Every founder eventually asks the same question: which cloud will pay for my first year? AWS, Google Cloud, and Microsoft Azure all run startup programs, and all of them advertise large numbers. The headline figures are real. They are also not directly comparable.
Comparing AWS vs Google Cloud vs Azure startup credits fairly means looking past the maximum. Each program has different tiers, eligibility rules, expiry dates, and limits on what the credits cover. A $350,000 ceiling means little if you do not qualify for it, and a $5,000 credit may be exactly what a two person team needs.
This guide lays the programs side by side in plain language. You will learn:
- What startup cloud credits are, and what they never are
- How each program works, from the entry tier to the top tier
- A comparison of amounts, eligibility, expiry, and AI coverage
- How to choose based on your stage, your stack, and your investors
- A step by step application plan, plus mistakes to avoid
Amounts and rules change often, and each provider states them on its own startup pages. The numbers here are the latest published figures at the time of writing. Confirm them in the official program terms before you build a plan around them.
What Are Startup Cloud Credits?
Startup cloud credits are promotional balances from a cloud provider that offset eligible usage charges for early-stage companies. AWS, Google Cloud, and Microsoft Azure each run programs with different amounts, eligibility rules, and expiry dates, and credits are not cash. Unused credits expire, and you pay standard rates afterward.
The providers have an obvious reason to offer them. Startups that build on a cloud tend to stay there as they grow. For you, the benefit is real runway, as long as you understand the rules.
Three ideas apply to every program:
- Credits are not cash. You cannot withdraw them, and they only offset eligible services.
- Credits expire. Each program sets its own validity period, and none of the three lets you stretch it freely.
- Credits end in a normal bill. When they run out, your card or invoice takes over at standard rates.
AWS’s promotional credit terms state the pattern clearly: credit is valid for a limited time, and credit not used before its expiration date is forfeited. The other two providers publish similar terms on their own startup pages.
AWS Activate: What You Get
AWS Activate is the AWS startup program. Eligible startups apply through the AWS Activate Credits page, and AWS describes the full process in its application guide.
Tiers and Amounts
- Activate Founders: $1,000 in credits for self-funded startups, and AWS says select startups may qualify for up to $5,000 over time. No Org ID is needed.
- Activate Portfolio: for startups connected to an Activate Provider, such as an accelerator or investor. Credits depend on the provider and your stage. AWS’s credits page lists up to $100,000, while its June 2026 guide lists up to $200,000.
- Above $200,000: AWS says additional credits are available for AI startups ready to scale beyond Activate’s tiers, through your AWS Account Manager.
Eligibility
AWS lists these basics: a pre-Series B startup, a company website or profile, a company founded in the last 10 years, and an AWS account on the Paid Tier Plan. Portfolio applicants also need an Org ID from their provider and must be new to Activate credits or requesting more than before.
What the Credits Cover
AWS says Activate credits can offset more than 200 services, and they can be redeemed on third party foundation models in Amazon Bedrock from providers such as Anthropic, Meta, Cohere, Mistral AI, and others. Credits do not cover items such as AWS Marketplace purchases (with the Bedrock model exception), Route 53 domain registration, or upfront fees for Savings Plans and Reserved Instances.
Timing and Expiry
AWS says reviews usually take 7 to 10 business days. Activate credits typically expire within one to two years, depending on the package, and AWS says it cannot extend the expiration date. AWS emails members 60 and 30 days before expiry.
Google for Startups Cloud Program: What You Get
Google’s program has three main tracks. The figures below come from Google’s published program pages.
Tiers and Amounts
- Start tier: about $2,000 in credits. Google describes it for startups founded in the past two years that have a working MVP, a clear business model, and plans to seek venture funding soon.
- Scale tier: up to $200,000 over two years for venture-funded startups. Google says it is designed for pre-seed and seed rounds within the last five years, and Series A rounds within the last 12 months.
- Scale AI tier: up to $350,000 over two years for AI-first startups. Google says year one covers 100% of eligible usage up to $250,000, and year two covers 20% up to an additional $100,000.
Eligibility Details
Google states that non-venture funding, such as private equity, bank loans, government innovation programs, grants, crowdfunding, or friends and family money, does not qualify a startup for the Scale tier. The Start tier requires that you have not already received Google Cloud credits beyond the free trial.
What the Credits Cover
Google says program credits can be used on Google Cloud services, including Gemini and Gemma models and Vertex AI open models. Third party models are billed directly and are not covered by the standard credits, though AI tier startups may qualify for an additional $10,000 for partner models in Model Garden.
Extra Perks
Google lists perks that include 12 months of Google Workspace Business Plus for new signups, a monthly Google Maps Platform credit for one year, Google Skills training credits, and a Startup Success Manager for some tiers. Program materials also mention Enhanced Support credits for AI tier startups, so check the terms for the current amount.
Microsoft for Startups Founders Hub: What You Get
Microsoft’s program is built around four levels: Ideate, Develop, Grow, and Scale. Microsoft says you graduate to higher levels, and unlock more Azure credits, as your company and Azure usage grow.
Levels and Amounts
- Level 1, Ideate: up to $1,000 in Azure credits to start building.
- Level 2, Develop: up to $5,000 total after business verification, based on Microsoft’s startup credit offer.
- Level 3, Grow: $25,000 in Azure credits, per Microsoft’s benefits guide.
- Level 4, Scale: up to $150,000 in Azure credits.
Microsoft also runs an Investor Offer. If your investor belongs to Microsoft’s Investor Network, they can give you a referral code that unlocks a larger credit amount. Microsoft’s support answers say the startup must be investor-affiliated, have received fewer than $350,000 in lifetime Azure credits, and not be Series C or later.
What the Credits Cover
Microsoft says Azure credits can be used toward any Azure product, including Azure OpenAI Service. Founders Hub also bundles benefits such as Visual Studio, GitHub, and Microsoft 365 seats at some levels, though those bundled benefits have their own terms.
Timing and Expiry
Microsoft’s support answers describe short windows at the entry levels. The initial $1,000 is valid for 90 days after redemption. Credits unlocked through business verification expire 180 days after verification, and Microsoft says these periods cannot be extended. For the Investor Offer, support answers say credits must be activated within 90 days and are then available for two years.
Because the Azure entry credits are shorter than the others, plan your first project before you redeem them.
AWS vs Google Cloud vs Azure Startup Credits: Side by Side
This table puts the three programs next to each other using the figures published by each provider. Treat it as a map, and confirm every number before you apply.
| Factor | AWS Activate | Google for Startups Cloud | Microsoft Founders Hub |
| Entry credits | $1,000 (Founders), select startups up to $5,000 over time | About $2,000 (Start tier) | Up to $1,000, then up to $5,000 after verification |
| Mid level | Portfolio credits through a provider | Scale tier, up to $200,000 over two years | Level 3 at $25,000 |
| Top headline | Up to $200,000 (AWS guide), more for AI startups via account team | $350,000 for AI-first startups | $150,000 at Level 4 or via investor referral |
| Stage rules | Pre-Series B, founded in last 10 years | Start: founded in last 2 years. Scale: pre-seed to Series A | Investor Offer: not Series C or later |
| Investor link needed? | For Portfolio, yes (Org ID) | For Scale, venture funding | For the largest offer, yes (referral code) |
| Typical validity | Usually 1 to 2 years | Start about 1 year, Scale and AI over 2 years | 90 or 180 days at entry; 2 years for Investor Offer |
| Extensions | AWS says it cannot extend | Check program terms | Microsoft says entry credits cannot be extended |
| AI model coverage | Third party models in Bedrock | Gemini, Gemma, Vertex AI open models | Azure OpenAI Service and other Azure AI |
| Notable extras | Activate support resources | Workspace, Maps credit, Skills credits | Visual Studio, GitHub, Microsoft 365 seats at some levels |
The pattern is consistent. All three offer a small, easy entry credit and a much larger credit for startups with investors, a referral, or demonstrated usage.
Realistic Expectations by Stage
Headline numbers describe the best case. Here is what different teams can reasonably expect:
| Your situation | AWS | Google Cloud | Azure |
| Solo founder, no investor | Founders, about $1,000 | Start, about $2,000 if eligible | Levels 1 and 2, up to $5,000 |
| Accelerator or angel backed | Portfolio through the provider | Possibly Scale if funding qualifies | Founders Hub levels, or an Investor Offer if the investor is in the network |
| VC-backed AI startup | Portfolio, plus a talk with the account team | Scale AI, up to $350,000 | Higher Founders Hub levels or Investor Offer |
These are broad patterns, not promises. Each provider decides each application.
How Expiry Compares
Expiry is where the programs differ most, and where founders get hurt.
- AWS: credits usually last one to two years depending on the package. AWS says it cannot extend them, and it emails members 60 and 30 days before expiry.
- Google: Scale and AI tier credits run over two years, with year two covering a smaller share of usage for the AI tier. The Start tier is shorter.
- Azure: the initial $1,000 lasts 90 days, verified entry credits last 180 days, and Microsoft says neither can be extended. The Investor Offer gives a longer window once activated.
A short window means you should have a plan ready before you redeem. A long window means you should still track the date, because unused credits are forfeited on all three clouds.
How AI Coverage Compares
AI is now the biggest driver of startup cloud spend, so coverage matters more than the headline.
- AWS: Activate credits can be used on third party models in Amazon Bedrock, from providers such as Anthropic, Meta, Cohere, and Mistral AI. AWS says larger awards above $200,000 exist for AI startups through account managers.
- Google Cloud: program credits cover Gemini, Gemma, and Vertex AI open models. Third party models are billed directly, with a separate $10,000 option for AI tier startups.
- Azure: credits apply to any Azure product, including Azure OpenAI Service.
If your product depends on one model family, check which provider offers that model on terms you can use credits for. Model availability and pricing differ, so confirm it before you commit.
What None of the Programs Cover
Credits never cover everything, and each provider keeps its own exclusion list. AWS’s list includes Marketplace charges (except Bedrock third party models), Route 53 domain registration, and upfront Savings Plans and Reserved Instance fees. Google and Microsoft publish their own exclusions in their program terms, so read them before you design around any provider’s credits.
Application Effort and Speed
AWS says its review typically takes 7 to 10 business days and shows status on a Credit Application Status page. Google reviews applications on a rolling basis, and the program pages describe the tier requirements. Microsoft starts you at Level 1 on signup, then moves you up as you verify your business and as usage grows.
Plan for a few weeks in the worst case, especially if you need an investor code or an Org ID. Do not wait until your last credits expire to apply.
Turn Credits Into a Runway Number
A credit balance only matters when you convert it to time. Use this simple rule: runway in months equals your credits divided by your monthly eligible spend, capped by the validity period.
The table below uses hypothetical numbers to show how both limits work. Credits run out early when spend is high, and they are forfeited when spend is low.
| Credits | Monthly eligible spend | Validity | Result |
| $5,000 | $1,500 | 6 months | Credits run out after about 3.3 months |
| $5,000 | $400 | 6 months | You use $2,400, and $2,600 is forfeited |
| $50,000 | $3,000 | 24 months | Credits run out after about 16.7 months |
| $50,000 | $1,200 | 24 months | You use $28,800, and $21,200 is forfeited |
The lesson is the same on every cloud. A large credit with a long window is only valuable if your real usage can absorb it, and a small credit with a short window needs a plan on day one. Do this math for each offer you are considering, using your own expected spend.
How to Choose the Best Startup Cloud Credits for Your Company
The biggest number is rarely the best choice. Credits last months or a couple of years. Your cloud choice can last a decade. Use the questions below to decide with the long view in mind.
Start With Your Stack and Your Team
What do your engineers already know? Time spent learning a new platform is a real cost, and it often outweighs a few thousand dollars in credits.
- If your code, tools, and hiring pool already lean toward one provider, that provider has a head start.
- If your product depends on a specific managed service, check that it exists on each cloud, and what it costs.
- If you plan to use containers and infrastructure as code, you keep more freedom to move later.
A team fluent in one cloud will usually ship faster there, even if another program offers a larger headline credit.
Match the Credits to Your AI Plans
If AI is central to your product, coverage matters more than the maximum amount. Ask three questions for each provider:
- Which models do I need, and are they available on this cloud?
- Can credits pay for those models, or are they billed separately?
- What do those models cost at my expected usage after the credits end?
For example, AWS says Activate credits can be used on third party models in Amazon Bedrock. Google says its credits cover Gemini, Gemma, and Vertex AI open models, while third party models are billed directly. Microsoft says Azure credits can be used toward Azure OpenAI Service. Those differences can matter more than a $50,000 gap in ceilings.
Match the Program to Your Funding Situation
Your investors can change what you qualify for.
- Bootstrapped: all three offer entry credits without an investor, though the amounts are modest.
- Accelerator or angel backed: ask your provider for an AWS Org ID, a Google partner nomination, or a Microsoft Investor Network referral code. The larger awards often depend on this link.
- VC backed: check each program’s stage rules, such as AWS pre-Series B, Google pre-seed to Series A, and Microsoft’s limit of Series C or earlier for its investor offer.
Before you apply anywhere, ask your investors which clouds they have partner relationships with. A five minute question can unlock a much larger offer.
Price the Day After Credits End
Credits delay the bill. They do not remove it. Estimate your steady monthly cost on each cloud before you commit. AWS offers the AWS Pricing Calculator for this, and the other providers publish their own pricing tools.
Pay attention to the items that grow quietly: data transfer, managed database fees, logging, and support plans. A cloud that is slightly cheaper for compute can be more expensive once those add up. A good rule of thumb is to model the full price month, not the credit adjusted one.
Think About Portability
Lock-in is a spectrum. Using managed services speeds you up, and it also ties you to a provider’s way of doing things. You do not have to avoid that, but you should choose it on purpose.
- Package workloads in containers, and keep deployment scripts in infrastructure as code.
- Keep data in open formats where you can.
- Know what leaving would cost. AWS, for example, offers free data transfer out to the internet for customers moving all of their data off AWS, as described in its announcement.
Can You Use More Than One Provider?
Yes, many startups do. Each program has its own terms, so read them before you stack offers. A common pattern is to run the main product on one cloud and use another provider’s credits for a specific workload, such as model experiments.
Be realistic about the cost of splitting. Multi-cloud setups add networking charges between clouds, extra tooling, and more for your team to learn. For most small teams, one primary cloud is the simpler path.
A Simple Scorecard
If you want a structured comparison, score each provider from 1 to 5 on the factors below, multiply by the weight, and add the totals.
| Factor | Suggested weight | What to ask |
| AI model coverage | 25% | Can credits pay for the models I need? |
| Credit amount and validity | 20% | What will I actually qualify for, and for how long? |
| Fit with my stack and team | 20% | Where can we ship fastest? |
| Cost after credits | 20% | What is my full price monthly bill? |
| Support and perks | 10% | What help comes with the program? |
| Portability | 5% | How hard would it be to move later? |
The weights are a starting point. A deep learning company might raise AI coverage, while a regulated startup might raise support.
How to Apply for Startup Cloud Credits Step by Step
The process is similar across providers, even though the forms differ. Follow this order.
- Write down your facts. Collect your company website, legal name, founding date, funding stage, last funding round date, and a short product description.
- Set up a business email. Use an address on your company domain. Mismatched emails and websites are a common cause of delays.
- Ask your investors about partner links. Request an AWS Org ID, a Google partner nomination, or a Microsoft Investor Network referral code, if they have one.
- Create the cloud account first. AWS requires an account on the Paid Tier Plan, and the other providers also tie credits to an account or tenant.
- Start with the entry program. Apply for the free, no investor option on each cloud you might use, and keep your accounts and details consistent.
- Apply for the larger tier. Use the provider’s stage rules to pick the right one, and describe your product and expected usage honestly.
- Track the status. AWS shows progress on its Credit Application Status page, and the other providers send updates by email or portal.
- Redeem with a plan. Short validity windows, like Azure’s 90 day entry credits, reward teams who know what they will build first.
- Record every expiration date. Put each date in a shared calendar with reminders at 90, 60, and 30 days.
For AWS specifics, see the AWS Activate application guide. For Google Cloud and Azure, use each provider’s official startup program pages for the current form and requirements.
How to Make Startup Cloud Credits Last
Whichever cloud you choose, the same habits stretch credits furthest:
- Set budgets and alerts early. On AWS, AWS Budgets can warn you as spend approaches your credit balance, and the other clouds have similar tools.
- Tag resources. Tags by project, environment, and team make it clear where credits go.
- Turn things off. Stop development and test environments outside working hours, and delete experiments you no longer need.
- Right size. Start small and grow into larger instances only when metrics demand it.
- Watch data transfer. Moving data between regions or out to the internet is a classic hidden cost.
- Know your full price bill. Check your monthly run rate without credits, so the end of the program holds no surprise.
- Plan the next funding source. Apply for the next tier or program well before the old credits expire.
Decision Guide: Which Cloud Should You Start With?
- Your team already knows one cloud well: start there, and take its credits.
- You are bootstrapped and need a small, quick start: apply for all three entry offers, but build on one.
- You are VC backed and AI first: compare the Scale AI tier at Google, the AWS AI startup route through your account team, and the Azure investor offer. Pick based on model coverage and fit.
- Your investors are partners with one provider: ask for that provider’s referral or Org ID first.
- You depend on one AI model family: choose the cloud where credits can pay for it.
- You expect heavy data movement: model the networking and transfer costs before you decide.
- You are still unsure: run a one week proof of concept on your top two clouds and compare real bills and developer experience.
Checklist: Questions to Settle Before You Redeem Credits
Once you are approved anywhere, answer these before you spend a dollar:
- What is the exact amount, and when does each credit expire?
- Which services and models are eligible, and which are excluded?
- Can credits be used on the AI models my product needs?
- Does the credit apply to one account only, or to my whole organization?
- What happens to my support plan, licenses, or bundled perks when credits end?
- Can I move up a tier or reapply, and what are the conditions?
- Who owns the cloud account, and who has billing access?
Write the answers in one shared document next to the expiration dates. It takes ten minutes, and it prevents the most common credit mistakes.
Common Mistakes When Comparing Startup Credits
- Chasing the biggest number. The maximum applies to a narrow group of startups.
- Ignoring the validity period. Credits that expire in 90 days demand a different plan than two year credits.
- Assuming credits cover everything. Every program excludes some services.
- Skipping AI coverage checks. Some models are billed directly even when you hold credits.
- Applying with inconsistent details. Different names, dates, or domains slow down reviews.
- Splitting across three clouds too early. Multi-cloud complexity can cost more than the credits save.
- Forgetting the full price month. The day credits end, your bill changes overnight.
One honest limit applies. Programs, amounts, and rules change often, and each provider decides each application. Treat this guide as a comparison framework, and confirm the latest details with each provider before you commit.
Practical Examples
These three scenarios are illustrative, not real customer cases.
The solo founder. A two person team with no investors takes the free entry offers from all three providers. They build a prototype on the cloud their developers already know, and use the others only for short experiments.
The accelerator startup. A seed stage company in an accelerator asks the program manager for an AWS Org ID and a Microsoft referral code. It compares offers, then picks the one that covers the AI models its product needs.
The AI lab. A venture backed team building on large models compares Google’s AI tier, AWS’s route for larger AI awards, and Azure’s investor offer. It chooses by model availability and its full price monthly bill, not by headline credits.
Conclusion
Comparing AWS vs Google Cloud vs Azure startup credits comes down to fit, not size. AWS Activate, Google for Startups Cloud, and Microsoft Founders Hub all offer a small entry credit and much larger awards for backed or AI-first startups. They differ in eligibility, validity, extensions, and what the credits can pay for.
Pick the program that matches your stack, your funding, and your AI needs. Apply for the entry offers now, ask your investors for the right referral, and put every expiration date on a calendar. Then model the full price month, so the day your credits end is just another day.
