, or who simply want the application handled by someone who files these weekly.

Note what is not required. You do not need a famous accelerator. You do not need a priced Series A — late-seed companies clear Portfolio regularly. You do not need an existing AWS account with months of spend, although having one helps the reviewer's math. You do not need to be US-incorporated; Portfolio runs across the US, UK, EU, the Gulf (UAE, Saudi Arabia), India, and most of APAC.

And here is who quietly does not qualify, so you do not waste two weeks finding out the hard way. Pre-seed companies with no institutional capital top out near $30K, not $100K — the conversation changes the moment a priced round closes. Solo hobby projects and unincorporated side projects are not eligible for Activate at all; they get the standard free-tier allowance and nothing more. Direct AWS competitors — alternative clouds, certain AI-infrastructure substitutes — are declined as a matter of policy. And if your spend is mostly AWS Marketplace billings (Datadog, Snowflake, MongoDB Atlas billed through Marketplace), credits will not cover those line items, so the realistic award shrinks to whatever AWS-native consumption remains. None of this is hidden, but none of it is on the marketing page either.

IIIThe four routes to a $100K-class pool

There is not one path to $100K; there are four, and they suit different companies. Picking the wrong one is the most common way founders lose two to four weeks. Read these as a decision, not a menu.

Three of the four reach a $100K-class number for a venture-backed startup, and one is a competitive program that can reach far higher but takes far longer. The honest framing: for roughly three out of four founders reading this, the answer is Route A, and the other routes are situational.

IVThe step-by-step application, start to finish

This is the part the cluster pages summarize and this guide does not. Below is the actual sequence for the default route — partner-filed Portfolio — from the moment you decide to pursue it to the moment credits hit your billing console.

The sequence is short because the heavy lifting — the ACE submission, the funding-source selection, the follow-up with the partner-development manager — is not yours to do. Your job is four inputs and one decision. Everything else is handled by whoever files the record.

VWhat to put in the use-case — the paragraph that gets approved

If there is one thing to take from this entire guide, it is this section. The use-case paragraph and the spend projection attached to it are what a reviewer reads first and weighs most. The difference between a $100K approval and a $25K downgrade is frequently nothing more than how this paragraph is written.

A reviewer is pattern-matching for three things in the use case: that the workload is real, that it is AWS-native (it consumes AWS compute and services in volume, not just object storage), and that the spend projection attached to it is believable. A good paragraph satisfies all three in a few sentences. A weak one fails on at least one and invites a downgrade — the reviewer approves a smaller, "safer" number rather than rejecting outright.

Compare the two versions below. They describe the same company, but only one reads like a fundable AWS workload.

VIWhat the AWS reviewer is actually checking

Reviewers do not deliberate over each application like a grant committee; they run a mental checklist and pattern-match. Knowing the checklist demystifies why some applications sail through in eleven days and others get trimmed.

These are the questions, in roughly the order a partner-development manager works through them. None of them are secret, but they are rarely written down in one place.

• Is this a real, verifiable company? — A working website, a findable team, and a verifiable funding signal. The overwhelming majority of submissions clear this; the few that fail tend to have no web presence at all.
• Is the workload genuinely AWS-native? — The reviewer wants to see compute and managed services in the projection — ECS, EC2, Lambda, RDS/Aurora, Bedrock — not a static site on S3. A workload that barely touches AWS does not justify a large award.
• Is the spend projection believable? — The number has to fit the stage. Implausibly large projections get downgraded for lack of credibility; implausibly small ones undercut the case for the credits. This is the lever Section V exists to help you pull.
• Does the filer have a track record? — Partners with high ACE close rates are effectively pre-trusted; their applications move faster. A first-time filer with no history gets extra scrutiny. This is one concrete reason the channel you file through matters.
• Is there AWS billing history to corroborate the story? — An existing account with a few months of spend lets the reviewer check the math against reality. No history is not disqualifying — it just shifts more weight onto the filer's attestation.
• Is the company an AWS competitor? — AWS does not fund direct competitors — alternative clouds, certain AI-infrastructure substitutes. A small share of applicants trip this without realizing their product overlaps an AWS service; it is worth checking before you file.

VIIThe real timeline — and where it slips

Founders plan around "credits in two weeks" and are then surprised when it is five. The two-week figure is real for a clean partner-filed application; the slippage almost always comes from a small number of identifiable causes.

The table below sets the partner-filed and VC-filed paths side by side, then the notes underneath name the things that actually move the clock. The headline: the mechanics are fast; the delays are human.

VIIIThe common failure modes (and how to avoid each)

Most applications that fail or get trimmed fail for the same handful of reasons. None of them are mysterious, and all of them are avoidable once you know the pattern.

Read this as a pre-flight checklist. If none of these describe your situation, your application is in good shape; if one does, fix it before you file rather than after you are declined.

• The vague use case — The number-one cause of downgrades. A paragraph of adjectives with no named services and no spend figure gives the reviewer nothing to approve. Fix: write the concrete version from Section V before you do anything else.
• The fantasy projection — Claiming enterprise-scale spend at seed stage reads as either naïve or padded, and reviewers respond by trimming the award. Fix: anchor to a current number and a defensible trajectory in the $3K–$15K/month band.
• Applying through the wrong door — Filling out the $5K self-serve form and expecting $100K. The self-serve form only ever yields the Founders tier. Fix: use the Portfolio channel — VC submission or a partner ACE filing — for the $100K number.
• Double-filing with two partners — ACE records are visible across AWS's partner team. Two partners filing the same opportunity creates a conflict that gets the second submission rejected and flags both filers. Fix: choose exactly one filer per application.
• Stacking with no real second workload — Adding a Build for Startups or Bedrock POC application on top of Portfolio when there is only one product. The reviewer approves Portfolio and silently zeroes the duplicate. Fix: only stack when there is a genuinely distinct second workload — see Section IX.
• Unwittingly competing with AWS — A product that substitutes for an AWS-native service triggers the competitor filter. Fix: if your product overlaps an AWS service, get a read on eligibility before investing two weeks in an application.

IXHow to push toward $150K — without getting your stack rejected

The $100K Portfolio award is a base, not a ceiling. Two additive tracks stack cleanly on top — but only under one condition, and violating that condition is itself a common failure mode.

The condition is simple to state and easy to ignore: additive credits require additive workloads. AWS will not approve $100K of Portfolio plus $25K of Build for Startups if both records describe the same product — that is double-counting, and the reviewer sees it instantly. They will approve $100K of Portfolio for your general infrastructure, plus $25K of Build for Startups for a clearly distinct project (a media pipeline on MediaConvert, say), plus $25K of a Bedrock POC for a clearly distinct AI initiative (a support agent on Claude). Distinct is the operative word.

So the honest test before you stack is one question: is there a real second workload here, or am I describing one product three ways? If the answer is the latter, file Portfolio alone and stop — adding paperwork will not add dollars, it will only add a silent $0 line and a slightly more skeptical reviewer. If the answer is the former, the additive applications are worth filing in the same week, and the realistic combined ceiling at Series A lands near $150K.

The mechanics of stacking — which funding source each record requests, how to phrase two genuinely separate use cases so neither cannibalizes the other — are where an experienced filer earns their place. But the strategic decision is yours, and it is binary: one workload, one application; multiple workloads, stack with intent.