$75K is not a published Activate tier. It is the mid-range outcome inside the seed-stage Portfolio range ($50K–$100K), awarded when the institutional vouch and the projected consumption are both moderate rather than maximal. This page walks through the reviewer math that lands an application at $75K specifically, the stacking moves that turn $75K into $100K–$125K, and the downgrade scenario where teams targeting $100K land here instead.
If you read the Activate page, you will find Builders ($1K), Founders ($5K self-serve / $25K partner-filed), and Portfolio (no dollar figure, gated to institutionally-funded companies). $75K appears nowhere. Yet a non-trivial share of CloudRoute-routed engagements land at exactly $75K. The reason is that Portfolio is a discretionary range, not a flat number, and $75K is the most common mid-range outcome inside it.
AWS Activate Portfolio is documented as "for venture-backed startups." The dollar amount is intentionally absent from the public page because the award is set per-application by the reviewing partner-development manager. The internal range is widely understood to be $25K at the floor, $100K at the typical ceiling, with outlier cases higher when there is a substantial migration or a specific AI cohort signal layered in.
Inside that range, $50K, $75K, and $100K are the three most common landing numbers. The reason is mundane — AWS reviewers round to $25K increments. They do not award $63K or $87K. They look at the application, decide it scopes to "the lower half," "the middle," or "the upper end" of Portfolio, and assign the corresponding round number.
So when a founder asks "how do I get $75K in AWS credits," the honest answer is: you do not specifically apply for $75K. You apply for Portfolio with a use case that scopes to the mid-range, and the reviewer lands you at $75K. The application form is the same one that produces $50K outcomes for floor-end applicants and $100K outcomes for ceiling-end applicants. The difference is in the inputs.
This page is about the inputs that produce a $75K outcome specifically. It also covers the downgrade case (applicants targeting $100K landing at $75K) and the upgrade case (applicants targeting $50K landing at $75K because their projection math was stronger than they assumed).
The AWS partner-development manager reviewing a Portfolio ACE record evaluates four factors. The combination of "moderate" answers across all four — rather than "strong" or "weak" — is what produces the $75K mid-range outcome.
Two variables push the award upward (toward $100K). Two variables hold it down (toward $50K). When all four are mid-range, the application lands at the middle of the Portfolio range: $75K.
Floor ($50K) signal: projection $1K–$2K/month. Reviewer reads this as a "small startup with modest infrastructure." Pool sized to ~24 months of consumption at floor end.
Mid ($75K) signal: projection $3K–$4K/month. Reviewer reads this as "seed-stage building toward production scale." Pool sized to ~18 months at this burn.
Ceiling ($100K) signal: projection $5K+/month with itemized service breakdown (ECS Fargate at $1.8K, Aurora at $1.2K, Bedrock at $1.5K, etc.). Reviewer reads this as "scaling infrastructure with credible engineering plan."
The projection is the single largest input. A team that under-projects to $1.5K/month (because they have not modeled production load) will be sized down to the $50K floor even if every other variable is strong. A team that over-projects to $20K/month at seed-stage will be flagged as implausible and downgraded.
Floor ($50K) signal: partner-filed ACE record with no specific VC named, or with a VC that is not in the AWS Portfolio Sub-Program. Treated as "the partner believes the use case is real but AWS cannot independently verify the funding signal."
Mid ($75K) signal: partner-filed with a named institutional lead — regional VC, family office, or strong angel syndicate at $1M+ aggregate. Reviewer accepts the partner attestation plus the named investor as a moderate vouch.
Ceiling ($100K) signal: VC has direct Portfolio Sub-Program access OR the partner-filed vouch references a tier-1 VC (Sequoia, a16z, Accel, Lightspeed, Index, etc.) as the lead. Reviewer treats this as full institutional vouch.
The vouch quality is the second-largest input. A bootstrapped team filing through Portfolio mechanic (which they should not — they should file Founders) will be rejected outright. A team with a real but mid-tier institutional lead lands in the middle.
Floor / mid signal: greenfield account, or account with less than 3 months of consumption history. Reviewer cannot verify the projection against actual usage patterns.
Mid signal: 3–6 months of history at $1K–$2K/month. Enough to verify the team is using AWS, not enough to confidently extrapolate to the higher tier.
Ceiling signal: 6+ months of history at $2K+/month with steady upward trend. Reviewer can extrapolate the projection from actual data.
AWS account history matters more than founders realize. A team with 6 months of $3K/month real consumption asking for $100K Portfolio gets approved almost automatically — the math is verifiable. A greenfield team asking for the same $100K gets cut to $75K or $50K because the projection has no historical anchor.
Floor signal: generic "we will run our SaaS on AWS." Reviewer reads this as "any cloud would work; AWS is one option."
Mid signal: specific services named (ECS, Aurora, S3, CloudFront) with reasonable rationale for each. Reviewer accepts the team has thought through the architecture.
Ceiling signal: specific services named PLUS a distinct second workload identified (AI/ML, video transcoding, IoT, etc.) that justifies the Bedrock POC or Build for Startups layer. Reviewer can see why the full Portfolio range is warranted.
When the workload is single-product (only "the SaaS"), Portfolio tops out at $75K — there is no distinct second use case to justify the ceiling. When the workload has a clearly additional component, Portfolio reaches $100K and the secondary application stacks on top.
Aggregating Portfolio approvals at the $75K mark, three profiles appear most often. If you fit one of these, $75K is your statistically likely outcome — not $50K, not $100K.
Who: seed-stage startup with $1.5M–$5M raised, lead investor is a regional VC, vertical-specialist fund, or family office. The VC is not in the AWS Portfolio Sub-Program directly. Projected AWS consumption $3K–$4K/month after migration to production.
How they land at $75K: the partner files Portfolio ACE with the named investor as the vouch. The reviewer accepts the institutional signal but applies a partial discount because the vouch is partner-mediated rather than VC-direct. The projection ($3K–$4K/month) sizes the pool to the mid-range. Result: $75K Portfolio approval, 13–17 days to credits.
Layering: almost always layered with a $25K Bedrock POC if the team has an AI workload (now common — most seed-stage SaaS has at least a chatbot, recommendation engine, or content-generation feature). Total stack: $100K. Some teams add Build for Startups for a distinct second workload, reaching $125K total.
Who: late-seed startup ($3M–$8M raised) where the lead VC does have Portfolio Sub-Program access — but the team has a conservative AWS consumption projection. Typical reason: the team is migrating from a managed platform (Vercel, Heroku, Render) and the post-migration baseline is still being established.
How they land at $75K: the VC submits Portfolio directly. The vouch is at the ceiling-quality (Portfolio Sub-Program access is the strongest possible signal). But the projection sizes the pool down — at $3K/month projected, the reviewer cannot justify the $100K ceiling because the math does not work. Result: $75K Portfolio with the implicit suggestion to re-apply for additional credits once consumption ramps.
Layering: these teams often add Build for Startups ($25K) for the migration-specific workload — IaC tooling, observability setup, security baseline. Combined: $100K. Bedrock POC layers on top if there is an AI initiative, reaching $125K.
Who: pre-seed startup in a top-tier accelerator (Y Combinator, Techstars, 500 Global, Antler) where the accelerator brand pushes the application beyond what raw funding stage would justify. Typically these teams have closed a small pre-seed ($500K–$1.5M) and the accelerator is acting as the institutional vouch.
How they land at $75K: the partner files Portfolio citing accelerator membership as the institutional signal. The reviewer accepts the accelerator brand as a partial vouch — equivalent to a mid-tier VC — and lands the award in the mid-range. Result: $75K Portfolio when the accelerator carries enough weight to push beyond the $50K floor but not enough to claim the $100K ceiling.
Layering: Bedrock POC is the natural addition (most accelerator cohorts now lean AI-heavy). The full stack reaches $100K. Build for Startups is harder to justify at pre-seed because distinct second workloads are uncommon.
A meaningful share of $75K outcomes are downgrades. The team applied for Portfolio expecting $100K — and the reviewer cut the award to the mid-range. Understanding why this happens lets you avoid the cut on your application.
Across CloudRoute-routed engagements, the most common reasons a $100K Portfolio request gets downgraded to $75K cluster into four causes. The pattern is predictable enough that vetted partners now pre-check the application for these before filing, which is the practical reason the partner-filed path has a lower downgrade rate than VC-direct submissions where there is no pre-check.
Cause 1 — Projected consumption math is conservative. The team projects $2K–$3K/month rather than $5K+/month. Reviewer cannot mathematically justify the $100K ceiling because the credits would last 40+ months — beyond the typical 24-month Portfolio validity. The cut is automatic at this point: the reviewer rounds down to the pool size that fits the projection. Fix: realistic projection that itemizes services and accounts for production-load growth (not the current pre-launch baseline).
Cause 2 — No distinct second workload identified. The application describes a single product (the SaaS) with no additive use case (AI feature, video pipeline, data lake, etc.). Reviewer reads this as "one workload" and sizes the pool accordingly. The $100K ceiling is reserved for applications where the workload composition justifies it. Fix: if there is a real second workload, include it in the application; if there is not, accept that $75K is the appropriate cap rather than push for an unjustified ceiling.
Cause 3 — Partner-filed vouch with non-tier-1 VC. The institutional vouch is real but mid-tier. Reviewer applies the partial discount discussed above. Fix: nothing you can do directly — your VC is your VC. But you can offset this by strengthening the projection and workload-specificity inputs to compensate. Teams that do this often recover to the $100K ceiling despite a mid-tier vouch.
Cause 4 — AWS account is greenfield with no consumption history. Reviewer cannot verify any projection because there is no historical anchor. Default behavior is to discount the projection by ~25% and size the pool accordingly. Fix: if you have time, run 60–90 days of real AWS consumption at $1K–$2K/month before applying. Even a modest history changes the reviewer evaluation significantly.
Across CloudRoute-routed Portfolio applications, approximately 22% of $100K requests land at $75K instead. Approximately 8% land at $50K (the floor). Approximately 70% land at $100K as requested. The partner pre-check eliminates most of the avoidable downgrade causes; the residual rate is driven by structural factors (vouch quality, projection math, account history) that are not always fixable in the moment.
A bare $75K Portfolio approval is unusual in practice. Almost every $75K Portfolio is layered with at least one secondary application — most commonly Bedrock POC, sometimes Build for Startups. This section walks through the realistic stacking compositions.
The stacking logic is the same as for the $100K Portfolio: secondary applications must describe distinct workloads, not the same workload as the Portfolio base. If Portfolio covers the general SaaS infrastructure, Bedrock POC must cover a specific AI initiative (a customer-support agent, a content-generation pipeline, a retrieval-augmented Q&A feature, etc.) — not "AI for our SaaS in general."
Build for Startups must similarly cover a clearly distinct second workload. The most common Build for Startups workloads alongside a Portfolio base are: a video transcoding pipeline (MediaConvert), a data lake (Glue + Athena + Lake Formation), an IoT ingestion pipeline (IoT Core + Kinesis), or an analytics platform (Redshift + QuickSight). Generic "we want more credits for our existing SaaS" submissions get silently downgraded to $0 because they double-count the Portfolio scope.
The most common $75K stack. Portfolio covers general infrastructure (ECS, Aurora, S3, CloudFront). Bedrock POC covers a specific AI feature with an eval methodology — typically a customer-support agent built on Claude Sonnet, a content-generation pipeline, or a retrieval-augmented internal Q&A.
Timeline: Portfolio approves at days 11–18. Bedrock POC takes an additional 4–7 days because the eval requirement adds a review step. Combined time-to-balance: 18–25 days.
Validity: Portfolio 24 months. Bedrock POC 12 months with a 6-month POC-completion checkpoint. If the team abandons the AI workload before the checkpoint, the unused Bedrock credits forfeit.
Risk profile: ~92% combined approval rate when the AI workload is well-specified.
The full-stack version. Portfolio covers general infrastructure. Build for Startups covers a distinct secondary workload (video pipeline, data lake, IoT ingestion, etc.). Bedrock POC covers a specific AI feature.
Who this fits: seed-stage startups with genuinely complex architecture — typically B2B SaaS with an analytics product, healthtech with a data-processing pipeline, fintech with a real-time risk-scoring engine alongside the core platform.
Timeline: all three records submitted within the same week. Combined time-to-balance: 18–25 days.
Risk profile: ~80% combined approval rate. The Build for Startups layer has the highest individual rejection rate (~15%) because the distinct-workload requirement is strictly enforced.
When this happens: the team has no AI workload, no distinct second workload, and the application scope is purely general infrastructure. The partner files Portfolio alone, lands at $75K, and stops.
Why this is rare: almost every seed-stage SaaS in 2026 has at least an AI feature somewhere (chatbot, recommendation, classification, summarization). When there is an AI feature, the Bedrock POC is the natural layer.
When to actually use this stack: teams in regulated industries that explicitly do not deploy AI yet (e.g., a healthtech compliance-tooling startup waiting for FDA pathway clarity before adding ML).
Five concrete inputs separate the three Portfolio landing points. Founders who understand these can self-assess where their application will land before they file — and either accept the realistic outcome or strengthen the weak inputs.
| Input | $50K floor | $75K mid | $100K ceiling |
|---|---|---|---|
| Projected monthly consumption | $1K–$2K | $3K–$4K | $5K+ |
| AWS account history | Greenfield | 0–6 months at $1K–$2K | 6+ months at $2K+ |
| Institutional vouch quality | Partner-mediated, no named VC | Named regional/vertical VC, partner-filed | Tier-1 VC OR VC with Portfolio Sub-Program direct |
| Workload composition | Single product, generic | Single product, specific services | Single product + distinct secondary workload |
| Typical layering pattern | Founders + Bedrock POC variant | Portfolio + Bedrock POC ($100K total) | Portfolio + Build for Startups + Bedrock POC ($150K total) |
Three founder situations make $75K Portfolio the honest right target. Three other situations suggest pushing for the $100K ceiling. Knowing which one you are in matters because over-targeting wastes a submission slot and under-targeting leaves credit on the table.
$75K is the right target if:
Push for $100K instead if:
The process for landing a $75K Portfolio is mechanically identical to the process for $100K Portfolio — the same ACE submission, the same partner-development-manager reviewer, the same wall-clock. The difference is purely in the outcome the reviewer assigns.
Hour 0: founder submits a CloudRoute inquiry. Company name, funding stage, AWS use case in one sentence. The form takes 90 seconds.
Hour 0–24: CloudRoute scores the inquiry and routes it to a vetted AWS partner — Advanced or Premier tier, with ACE submission rights, matched to your region (US, EU, MENA, APAC) and to the stack you are building (Bedrock-specialist, migration-specialist, etc.).
Hour 24–48: partner discovery call. ~30 minutes. The partner confirms eligibility, walks through what the application looks like for your situation specifically, and outlines the timeline. You decide whether to proceed.
Hour 48–72: if proceeding, founder spends ~30 minutes filling in the application worksheet — company info, AWS account ID, projected consumption, deck, the named investor (if any). Partner submits the ACE record within 24 hours of receiving inputs.
Day 11–18: Portfolio approval. AWS sends a notification to the company's AWS account email; credits appear in the billing console. The reviewer's assigned dollar amount is the line that matters — $50K, $75K, or $100K depending on the input variables discussed above.
Day 12–25: if a Bedrock POC layer was submitted alongside, it approves 4–7 days after Portfolio. Build for Startups (if submitted) approves at roughly the same window.
Across CloudRoute-routed engagements at the $75K Portfolio landing point, the average founder time investment is ~30 minutes for the application worksheet plus ~30 minutes for the partner discovery call. Total: ~1 hour. The wall-clock is 11–18 days for Portfolio alone, extending to 18–25 days if Bedrock POC or Build for Startups are layered.
The three Portfolio landing points compared on the dimensions that matter operationally.
| Variable | $50K (floor) | $75K (mid) | $100K (ceiling) |
|---|---|---|---|
| Most common profile | Seed-stage with partner-mediated vouch, modest projection | Seed-stage with named VC or accelerator brand, mid projection | Series-A with tier-1 VC or Portfolio-Sub-Program VC, full projection |
| Projected monthly AWS spend | $1K–$2K | $3K–$4K | $5K+ |
| AWS account history needed | None | Helpful, not required | Strongly favored |
| Typical stack composition | Founders + Bedrock POC OR Portfolio alone at floor | Portfolio + Bedrock POC ($100K total) | Portfolio + Build for Startups + Bedrock POC ($150K total) |
| Validity | 24 months (Portfolio) / 12 months (Founders variant) | 24 months (Portfolio) + 12 months (POC layer) | 24 months (Portfolio) + 12 months (each additive layer) |
| Wall-clock to first credit balance | 14–21 days | 11–18 days | 11–18 days |
| Risk of further downgrade | ~8% to lower outcome | ~12% to floor | ~22% to $75K, ~8% to $50K |
| Realistic 12-month coverage | 17–28 months of consumption | 18–24 months | 18–24 months |
Situation: Seed round of $4M led by a regional VC with strong vertical SaaS track record but no AWS Portfolio Sub-Program access. Company had been running on Vercel + Supabase for 14 months at ~$1.5K/month combined. Migrating to AWS for production scale and to enable a customer-support AI agent built on Claude Sonnet via Bedrock. Projected post-migration AWS spend at $3K/month general + $1.5K/month Bedrock inference once the agent rolled out.
What CloudRoute did: Routed within 19 hours to a US-East partner with Bedrock specialization and a track record of mid-range Portfolio approvals. Partner discovery call confirmed two distinct workloads (general SaaS infrastructure + the AI agent). Partner filed Portfolio ACE with the regional VC named as lead investor and $3K/month projection itemized across ECS Fargate, Aurora PostgreSQL, S3, and CloudFront. Filed Bedrock POC ACE separately for the customer-support agent with eval methodology documented.
Outcome: Portfolio approved at $75K on day 14 (mid-range outcome: regional VC vouch + moderate projection + greenfield AWS account → reviewer landed in the middle of the range). Bedrock POC approved at $25K on day 19. Total credits applied: $100K within 19 days. Founder time across the engagement: ~1.5 hours including discovery call. Customer cost: $0; CloudRoute commission paid by partner from AWS engagement funding.
engagement window: 19 days · founder time: ~1.5 hours · credits secured: $100K ($75K Portfolio + $25K Bedrock POC) · cost to customer: $0
CloudRoute routes seed-stage and accelerator-backed founders to AWS partners who file the Portfolio ACE submission, then layer Bedrock POC and Build for Startups on top. Customer pays $0; AWS funds the engagement.