Series-B is the funding stage where the credit conversation stops being about Activate Portfolio and starts being about migration credits and committed-spend discounts. The $150K Activate stack that funded 18–24 months at Series-A burn now funds 2–4 months at Series-B burn. The structural shift forces a different stack: MAP for migrations ($250K–$500K), EDP for committed-spend discounts (10–25% on $100K+/year), PPA for negotiated per-service rates. This page walks through the Series-B-specific credit landscape, the math behind the shift, and the AWS account team relationship that becomes the primary engagement at this stage.
Series-A and Series-B sound adjacent in fundraising terms but they're distinct in AWS economic terms. The credit programs that dominate at Series-A — Activate Portfolio, Build for Startups, Bedrock POC — cease to be the primary funding mechanism at Series-B. The reason is not eligibility; it's scale.
At Series-A, the typical AWS spend is $5K–$25K/month. A $100K Activate Portfolio credit pool funds 4–20 months of that consumption. Founders treat credits as a meaningful runway extension. The math is straightforward: get the credits, burn them, move to paid status when they expire.
At Series-B, the typical AWS spend is $50K–$200K/month. The same $100K Portfolio funds 2–4 weeks of consumption at the high end, 2 months at the low end. The credit-as-runway-extension framing collapses. Founders at Series-B don't think about credits as covering operating costs — they think about credits as covering discrete projects (migrations, new feature builds, AI POCs) while the bulk of the AWS bill gets reduced through discount programs.
AWS's program design reflects this. The Activate program is calibrated for pre-seed through Series-A startups; the documentation, the application process, and the credit ceilings all assume monthly spend under $25K. The Migration Acceleration Program (MAP) is calibrated for migrations with $300K+ total cost — exactly the scale that Series-B companies routinely encounter when moving off Heroku Enterprise, GCP, or Azure. The Enterprise Discount Program (EDP) is calibrated for sustained spend at $100K+/year — exactly the scale that Series-B companies hit in normal operation.
The result: a Series-B founder who routes the conversation through Activate Portfolio is using a Series-A tool for a Series-B problem. The dollars don't scale. The Series-B-appropriate stack is MAP + EDP + PPA, with residual Activate-family credits (Bedrock POC for new AI workloads, occasional Build for Startups for discrete projects) as add-ons rather than primary instruments.
This page is structured around that shift. The first half covers what remains of the credit conversation at Series-B (the residual Activate stack plus MAP). The second half covers the bill-discount mechanics (EDP and PPA) that become the dominant economic instrument.
Three economic instruments dominate the Series-B AWS conversation. Each handles a distinct scope; they layer rather than compete. Understanding which instrument addresses which spend category is the threshold question for any Series-B AWS planning.
The first instrument is MAP — the Migration Acceleration Program. MAP funds the migration itself, calibrated to subsidize 25% (Mobilize phase) to 50% (Migrate phase) of correctly-tagged AWS resource cost during the migration. For a Series-B migration with $1M+ total cost, MAP routinely returns $250K–$500K of credits across the migration phases. The MAP credits cover migration-window consumption; they don't cover steady-state post-migration spend.
The second instrument is EDP — the Enterprise Discount Program. EDP is a committed-spend discount: the company commits to a minimum annual AWS spend on a multi-year contract; AWS returns a tiered discount on that committed spend. Typical EDP eligibility kicks in at $100K+/year of sustained spend. Discounts range from 10% on the entry tier to 25% on larger committed-spend tiers. EDP is permanent in the sense that it doesn't expire — every dollar spent during the contract earns the discount.
The third instrument is PPA — the Private Pricing Agreement. PPA handles per-service negotiation when EDP's flat-discount structure doesn't fit. Common PPA cases at Series-B: Bedrock inference at scale (where per-token rates can be negotiated below list pricing for committed volume), OpenSearch for large search workloads, EC2 reserved instances at unusual scale, S3 for large data lakes. PPA is layered on top of EDP — it doesn't replace the EDP discount; it negotiates additional service-specific rate reductions.
Residual Activate-family instruments still apply. Bedrock POC funding ($10K–$75K at Series-B scale) covers new AI workloads. Build for Startups ($25K) occasionally covers discrete projects that don't fit the migration or new-AI framing. The $5K Founders tier and $25K partner-filed tier from Activate are mostly irrelevant — Series-B spend exhausts them in days, not months.
The AWS account team relationship is the fourth element, less an instrument than a delivery channel. At Series-A the partner files the credit application and AWS reviews it administratively. At Series-B the AWS account team — typically a Customer Solutions Manager plus an Account Manager plus solutions architects — runs the EDP and PPA negotiations directly with the company. The partner remains valuable for MAP delivery and any residual credit applications, but the EDP/PPA conversation is direct between the company and AWS.
Most Series-B companies have already exhausted Activate Portfolio at Series-A. The standard arrival position is "$100K Portfolio credits consumed, possibly some Build for Startups balance remaining, possibly some Bedrock POC balance remaining." The Series-B credit conversation starts from that depleted state.
The Activate Portfolio credit pool issues against a 24-month validity window. A Series-A company that secured $100K Portfolio in month 0 of Series-A and spends $8K/month on AWS will exhaust the balance somewhere around month 12–14. By the time the same company raises Series-B (typically months 18–24 post-Series-A), the Portfolio balance is gone.
Build for Startups credits ($25K) issue against a 12-month window. They typically exhaust within the first year of Series-A. By Series-B, the Build for Startups balance is gone and not eligible for renewal — Build for Startups is a one-time per-company instrument, not a recurring grant.
Bedrock POC credits ($10K–$50K at Series-A) issue against a 12-month window with a 6-month POC completion checkpoint. They typically exhaust during the POC window. A Series-B company that ran a Bedrock POC at Series-A may have residual balance if the POC underran, but the typical position is exhausted.
This is the structural reason the Series-B credit conversation centers on MAP and bill discounts rather than fresh Activate applications. The Activate pools are largely spent; the remaining instruments are migration-funded credits (which scale to Series-B migration sizes) and committed-spend discounts (which scale to Series-B sustained spend). Asking AWS for another $100K of Portfolio credits at Series-B is structurally non-fit — the program isn't designed for it.
There are exceptions. A Series-B company that was on AWS at pre-seed but only became eligible for Portfolio at Series-A (because Portfolio requires institutional funding) may have received Portfolio late in the Series-A cycle and still have residual balance entering Series-B. A Series-B company that migrated late from another cloud may have used Portfolio for the migration-window consumption only and have balance remaining. These cases are minority — the modal Series-B position is Activate-exhausted.
For the residual credit conversation, the meaningful instrument at Series-B is a fresh Bedrock POC application for any new AI workload that wasn't scoped during Series-A. Bedrock POC awards at Series-B scale can reach $50K–$75K because projected Bedrock spend is materially larger; some Series-B Bedrock POCs reach $75K when the partner can argue substantial scale. Build for Startups occasionally fits a discrete project scope, particularly for FinTech compliance work, HealthTech HIPAA platform builds, or MediaTech transcoding pipelines.
The Generative AI Accelerator is calibrated for pre-Series-B AI startups — the cohort selection criteria favor companies that are still building toward their first production AI revenue, not companies that have already scaled AI to material spend. Series-B AI startups that did not enter the Accelerator during Series-A typically cannot enter at Series-B. The path for Series-B AI workloads is the standard stack (Bedrock POC + EDP for Bedrock-included spend) plus PPA for per-token rate negotiation at scale.
MAP at Series-B scale operates on the same Assess / Mobilize / Migrate phase structure as smaller migrations, but the absolute dollar values are materially larger. The 25% Mobilize and 50% Migrate calibrations apply to phase costs that routinely cross $500K each, which means MAP credit issuance routinely crosses $250K total per engagement.
The typical Series-B migration scope: a 40–80 engineer team moving off Heroku Enterprise ($30K–$100K/month source spend), or off GCP/GKE ($50K–$200K/month), or off Azure ($50K–$150K/month), or off a substantial on-prem datacenter footprint. The total migration cost — engineering labor, partner labor, AWS parallel-run cost, data transfer, post-migration optimization — routinely lands in the $1M–$3M range.
MAP credit issuance scales with that cost. A migration with $200K Mobilize phase cost returns $50K in Mobilize credits (25% calibration). A migration with $600K Migrate phase cost returns $300K in Migrate credits (50% calibration). Add the $25K Assess credit at the front and total MAP issuance lands at $375K for the engagement. Larger migrations push past $500K total MAP issuance; the cap is set by the partner-engagement scope rather than a fixed per-customer ceiling.
The Mobilize phase at Series-B scale is structurally different from Mobilize at smaller scale. A Series-B Mobilize phase typically covers 10–20 microservices migrating from the source platform, plus a primary database tier, plus the observability and CI/CD infrastructure that supports them. The pilot scope is bigger because the eventual Migrate scope is much bigger — Mobilize has to prove the playbook at a meaningful fraction of the target architecture.
The Migrate phase at Series-B scale runs 16–36 weeks of active cutover work. Workloads move in waves: stateless services first, then stateful services with cutover windows, then the long-tail of less-critical batch and reporting workloads. Each wave consumes tagged AWS resources that accrue toward MAP credit calculation. The tagging discipline that protects credit issuance at smaller scale becomes operationally critical at Series-B scale because the absolute dollars at risk from tagging drift can exceed $50K.
Partner economics also shift at Series-B scale. The Migration Competency Partners that take on $1M+ migrations are typically larger consulting organizations with dedicated AWS migration practices, not solo practitioners. The partner-labor cost is real (paid by AWS via the MAP partner-funding mechanism, not by the customer), and the partner's margin depends on delivering the migration on the projected scope and timeline. Partners take on Series-B migrations selectively, prioritizing companies with clear architectural targets and engineering bandwidth to participate in the cutover.
EDP is the most economically significant instrument in the Series-B AWS toolkit. Unlike credits, which are bounded one-time pools that expire, EDP discounts compound across the contract term. A Series-B company spending $500K/year on AWS at an 18% EDP discount returns more value over the contract than the entire $150K Activate stack returned at Series-A. Understanding the EDP mechanics is the first economic priority at Series-B.
EDP is structurally a committed-spend contract between the customer and AWS. The customer commits to a minimum annual AWS spend across the contract term (typically 1–5 years; 3 years is the modal contract length). In exchange, AWS applies a discount on the committed-spend tier. The discount rate scales with the committed amount: higher commitments earn higher discounts.
Typical EDP discount tiers in 2026: 10% discount at the entry tier (typically $100K/year committed); 12%–15% at the mid tier (typically $250K–$500K/year committed); 18%–22% at the upper tier (typically $1M+/year committed); 25% at the strategic tier (typically $5M+/year committed or accompanied by strategic-account positioning). The exact tier thresholds vary by negotiation; the percentages are typical ranges from CloudRoute's observed engagements.
The compounding math is what makes EDP qualitatively different from credits. A $500K/year AWS spend at 18% EDP discount saves $90K/year. Across a 3-year contract, the cumulative saving is $270K. Across a 5-year contract, $450K. This is direct bill reduction — the customer's AWS invoice arrives at the discounted rate; the saving doesn't require credit-balance tracking or expiration management.
EDP eligibility filters on sustained spend rather than fundraise stage. A Series-B company that crosses $100K/year of AWS spend (about $8.3K/month) becomes EDP-eligible. Most Series-B companies cross this threshold within the first quarter post-Series-B raise; many cross it during Series-A operations. The fundraise stage matters for AWS account team assignment and for the negotiation posture, but the EDP-mechanics gate is sustained spend.
The negotiation process is direct AWS, not partner-filed. The AWS account team — Account Manager plus Customer Solutions Manager plus the EDP specialist — runs the conversation. The customer's side is typically the CFO or VP of Engineering (sometimes both), plus the FinOps or platform engineering lead who can validate the projected spend. The negotiation runs 4–10 weeks from initial discussion to signed contract; the contract takes effect on a designated start date and the discount begins applying from that date forward.
The honest math for Series-B founders: at $300K/year of AWS spend, EDP at 12% returns $36K/year (~$108K across 3 years). At $500K/year, EDP at 18% returns $90K/year (~$270K across 3 years). At $1M/year, EDP at 22% returns $220K/year (~$660K across 3 years). At $2M/year, EDP at 25% returns $500K/year (~$1.5M across 3 years). The discount value scales faster than the Activate stack ever could.
EDP is a commitment. The customer commits to the minimum annual spend; if actual spend falls short, the customer typically owes the shortfall to AWS. This is the operational risk of EDP and the reason the commitment level matters more than the discount percentage in the initial negotiation.
A Series-B company that commits to $500K/year and actually spends $400K/year owes the $100K shortfall (or some negotiated equivalent). The mitigation is conservative commitment sizing: project the floor of expected spend across the contract term, commit to that floor, and let actual spend overage simply earn additional discount at the marginal rate.
AWS understands this risk and the EDP specialists typically guide customers toward commitment levels that the customer is likely to meet. But the conservative-commitment posture is the customer's responsibility — AWS's incentive is to maximize the commitment, not minimize the shortfall risk.
EDP is a contract-level discount that applies to all eligible AWS spend. Reserved Instances (RIs) and Savings Plans (SPs) are service-level discount mechanisms that apply to specific compute commitment (EC2 instances, Fargate capacity, Lambda compute). EDP and RIs/SPs stack — a customer with an 18% EDP discount and a 30% Savings Plan discount on EC2 gets the EDP discount applied to the already-Savings-Plan-discounted bill.
The stacking math means a Series-B company optimizing aggressively can land at a 40%+ effective discount on the EC2 portion of the bill (Savings Plans + EDP combined). The bulk of the credit-and-discount value at Series-B comes from this stacking, not from any single instrument.
PPA enters the conversation when even the stacked discount structure doesn't reach the customer's cost-target on a specific service. The most common PPA case at Series-B is Bedrock inference at substantial scale, where per-token rate negotiation can reduce effective cost by 20%–40% below list pricing even after EDP discount.
EDP is a flat discount on committed spend. It works well for distributed AWS workloads where the spend is spread across many services. When the spend concentrates heavily on one or two specific services, PPA — Private Pricing Agreement — handles the per-service rate negotiation that EDP can't. PPA is the layered instrument that addresses concentrated high-spend cases at Series-B scale.
PPA is a negotiated rate agreement on specific AWS services. Unlike EDP's flat discount on committed spend across services, PPA reduces the per-unit rate on specific services for a committed volume. Common PPA cases at Series-B in 2026:
Bedrock inference at scale. A Series-B company committing to $50K+/month of Bedrock spend on a specific model family (typically Claude, sometimes Llama or Mistral) can negotiate a PPA that reduces the per-million-token rate on input and output tokens below list pricing. Effective rate reductions of 15%–30% are typical, on top of any EDP discount that would otherwise apply to Bedrock spend.
OpenSearch at scale. Companies running large search workloads (typically $30K+/month OpenSearch spend, often supporting product search or log search) can negotiate PPA on OpenSearch cluster hours or storage volumes. Effective rate reductions of 10%–25% are typical.
Large EC2 fleets with specific instance-type commitment. PPA on EC2 typically appears when the company's compute commitment is concentrated on a specific instance family (e.g., M7i or C7g for general-purpose compute, P5 or H100 for AI inference, R7i for memory-bound workloads) at substantial scale. PPA on EC2 stacks with Savings Plans and EDP, often pushing effective rates 30%–45% below list.
S3 for large data lakes. Companies running multi-petabyte S3 footprints can negotiate PPA on storage class rates and request charges. The discounts tend to be modest (5%–15% on storage class rates), but at petabyte scale the absolute dollars are material.
The PPA negotiation is direct AWS, similar to EDP. The AWS account team coordinates with the service-team product manager (Bedrock PM, OpenSearch PM, EC2 PM as relevant) to scope the rate reduction. The contract attaches to the customer's overall AWS agreement; the rate applies for the contract term.
PPA at Series-B is more often a "yes when applicable" than a default. EDP covers the bulk of the bill-reduction math at Series-B; PPA addresses the specific concentrated-spend cases that EDP can't. CloudRoute's observed pattern: roughly 30% of Series-B engagements involve a meaningful PPA layer; the other 70% are well-served by EDP alone.
Build for Startups and the partner-labor-equivalent instruments that AWS funds for specific build engagements still apply at Series-B, particularly for vertical-specific work that requires partner expertise. The dollar amounts at Series-B can reach $100K–$300K of partner labor funded by AWS for the right engagement scope.
Build for Startups at Series-B is typically applied to vertical-specific build engagements that the company couldn't reasonably scope in-house. The standard cases:
FinTech compliance work. A Series-B FinTech building toward SOC 2, PCI DSS, or specific regulatory frameworks (like the EU's DORA or the US's Reg DD for digital deposit products) routinely engages AWS partners with FinTech competency to build the compliance telemetry, audit logging, and IAM-policy frameworks that the regulator expects. AWS funds the partner labor through a combination of Build for Startups credits and dedicated partner-engagement funding; the customer pays $0 for the partner labor, and the total funded labor value can reach $100K–$200K.
HealthTech HIPAA platform builds. A Series-B HealthTech building a HIPAA-compliant clinical or patient-data platform engages AWS partners with HealthTech competency to build the BAA-eligible infrastructure, the audit logging, and the data-handling patterns. The partner labor is typically larger than FinTech because the HealthTech compliance surface is broader; total funded labor value can reach $150K–$300K.
MediaTech transcoding pipelines. A Series-B media company building a video transcoding pipeline (typically on MediaConvert, MediaLive, or a custom EC2-based encoding farm) engages partners with media competency to build the pipeline. Total funded labor can reach $100K–$200K depending on the encoding scope and ingestion volume.
The instrument structure for these engagements: Build for Startups credits ($25K) cover the AWS resource cost during the build; AWS partner-engagement funding covers the partner labor. The customer's out-of-pocket cost is $0; the partner is paid by AWS through the program-funding mechanism. The engagement deliverable is the built infrastructure plus a co-branded readiness assessment document.
CloudRoute routes Series-B build engagements to partners with the relevant vertical competency. The match criteria include AWS Competency (FinTech, Healthcare, Media), prior engagement count in the vertical, and regional proximity (US, UK, EU, MENA, etc.). The engagement runs 8–24 weeks depending on scope; the credits and partner-labor funding issue across the engagement on milestone delivery.
Bedrock POC funding at Series-B operates on the same partner-filed mechanism as at Series-A, but the projected Bedrock spend is materially larger, which pushes POC awards toward the upper end of the calibrated range. Series-B Bedrock POCs routinely reach $50K and occasionally $75K when the partner can argue substantial scale.
The Bedrock POC funding program calibrates award amounts against projected Bedrock spend. At Series-A, projected Bedrock spend during the POC window is typically $5K–$25K/month, which calibrates to a $10K–$50K POC award. At Series-B, projected Bedrock spend during the POC window is typically $20K–$80K/month, which calibrates to $25K–$75K POC awards.
The Series-B POC scope is typically more architecturally substantial than the Series-A scope. A Series-A Bedrock POC often covers a single feature (a chatbot, a document summarization endpoint, a classification model); a Series-B Bedrock POC often covers a full product surface (a customer-facing AI product line, a workflow automation layer across multiple internal tools, a multi-tenant inference platform). The eval methodology is correspondingly more substantial — N=2,000 examples is typical at Series-B versus N=500 at Series-A.
The path to the upper $75K POC award: a clear customer or product surface with material projected revenue (not exploratory); a defined model family (typically Claude Sonnet or Claude Opus for the high-quality tier, with Claude Haiku or Nova Lite for the cost-efficient tier in a routing setup); a rigorous eval harness with held-out test sets and weekly accuracy reporting; an explicit projected budget in the $30K–$80K/month range during the POC window; a 60–90 day POC window with a clear go/no-go decision criterion.
CloudRoute's observed pattern at Series-B: ~70% of well-scoped Bedrock POC applications approve in the $50K–$75K range. The remaining 30% approve at the floor ($25K) or are downgraded to standard Activate-family credits when the POC scope doesn't clear the partner-filed bar.
Bedrock POC credits are Bedrock-earmarked. They cover Bedrock inference, embeddings, and the directly-supporting infrastructure (OpenSearch for vector search, S3 for prompt logs, Lambda for orchestration glue). They don't cover unrelated EC2 or RDS spend. At Series-B, the typical Bedrock POC credit pool exhausts during the 60–90 day POC window, transitioning the workload to paid Bedrock consumption at the EDP-discounted (and potentially PPA-negotiated) rate post-POC.
At Series-A, the partner is the primary AWS engagement channel — they file ACE records, they handle the credit application, they interact with AWS reviewers. At Series-B, the AWS account team typically becomes the primary engagement for EDP, PPA, and the broader strategic conversation, while the partner remains the delivery channel for MAP and specific build engagements.
AWS assigns account teams based on company spend and strategic profile. The default Series-B account team typically includes: an Account Manager (commercial owner of the customer relationship), a Customer Solutions Manager (CSM, technical and operational owner), and one or more solutions architects (deep-dive technical advisors for specific service categories). Larger Series-B engagements add specialists: an EDP specialist for the discount-negotiation track, a Bedrock specialist for AI workloads, a security specialist for compliance work.
The relationship begins to form once sustained spend crosses roughly $20K/month, which most Series-B companies hit within a quarter or two of the Series-B raise. The account team typically reaches out proactively — AWS's commercial pipeline systems flag companies crossing the threshold and route account team coverage. Companies that don't hear from an account team within a quarter of crossing the threshold can request coverage; the request routes through the AWS partner network or the company's existing AWS contacts.
The account team conversation centers on EDP and PPA. Initial conversations typically cover: the customer's projected AWS spend trajectory (informed by the Series-B operating plan); the workload mix (which services dominate the spend); the migration status (any ongoing migrations from other clouds or platforms); the AI strategy (any Bedrock or Sagemaker workloads in scope); the multi-region or compliance posture (any geographic distribution or regulated-vertical requirements). The output of these conversations is the EDP/PPA proposal that the customer then negotiates.
Partner routing remains valuable at Series-B for specific engagement types. MAP migrations are partner-led even when the EDP conversation is direct AWS — the Migration Competency Partner runs the migration, files the milestones, and earns the MAP credit issuance for the customer. FinTech/HealthTech/MediaTech build engagements are partner-delivered because the vertical competency requires partner expertise. Bedrock POC applications remain partner-filed because the POC scope requires the partner's eval-methodology framing.
The honest framing for Series-B founders: the AWS account team is the primary economic conversation at this stage. Partner routing is valuable for delivery (MAP, vertical builds, Bedrock POC) but the discount-negotiation conversation is direct AWS. Founders who route all AWS conversations through partners at Series-B miss the EDP/PPA mechanics that handle the bulk of the bill-reduction math.
The Series-B economic position is best understood through a composite example. Across CloudRoute's routed Series-B engagements, a typical shape is a B2B SaaS spending $400K/month on AWS after completing migration off Heroku Enterprise. Here's how the stack composes.
The starting position: 60-engineer B2B SaaS, Series-B with $40M raised, completing migration off Heroku Enterprise. Pre-migration Heroku spend was $85K/month (compute dynos, Heroku Postgres at 4.8 TB, Heroku Redis Premium, Apache Kafka cluster). Post-migration AWS spend stabilized at $400K/month — higher in absolute terms than Heroku because the AWS scope includes parallel-run cost during migration, broader multi-region deployment for latency reduction, and a new AI feature line built on Bedrock.
The MAP credits: total MAP credit issuance reached $250K. Assess phase returned $25K at week 4 of the engagement. Mobilize phase (covering 8 microservices and the Postgres tier as pilot) consumed approximately $300K of tagged AWS resources over 10 weeks; MAP returned $75K at 25% calibration. Migrate phase (covering the remaining workloads across 18 weeks) consumed approximately $300K of tagged AWS resources; MAP returned $150K at 50% calibration. Cumulative MAP issuance: $250K, issued across weeks 4 to 24 of the engagement.
The Bedrock POC: $50K Bedrock POC funded the new AI feature line during a 75-day POC window. The POC covered a customer-facing AI product surface (workflow automation across 12 internal tools, with routing between Claude Sonnet for complex queries and Claude Haiku for high-volume simple queries). Eval methodology: N=2,400 examples weekly, accuracy reporting to the product team, go/no-go decision at day 75. The POC graduated to production; post-POC Bedrock spend stabilized at $35K/month, transitioning to EDP-discounted and PPA-negotiated rate post-POC.
The EDP contract: 3-year EDP signed in month 8 of the engagement, after the migration stabilized post-migration spend at $400K/month. Committed spend: $15M total across 3 years ($5M/year, slightly above the $4.8M/year projected baseline to leave headroom for growth). Negotiated discount: 18% on committed-tier spend. Annual saving from EDP: $900K/year ($5M × 18%). Cumulative saving across the 3-year contract: $2.7M.
The PPA layer: PPA negotiated on Bedrock inference for the AI feature line. Bedrock spend at scale crossed $35K/month, qualifying for PPA conversation. Negotiated rate reduction: 22% off list pricing on Claude Sonnet input and output tokens for committed volume. Annual saving from PPA layer (on top of EDP): approximately $90K/year on the Bedrock-specific portion of the bill.
The total economic position: $250K MAP credits (one-time, deployed across migration window). $50K Bedrock POC credit (one-time, deployed during POC). $2.7M EDP savings across 3-year contract. ~$270K PPA savings on Bedrock across 3-year contract. Cumulative 3-year economic value: ~$3.27M. The Activate-family credits that defined the Series-A conversation are absent from this picture — the company exhausted them during Series-A operations.
The cost to the customer: $0 for partner routing, $0 for MAP delivery, $0 for Bedrock POC application. The EDP and PPA contracts represent customer-side commitments on spend, but no out-of-pocket fees beyond the AWS bill itself (now at discounted rates).
The honest comparison between the two adjacent funding stages.
| Variable | Series-A stack | Series-B stack |
|---|---|---|
| Dominant instruments | Activate Portfolio + Build for Startups + Bedrock POC | MAP + EDP + PPA (Activate-family residual) |
| Typical AWS spend | $5K–$25K/month | $50K–$200K/month |
| Activate Portfolio role | Primary credit source ($100K) | Already exhausted in most cases |
| MAP role | Rare (small migrations don't fit MAP economics) | Dominant for migration scope; $250K–$500K typical |
| EDP eligibility | Below threshold ($100K/year) | Above threshold; primary bill-reduction instrument |
| PPA relevance | Not applicable | Layered on EDP for concentrated high-spend services |
| Bedrock POC ceiling | $10K–$50K | $25K–$75K (larger projected scale) |
| Build for Startups | $25K, often used for discrete projects | $25K + partner-labor funding up to $300K for vertical builds |
| Generative AI Accelerator | In scope (pre-Series-B program) | Window closed for established companies |
| Primary AWS engagement channel | Partner via ACE program | AWS account team (direct) + partner for delivery |
| Credit-value framing | $100K–$150K total credits | $250K–$500K MAP + $1M–$2.7M EDP savings |
| Time horizon | Credits fund 18–24 months operating cost | EDP discount compounds across 3-year contract term |
| Cost to customer | $0 | $0 |
Situation: B2B SaaS, Series-B with $40M raised, 60-engineer team across US (primary) and EU (secondary) regions. Completing migration off Heroku Enterprise at $85K/month pre-migration source spend. Heroku scale limits hit on Postgres at 4.8 TB and Heroku Connect for the Salesforce integration. Migration driven by scale ceiling and unit-economics pressure (Heroku Enterprise pricing unfavorable above $50K/month). Total projected migration cost: $1.2M (engineering labor, partner labor, AWS parallel-run, data transfer, multi-region setup). Post-migration AWS spend projection: $400K/month, including a new AI feature line on Bedrock projected at $35K/month.
What CloudRoute did: Routed within 22 hours to a Migration Competency Partner with Heroku-to-AWS expertise plus multi-region deployment experience. Discovery confirmed MAP eligibility (migration scope $1.2M, post-migration AWS projection $400K/month, 60-engineer team capable of supporting tagging discipline and Migration Hub maintenance). Partner registered the MAP engagement in Migration Hub on day 4; tagging convention installed across both US and EU regions before any Mobilize-phase resource was spun up. Bedrock POC application filed in parallel ($50K against projected $35K/month Bedrock spend during 75-day POC window). AWS account team assigned in month 3 once sustained AWS spend crossed $50K/month; EDP conversation initiated by the account team in month 6 of the engagement. PPA conversation on Bedrock inference initiated in month 8 once Bedrock spend stabilized at scale.
Outcome: Total economic position across the engagement and the resulting EDP contract: $250K MAP credits issued across weeks 4–24 (Assess $25K, Mobilize $75K at 25% of $300K Mobilize phase tagged consumption, Migrate $150K at 50% of $300K Migrate phase tagged consumption). $50K Bedrock POC issued during the 75-day POC window, transitioning to paid Bedrock at PPA rate post-POC. 3-year EDP contract signed in month 8 at $5M/year committed spend with 18% discount, returning $900K/year in direct bill reduction ($2.7M cumulative across the contract). PPA layer on Bedrock at 22% rate reduction on Claude Sonnet committed volume, returning approximately $90K/year ($270K cumulative). Cumulative 3-year economic value: ~$3.27M. Migration complete week 26; final Heroku Enterprise contract terminated week 30. Post-migration AWS spend stabilized at $398K/month, within 1% of projection.
engagement window: 30 weeks · founder time: ~24 hours total across migration + EDP/PPA negotiation · cumulative 3-year value: ~$3.27M · cost to customer: $0
CloudRoute routes Series-B companies to Migration Competency Partners for MAP delivery and coordinates with AWS account teams for the EDP/PPA conversation. MAP credits arrive across migration phases over 6–24 weeks; EDP contracts sign in 4–10 weeks; PPA layers on top within the same window.