Series-A startups are inside the strongest credit eligibility band AWS has — but the headline tier ($100K+) is partner-filed, not self-serve. This page walks through every credit track a Series-A team qualifies for in 2026, what each track actually pays for, the application paperwork, the timeline, and which combinations stack.
AWS Activate gates credit ceilings on signal — primarily on funding stage, secondarily on a vetted "vouch" (an investor, an accelerator, or a CloudRoute-style routing partner). Series-A is the band where AWS's confidence in the company's longevity peaks against the credit cost being still small relative to the eventual AWS bill.
At pre-seed, AWS's default credit offering caps at $5K self-serve and $25K through a partner. The reasoning is brutal but rational: ~70% of pre-seed companies will not be on AWS in 18 months. Burning $100K credits on each one is unprofitable for AWS.
At seed, the ceiling lifts to roughly $50K partner-filed because the vetting signal — institutional money — is stronger. But seed-stage workloads tend to run < $1K/month on AWS, so even a $50K credit pool will last 4+ years at typical burn. AWS's tolerance for $50K credits is high; for $100K credits is moderate.
At Series-A, the math flips. Your AWS spend in the next 24 months is likely $5K–$25K/month. AWS's $100K credit investment is recovered the moment you graduate off credits and become a paying customer — typically months 18–24 post-Series-A. AWS knows this. Their credit pool for Series-A companies is the largest single tier in the Activate program.
The catch is that the $100K tier doesn't appear on the public Activate page. The public page shows the $5K self-serve tier — the $1K Builders tier — and a vague link to "Activate Portfolio for VC-backed companies." The portfolio track is real, but applying for it requires either (a) your VC submitting on your behalf, or (b) an AWS partner submitting on your behalf via the ACE program. CloudRoute routes Series-A founders to the second path because most VCs are slow.
There are five distinct credit pools a Series-A startup can apply for. Two are self-serve; three require a partner or investor sponsor. They stack — meaning a single Series-A company can claim from multiple pools simultaneously up to a soft ceiling around $150K.
The stack matters because no single pool reaches $150K alone. The headline tier — Activate Portfolio — caps at $100K. To get the rest, you layer Build for Startups ($25K) and Bedrock POC ($10K–$50K) on top. AWS reviewers know this and allow stacking when the use cases are non-overlapping (you're using portfolio credits for general infra; Build for Startups for a new vertical; Bedrock POC for an AI feature). That's the trick.
Most published "AWS credits for startups" guides miss this. They show the $5K self-serve tier and stop. The real game is layering three or four tracks until the ceiling is hit. The next section breaks down what each layer covers and how to apply.
At pre-seed, self-serve is mostly fine — $5K is enough for tinkering. At Series-A, self-serve becomes a structural cap on your credit ceiling. Understanding why means understanding how AWS's reviewer pool actually works.
When you submit a self-serve Activate application, the form goes to a low-touch reviewer queue. They check three things: did the company exist on paper? Does the website work? Is the use case AWS-compatible (i.e. not a competitor)? Approval is usually within 48 hours. Credits land in 3–7 days. Ceiling: $5K.
When you submit through a partner via the ACE (APN Customer Engagements) program, the application goes to a different reviewer queue — one that handles partner-attested opportunities. The reviewers there have higher trust in the documentation, because the partner has a reputation cost if they misrepresent. The reviewer ceiling shifts from $5K (self-serve) to $25K–$100K (partner-filed), depending on which sub-track the partner files under.
The difference isn't in the form fields. The difference is in who's vouching for the legitimacy of the application. A self-attested form caps low. A partner-attested form caps high. AWS's system is rational: trust scales with reputation.
At Series-A, your investor could also vouch for you via the Portfolio Sub-Program. About 30% of major VCs (a16z, Sequoia, Bessemer, Greylock, Founders Fund, NEA, Khosla, Index, GV, Lightspeed, Accel, Insight, Tiger, General Catalyst) have Portfolio Sub-Program access. If your VC has it, you can ask them to submit. The catch: most VCs are not in the habit of doing this on behalf of every Series-A portfolio company, and getting them to do it takes 3–6 weeks of back-and-forth.
CloudRoute's short answer: if your VC offers to submit within a week, take it. If they don't respond in a week, route through a partner via ACE — same ceiling, faster.
You don't fill in an Activate form. The partner does. They're the one logged into Partner Central — the gated portal AWS partners use — submitting an opportunity record with your company name, use case, and projected AWS consumption. You provide them with: company info (4 sentences), a deck (10 slides max), the use case (1–2 paragraphs), and your AWS account ID.
AWS reviews the partner's submission. If approved, the credits show up in your AWS account billing console under "promotional credits." They burn down as you incur charges, starting with the oldest first. The partner gets attribution credit — which is part of why they're willing to do this for free; they're working toward AWS partner tier requirements.
You need credits in 48 hours. You have an AWS account already and just need a small buffer. Your use case is genuinely small ($1K–$5K monthly). You don't want to talk to a partner.
In all these cases, the self-serve form at aws.amazon.com/startups/credits/ is the right move. It exists because some small needs don't justify partner overhead.
Portfolio is the credit pool AWS reserves for institutionally-funded startups whose investors or vetted partners can vouch for them. It's the tier most Series-A founders are eligible for but don't know exists in its full form.
Eligibility: institutional funding within the last 24 months (typically Series A, B, or notable seed-with-tier-1-VC). The funding doesn't need to be from a VC with Portfolio Sub-Program access — but the application either needs the VC to be in the program or a partner to file on your behalf.
Coverage: $100K total. Allocated across the AWS services you'll actually use (compute, storage, database, networking, ML). Excludes Marketplace purchases and AWS Professional Services. The credits are not earmarked — you spend them across any covered service freely.
Application time: ~3 weeks from initial submission to credits landing in your AWS account, when partner-filed. Faster (10–14 days) if your VC has internal AWS contacts and submits directly. Slower (4–6 weeks) if you're doing it self-serve through a contact form.
Validity: 24 months from issue. Unused credits expire — but at Series-A burn rates, $100K will typically be exhausted in 18–22 months of normal operation.
VC submits: slightly faster if they're responsive (10–14 days), but most VCs aren't. Higher ceiling possible in edge cases ($150K+) if VC and AWS have a co-investment history. Risk: 4–6 week delay if VC is slow.
Partner submits: consistent 11–18 day timeline. Ceiling stable at $100K. The partner handles the paperwork; you spend ~30 minutes on the application. CloudRoute routes you to a partner with Series-A experience based on your stack + region.
Build for Startups is a newer pool launched in 2024 — designed to fund specific workloads (a new product line, a new feature, a new vertical). It stacks on top of Portfolio because the use case is distinct.
Eligibility: any AWS-eligible startup (no funding-stage requirement); the workload has to be a discrete project you can describe in one paragraph. AWS reviewers approve if the project is real and the AWS services used are clearly itemized.
Coverage: up to $25K. Best used for: a specific feature launch (e.g., "we're adding a video transcoding pipeline using MediaConvert"), a specific compliance push (e.g., "we're building SOC 2 telemetry into CloudWatch"), or a specific new vertical (e.g., "we're launching a B2B portal that needs Cognito + API Gateway").
Application: partner-filed via ACE. Submitted as a separate opportunity from your Portfolio application. Approval takes 2–3 weeks.
The honest take: Build for Startups is rarely worth applying for in isolation. But layered on top of Portfolio, it's additive — and the paperwork is similar enough that filing both at once costs 30 minutes total (vs. 30 minutes for either alone).
In 2024, AWS launched a separate POC-funding pool specifically for teams running workloads on Amazon Bedrock (Claude, Llama, Mistral, Titan, Nova). The pool is partner-filed only — there's no self-serve form. This is the credit stream most AI startups don't know about.
Eligibility: any startup running or planning to run inference on Bedrock. Doesn't matter if you're already on OpenAI direct or running self-hosted Llama — what AWS cares about is your commitment to a Bedrock proof-of-concept of meaningful scope. "Meaningful" generally means: > $2K/month of projected Bedrock spend, a real use case (not a toy), and a clear evaluation plan.
Coverage: $10K minimum, $50K typical, ~$100K for larger ambitious POCs. The credits are Bedrock-earmarked — they can only be spent on Bedrock inference, embeddings, and the storage/compute that directly support the Bedrock workload (OpenSearch for vector search, S3 for prompt logs, Lambda for the orchestration glue). They cannot be spent on unrelated EC2 or RDS.
Application time: 2–4 weeks. Partner submits the POC plan + the projected inference budget + the eval methodology. AWS reviewer (Bedrock-team-adjacent) approves.
Why this exists: Anthropic, Meta, and Mistral are paying AWS marketing dollars to drive Bedrock adoption. AWS converts those dollars into credit funds for startups. The supply is real but bounded — apply earlier in the year rather than later.
A passing POC: a clear customer or product use case (not "exploring AI"); a chosen model with reasoning ("Claude Sonnet 4 because it balances cost and quality for our scope"); an eval harness ("we'll measure accuracy on N=500 examples each week against a held-out test set"); an explicit budget ("$15K/month at peak"); a defined POC window ("60 days from credits-applied to go/no-go decision").
A failing POC: "we want to explore generative AI"; "we'll use whatever model is cheapest"; "we don't have eval infrastructure yet but we'll add it"; "we're not sure of the budget"; "indefinite POC window."
CloudRoute's reviewer experience: ~85% of well-scoped POCs at Series-A get approved within 3 weeks. ~30% of vague POCs get approved at all, and the ones that do tend to land at the bottom of the range ($10K).
| Track | Ceiling | Filed by | Time-to-balance | Founder hours | Stackable? |
|---|---|---|---|---|---|
| Activate Builders (self-serve) | $1K | You | 24 hours | ~5 min | No (don't bother at Series-A) |
| Activate Founders (self-serve) | $5K | You | 3–7 days | ~30 min | Yes, with Portfolio |
| Activate Portfolio — VC submits | $100K | Your VC | 10–28 days | ~30 min | Yes, with Build + Bedrock |
| Activate Portfolio — Partner submits | $100K | Partner via ACE | 11–18 days | ~30 min | Yes, with Build + Bedrock |
| Build for Startups | +$25K | Partner via ACE | 14–21 days | ~10 min (additive) | Yes — adds on top |
| Bedrock POC funding | +$10K–$50K | Partner via ACE | 14–28 days | ~30 min (POC plan) | Yes — Bedrock-earmarked |
| Generative AI Accelerator | up to $1M | You (after acceptance) | 60–90 days | application + interview | Standalone (mutex with smaller tiers) |
A typical Series-A engagement looks like this. Numbers are pulled from CloudRoute's routed pipeline; partner-specific timelines vary by ±3 days depending on whether your AWS account already exists.
Day 0 — You submit an inquiry to CloudRoute (3 minutes). We route within 24 hours to a partner matching your stack, region, and timing constraints. You get a calendar link.
Day 1–2 — 30-minute discovery call with the partner. They confirm your use case maps to a credit track (it almost always does — Series-A is the easiest band to qualify). They share the application doc.
Day 3–4 — You spend ~30 minutes filling in: company info, AWS account ID (or "I need to create one"), use case paragraph, attach your deck. Email it back. If you don't have an AWS account, the partner walks you through creating one in 15 minutes.
Day 5–7 — Partner submits the ACE opportunity record for Portfolio. Same day, they submit the additive Build for Startups record if applicable. Same day, they submit the Bedrock POC record if applicable.
Day 8–10 — AWS reviewer queue assigns the records. Often a low-touch approval if the partner has handled similar Series-A applications before. Occasionally a single clarifying question from the reviewer.
Day 11–14 — Credits show up in your AWS billing console under "promotional credits." Bedrock POC credits show up tagged for Bedrock use only. Build for Startups credits show up untagged, with a budget reminder.
Day 14–18 — Partner starts the actual work (if applicable). If you only wanted credits without the partner doing build work, the engagement is over at day 14 — you got your credits, no further obligation.
It can be. ~15% of Series-A engagements land credits in under 10 days. The variables: do you have an AWS account already? Is your use case unambiguous? Is the partner experienced with Series-A applications? The mean is 14 days; the median is 13; the 90th percentile is 22.
Mistake 1: Applying only for the self-serve tier. The most common one. Founders see the $5K Activate Founders form, fill it out, get $5K in 3 days, and think the credit conversation is over. They never realize they were eligible for $145K more. Fix: always check whether you qualify for Portfolio before settling for self-serve.
Mistake 2: Filing through the VC and then doing nothing for 4 weeks. VCs say "yes I'll submit" and then disappear. You wait 4 weeks, no credits. By the time you realize they forgot, you've burned the window. Fix: give your VC 7 days; if no progress, route through a partner in parallel. Worst case you withdraw one application.
Mistake 3: Submitting a vague Bedrock POC plan. "We want to use Claude for things." This gets rejected or downgraded to the floor ($10K) of the range. Fix: define the use case, the model, the evaluation method, the budget, and the timeline before submitting. CloudRoute partners have POC plan templates that pre-fill 80% of this.
Mistake 4: Forgetting to apply for Build for Startups when Portfolio is already approved. Build for Startups is additive — it stacks on Portfolio without conflict. But you have to ask for it; AWS doesn't auto-grant it. Fix: ask the partner to file both at the same time. Adds 10 minutes to the application; adds $25K to the credit pool.
Mistake 5: Burning credits on the wrong services. Activate Portfolio credits cover any AWS service except Marketplace and Professional Services. But they don't cover Marketplace SaaS purchases (Datadog billed via AWS Marketplace, etc.). Founders sometimes set up Datadog through Marketplace, expecting credits to apply, and discover the bill is real. Fix: bill Marketplace SaaS directly through Datadog's own portal until credits are exhausted.
There's a long history of partner relationships being murky — partners who promise a lot, deliver little, and bill aggressively. The CloudRoute model is structurally different: customers pay $0; AWS funds the partner via credit programs; the partner pays CloudRoute a commission on closed work. Here's what the partner does, concretely.
They file the ACE record — the partner-attested credit application. Without a partner (or a VC with Portfolio access), this isn't available to you.
They vet your use case for AWS-compatibility before filing — so you don't get a rejection that burns a future submission.
They act as the named technical partner on the file — meaning AWS's reviewer trusts the application carries some technical credibility. This matters more than founders realize; reviewers fast-track applications from known partners.
They provide the actual AWS implementation if you want it — production AWS account setup, IAM, VPC, the migration off Heroku or wherever, observability, deployment pipelines. Most Series-A engagements include this; the credits pay for the cloud spend during the engagement.
They provide a co-branded readiness assessment — a written document showing what they built, what trade-offs they made, what to monitor. This is the deliverable. You can show it to your board, your auditors, your next investor.
They do not own your AWS account. The account is yours. They have IAM access during the engagement; that access expires.
They do not lock you in. CloudRoute partners commit to honoring a 30-day handover at any time — no contract clauses preventing you from taking the work in-house or to a different vendor.
They do not bill you for the credits. Credits flow from AWS to your AWS account directly. The partner is paid via AWS's engagement-funding programs (not from your wallet).
The honest comparison most "AWS credits for startups" guides skip.
| Variable | Self-serve only | Portfolio only | Full stack (Portfolio + Build + Bedrock POC) |
|---|---|---|---|
| Credit ceiling | $5K | $100K | $150K |
| Time-to-balance | 3–7 days | 11–18 days | 14–21 days |
| Founder hours | ~30 min | ~30 min | ~70 min |
| Validity window | 12 months | 24 months | 24 months |
| Reviewer queue | self-attested (low ceiling) | partner-attested (high ceiling) | partner-attested + Bedrock track |
| Stacks with future credits? | Yes (Portfolio later) | Yes (Bedrock POC later) | Yes (Gen-AI Accelerator later — competitive) |
| Cost to founder | $0 | $0 | $0 |
| Risk of rejection | < 5% | ~8% | ~12% (Bedrock POC scrutiny adds risk) |
Situation: Hitting Heroku platform limits before a hospital pilot launch. No internal cloud lead. CTO had spoken to AWS sales and bounced — too slow, too generic. Needed HIPAA-aware production AWS setup with credits to cover ~14 months of runway.
What CloudRoute did: Routed within 22 hours to a UK partner with HIPAA + Heroku-migration competencies. Partner filed Activate Portfolio ($100K base) + Build for Startups ($25K HIPAA-specific telemetry work) on day 4. Bedrock POC ($25K) for an AI feature on the roadmap was filed day 5.
Outcome: Production AWS account live in 9 days (week 2). All three credit tracks approved within day 14. Total credits applied: $150K. Heroku decommissioned in week 4. Hospital pilot launched on schedule. Total founder time across the engagement: ~6 hours.
engagement window: 9 weeks · founder time: ~6 hours · credits secured: $150K
No procurement loop. No discovery theater. We route within 24 hours; the partner submits the credit application; credits land in 11–18 days.