You're a SaaS founder or engineer who has seen "AWS SaaS Factory" mentioned and wants to know whether it's worth your time. Your first action is cheap: open the AWS SaaS page and the SaaS Lens linked below, and decide whether your product needs AWS at all before you chase a program.
I'm an AI-augmented full-stack engineer, and my own products run on Firebase and Supabase, not AWS, so this isn't a hands-on program review. It's a fact-checked summary of what AWS publishes today, with the things I couldn't confirm marked as such. For the wider architecture picture, see the SaaS backend infrastructure guide.
What is AWS SaaS Factory, and is it still available?
AWS SaaS Factory is an AWS program offering software vendors guidance, reference architectures and expert access for building SaaS on AWS. Its old program page now redirects to AWS Marketplace, so confirm availability with AWS.
What AWS says it is
AWS's SaaS Journey Framework whitepaper describes the program as giving "direct access to technical and business content, best practices, solution architects, and SaaS experts" to help accelerate delivery of SaaS on AWS. The paper was published in October 2020 and now carries a banner saying it's for historical reference only, so some content may be outdated. The current SaaS on AWS page still references SaaS Factory as a way to access AWS experts with experience building and scaling SaaS applications, and points to the SaaS Journey Framework with four phases: design, build, launch and optimize.
What I couldn't confirm
On 30 September 2026 I requested the old program addresses, aws.amazon.com/partners/saas-factory/ and aws.amazon.com/partners/programs/saas-factory/, and both redirected to the AWS Marketplace partner page, which doesn't mention SaaS Factory. That means I can't confirm the current eligibility rules, how to apply, or whether it still costs nothing. Earlier versions of this post called it free. I've removed that claim because I couldn't verify it on an AWS page. If the program matters to you, ask your AWS contact or partner manager directly.
What AWS SaaS resources can you use without the program?
You can use the SaaS Journey Framework, the SaaS Lens in the Well-Architected Tool, AWS's tenant isolation guidance and a serverless reference solution on GitHub, all of which are public and need no program membership.
The four-phase journey
The Journey Framework's four phases are design, build, launch and optimize. AWS notes the framework isn't strictly linear, and some activities run at the same time. As a rough mapping of my own, not AWS's: design covers the problem sentence and scoping in a SaaS MVP plan, build is the first production version, launch is your first paying customers, and optimize is where cost and tenant tuning begin. Use it as a checklist to find gaps, which is how AWS describes it too, as a validation point you can use in whole or in part.
The SaaS Lens
The SaaS Lens for the Well-Architected Framework was published in April 2023 and focuses on designing, deploying and architecting multi-tenant SaaS workloads on AWS. It's available in the Lens Catalog of the AWS Well-Architected Tool. AWS's architecture fundamentals paper says that tool is available at no charge in the console, though that paper is also marked historical, so check the console before relying on it.
Tenant isolation
The most useful concept from the AWS material is that authentication isn't isolation. AWS's tenant isolation paper says a user "could be authenticated and authorized, and still access the resources of another tenant," and that isolation must be applied separately, using tenant context to limit access to resources. It describes dedicated stacks per tenant on one end and pooled resources with fine-grained policies on the other. The two ends of that range look like this:
| Model | How isolation is achieved | Trade-off |
|---|---|---|
| Dedicated resources per tenant | Whole stacks per tenant, with network or other coarse-grained policies blocking cross-tenant access | Stronger separation, more to run and pay for |
| Pooled resources | Tenants share resources, for example items in one DynamoDB table, with fine-grained policies controlling access | Cheaper to operate, and every policy has to be right |
This is a general principle, and it applies whether you're on AWS, Supabase or Firebase. My multi-tenant architecture guide covers the same trade-offs without tying you to one cloud.
A reference solution, with a warning
The serverless SaaS reference solution on GitHub shows a multi-tenant architecture built on serverless AWS services and the SaaS Builder Toolkit, with a control plane, tenant provisioning and a sample application. Its README is explicit that it's a reference, not production software: "You own the security of what you deploy." Read it to learn patterns, and don't deploy it as-is to hold customer data.
Is AWS the right place to build your SaaS MVP?
Choose AWS when your team has AWS skills, you need dedicated-tenant tiers or Marketplace distribution, and skip it for a first MVP where a managed backend like Supabase or Firebase gets you live faster.
That's my recommendation, based on stack trade-offs, not on an AWS study. The table shows how I'd weigh it:
| Situation | Better fit | Why |
|---|---|---|
| First MVP, small team, no AWS experience | Supabase or Firebase | Auth, database and storage come in one managed product, so there's less infrastructure to design |
| Customers demand dedicated infrastructure or data residency controls | AWS | You can build dedicated stacks per tenant |
| You plan to sell through AWS Marketplace | AWS | Marketplace listing is an AWS-native route |
| Existing AWS team and credits | AWS | You already pay the learning cost |
The Supabase vs Firebase comparison helps if you take the managed route.
A one-afternoon evaluation
If you're undecided, this sequence costs you a few hours and no commitment:
- Write down your tenancy requirement: will customers share infrastructure, or does any customer contract demand dedicated resources?
- Read the tenant isolation paper and note which model, dedicated or pooled, matches that requirement.
- Run the SaaS Lens questions against your draft architecture in the Well-Architected Tool, and note the gaps it flags.
- Price one month of your expected traffic using the vendors' own pricing pages, not a blog's figures.
- Compare that estimate with a managed backend's plan, including the hours your team would spend on operations.
The fifth step is the one teams skip. Infrastructure you run yourself costs engineering time as well as money, and on a small team that time is usually the scarcer resource. This is my judgment, not an AWS statistic.
What AWS costs to start
Prices change, so these are the figures AWS published on 30 September 2026. The AWS Lambda pricing page lists a free tier of one million requests and 400,000 GB-seconds per month, and a listed rate of $0.20 per million requests for x86, with rates varying by region. The AWS Free Tier page says new accounts on the Free plan get up to $200 in credits over 6 months, and the account closes after 6 months or when credits run out, whichever comes first. It also lists 30-plus services as always free within monthly limits. Free-plan credits and Lambda's free tier cover experiments. A real multi-tenant product still needs a cost model, so set budget alerts before you launch.
Where this leaves an early-stage product
I don't have verified SaaS Factory case studies, so I removed the percentage claims an earlier version of this post made about time to market and cost savings. AWS's own SaaS page carries a customer testimonial about halving time to market, but it's one company's quote and I can't generalize from it.
For an early-stage product, prove demand first, choose the simplest backend your team can run, and read the AWS material for the isolation ideas either way.
Questions to put to AWS before you plan around the program
Because I couldn't confirm the current terms, these are the questions I'd ask before building a plan around SaaS Factory:
- Is the program open to new applicants right now, and what are the eligibility rules?
- What does it cost, and are any engagements tied to spending commitments or AWS Partner status?
- Which engagements are still offered, such as architecture reviews, workshops or content access?
- Is there a named contact, and how long does onboarding take?
- Does anything in the program require you to build on a particular architecture or service?
Ask for the answers in writing, since program terms can change and a web page from 2020 won't tell you.
Which one of your customers would ask for dedicated infrastructure, and what would they be willing to pay for it? If you can't name one, the simpler backend is probably the right call for now.