guides
Before You Apply for Startup Credits
Startup credits can leave you with a larger bill than you expected. Check eligibility, coverage, expiry and migration costs before choosing an offer.
Using the full $100,000 benefit in the second year of Google’s Scale programme can require $500,000 of eligible spending. The published coverage is 20%, up to that cap. Your share at the maximum is $400,000, before other charges.1
That can be a substantial benefit for a company with the workload to use it. For a team expecting to spend $10,000, the same rate means $2,000 of support. The headline has not changed. Its value to the company has.
Sourcey’s startup programme catalogue brings offers together with eligibility, application routes, and source links. How we check those programmes explains what supports the published terms. Once an offer makes your shortlist, work out what it will pay for in your actual plan. The award, its usable value, and the cost of adopting the service are three different numbers.
Work out which route is open to you
AWS Activate illustrates why the application route deserves attention before the maximum award. Its programme page distinguishes an initial $1,000 award on the self-funded Founders route from up to $5,000 for selected participants. Portfolio advertises up to $200,000 and uses an Activate Provider route. The advertised maximum is no help if you cannot use the route that leads to it.2
If you have a qualifying partner relationship, identify the information the provider requires. If you are bootstrapped, use the route available to that situation. “Ask your VC” is not much of an answer to a founder who doesn’t have one. Free tiers are often part of that route, and We Are Founders compares the free tools bootstrapped founders can start on, with the limits each one sets.
The same distinction appears outside cloud infrastructure. Agora’s named VC programme uses a form for investors acting for portfolio companies. A startup can be relevant to the programme while the next action belongs to its investor.
Read the intended applicant, not just the word “Apply.” A direct application, partner referral, and portfolio nomination are different routes.
Read the percentage as carefully as the ceiling
Google’s Scale benefits separate first-year coverage from the second-year rate. Treat each period as its own calculation. The second year covers 20% of eligible spend, capped at $100,000, so the amount you use depends on the bill you incur.1
For a hypothetical eligible workload, the arithmetic looks like this:
| Eligible second-year spend | Benefit at 20% | Remaining spend before other charges |
|---|---|---|
| $10,000 | $2,000 | $8,000 |
| $100,000 | $20,000 | $80,000 |
| $500,000 | $100,000 | $400,000 |
These examples apply the published second-year rate and cap to hypothetical eligible bills. They show why “up to” belongs in the calculation, not just in small print beneath the headline.
A coverage schedule tells you more than a large total. Record the rate, qualifying period, and cap together. If an offer has several stages, model the stages separately.
Map the offer to the bill you expect
“Cloud” is too broad a category for this calculation. List the services your workload uses and check which charges the award covers.
AWS’s promotional-credit terms contain exclusions and exceptions, including categories of Marketplace charges, certain support, domains, and upfront commitments. Programme-specific authorisation matters. Read the applicable offer and terms together rather than assuming that every charge from the same provider qualifies.3
A useful working table is small:
| Item | What to record |
|---|---|
| Service | The actual product or billable component you plan to use. |
| Expected usage | A realistic month, with uncertainty made visible. |
| Coverage | The eligible proportion and any service-specific restrictions. |
| Timing | Start date, staged benefits, and expiry. |
| After the award | The ordinary cost of the same workload. |
Keep excluded charges alongside eligible ones. Otherwise, the forecast can make an award appear to cover an entire system when it pays for only one part.
For a simple planning example, suppose a month has $8,000 of eligible usage and $2,000 of excluded charges. At 20% coverage, with enough award remaining and no other adjustments, the benefit is $1,600. The cash bill is $8,400. It is not $8,000, and the excluded charges do not help consume the award.
Then put the move on the same sheet. If switching providers costs $12,000 in engineering time and saves $1,600 a month, it takes 7.5 months just to recover that work. That is a hypothetical break-even calculation, before post-credit price differences or exit costs. An award expiring in six months would not pay back the move on those assumptions.
The biggest offer may not be the best fit
Suppose one fictional offer advertises $50,000 but expires before you expect to consume more than $4,000 of eligible services. Another offers $8,000 against infrastructure you already use.
The second may be easier to use well. That still does not settle the decision. Compare normal pricing, reliability requirements, migration work, commitments, and what your team wants to operate after the credit ends.
Moving data, changing deployments, learning another provider, and rewriting integrations all take time. Estimate the work you can identify; keep the uncertainty visible where you cannot. A migration does not become free because the first infrastructure invoice is covered.
Sometimes a credit offer arrives at exactly the right moment for a technically sensible move. Sometimes it encourages a team to build a system it never needed. The offer should support the architecture, not quietly choose it.
Check what kind of benefit it is
Wysera’s startup programme gives a different example. Its offer describes a two-week setup programme, direct founder involvement, and discounted pricing for twelve months after conversion. It also advertises savings compared with a bundle of other tools. That comparison is not a credit balance deposited in your account.
The help may be valuable. A team trying to consolidate its work may care more about getting set up than receiving another small discount. But if you were never going to buy the comparison bundle, its advertised cost is not money you have recovered.
A spend credit pays eligible charges. A discount reduces the price of something you still buy. A free period gives you time. Setup help gives you another person working on the implementation. A list that adds all four into one enormous “total benefits” number has made them harder to compare.
For a pilot, check what happens when it ends. For a discount, check the price it applies to and its duration. The same goes for the negotiated deals that SaaSOffers collects outside the official programmes. For setup help, identify what the team still needs to provide. If setup is what has held the team back, help with implementation may be worth more to you than a credit balance against software nobody has configured.
Timing can change the decision
An award with an expiry is useful only when eligible spending happens inside its window. Find out whether the clock starts on approval, activation, account creation, or another event specified by the programme.
Also check prior-credit and existing-account conditions. A small award taken now may interact with a later application. The relevant programme should explain the relationship; do not assume awards stack because they appear on different landing pages.
For staged benefits, write down what releases the next stage and what the provider can decide. A possible future award belongs in a different column from a confirmed balance.
If you already use the eligible services, a suitable award can fund work you were going to do anyway. Set a reminder before the benefit ends, while you still have time to adjust usage or budget for the normal bill.
Keep a record you can revisit
For each serious candidate, save the official source, the tier, the eligible applicant, the access route, benefit type, coverage schedule, exclusions, expiry, prior-credit conditions, and the date you checked it.
Then add two estimates from your own plan: what you expect to use during the benefit period and what the same work costs afterwards. An unknown stays unknown until you have an answer. Missing expiry information is not permission to assume the offer lasts forever.
Programme terms checked 17 October 2026. Recheck the linked provider terms before applying; eligibility and awards remain the provider’s decision.
Keep that worksheet beside the offers you shortlist in Sourcey. When a provider changes an amount, you can update the calculation. When your workload changes, you can see which offers are still worth pursuing.