The Google for Startups Cloud Program gives eligible startups up to 200,000 dollars in Google Cloud credits over two years, and up to 350,000 dollars if you qualify for the AI-first tier. There is also a small self-serve tier of up to 2,000 dollars for bootstrapped companies with no funding at all. The credits cover Google Cloud Platform services, Firebase, and Google's own models including Gemini on Vertex AI, plus a year of Google Workspace Business Plus and access to Google Cloud startup engineers. It is non-dilutive: no equity, no repayment.
That is the short version. The longer version is that Google runs one of the most generous and also one of the most tiered startup programs of the big three clouds, and which tier you land in depends almost entirely on your funding situation. Here is how it breaks down in 2026.
The three tiers
| Tier | Credit amount | Who it is for |
|---|---|---|
| Start (bootstrapped) | Up to 2,000 dollars | Pre-funding startups with a website or app |
| Scale (funded) | Up to 200,000 dollars over 2 years | Equity-funded, pre-seed through Series A |
| AI-first | Up to 350,000 dollars over 2 years | Funded startups building AI as the core product |
A few details worth knowing about each:
- Start tier. Roughly 2,000 dollars in credits, self-serve, no investor required. It is small, but it stacks fine with the other self-serve cloud credits covered in our startup cloud credits guide, and it does not stop you from applying for a bigger tier later once you raise.
- Scale tier. Up to 200,000 dollars over two years. Google typically covers a high percentage of your usage in year one and a smaller percentage in year two, so the headline number assumes you actually spend at scale in both years.
- AI-first tier. Up to 350,000 dollars over two years. Per Google's published structure, that is 100 percent coverage up to 250,000 dollars in year one, then 20 percent coverage up to an additional 100,000 dollars in year two. The tier also comes with AI-specific extras: dedicated AI training, webinars, and earlier access to some Google AI products.
The year-two step-down matters. Google is not handing you a lump sum; it is subsidizing your bill on a declining schedule designed to convert you into a paying customer. Plan your architecture and spend around that curve, not the headline figure.
What the credits actually cover
This is where Google's program is genuinely differentiated for AI startups:
- Google Cloud Platform compute, storage, networking, BigQuery, GKE, the standard stack.
- Firebase costs are covered under the same umbrella.
- Vertex AI and Gemini. Credits apply to Google's own models, including the Gemini family and Gemma, whether you call them through Vertex AI or scale up Gemini API usage. If your product is built on Gemini, this program can effectively cover your model bill for a year. For what Gemini costs without credits, see our Gemini API guide and the live per-model pricing on our rankings, which we re-check weekly.
- What is not covered: third-party models served through Vertex AI's model garden (Anthropic, Meta, and others) are billed directly by those providers and do not draw down program credits. If Claude is your primary model, budget for that separately or look at Anthropic's own program in our Claude for Startups guide.
On top of credits you get 12 months of Google Workspace Business Plus for new signups, access to a Startup Success Manager and Google Cloud customer engineers at the higher tiers, and enhanced technical support coverage.
Eligibility: what Google checks
For the Scale and AI-first tiers, the pattern in 2026 looks like this:
- Equity funding from an institutional investor, anywhere from pre-seed to Series A. SAFEs from institutional investors count. Grants, prize money, crowdfunding, and friends-and-family rounds generally do not qualify you for the big tiers, though they may still get you a smaller one.
- Recency limits. Series A rounds are expected to be recent (roughly within the last 12 months), while earlier rounds have more room.
- Company age under about ten years.
- New to Google Cloud credits. If you have already received more than about 5,000 dollars in Google Cloud credits, you are typically out for the large tiers. This is the classic reason to sequence your applications carefully; see our startup credits checklist before you burn eligibility on a small credit.
- AI-first tier additionally wants AI to be the core of the product: model training, inference, retrieval pipelines, agents, that kind of thing. A SaaS app with one AI feature bolted on is a Scale-tier application, not an AI-first one.
The partner path
Like AWS Activate and Microsoft for Startups, Google's biggest numbers flow through partners. Many VCs, accelerators, and startup platforms are enrolled in the Google for Startups partner network, and applying through one of them does two things: it verifies your funding for Google, and it usually smooths the approval into the top tier you qualify for. If you are in an accelerator or your lead investor has a startup-perks portal, ask whether they are a Google Cloud partner before applying cold. The self-serve application works, but the partner route is how most startups land the 200k and 350k outcomes.
How to apply, step by step
- Pick your tier honestly. Bootstrapped means Start. Funded means Scale. Funded and AI-native means AI-first.
- Get your paperwork ready. Company website, a Google Cloud billing account, proof of funding (investor name, round, date), and a short description of what you are building.
- Check for a partner referral first. One email to your investor or accelerator can be worth six figures.
- Apply on the Google Cloud startups page. Review typically takes days to a few weeks.
- Once approved, spend deliberately. Credits run on the two-year clock from acceptance, and the year-two coverage drop means front-loading heavy workloads (training runs, migrations, load testing) into year one is usually the right call.
How it compares to AWS and Microsoft
Google's top self-qualifying number for AI startups (350,000 dollars) is the largest headline figure of the big three clouds, but the two-year declining structure and the Gemini-centric coverage make it a different bet than AWS Activate's portfolio path or Microsoft's investor-gated Azure tiers. We compare all three side by side in AWS vs Google vs Microsoft for startups. The short answer: pick the cloud whose models and stack you actually want to run, because all three programs are designed to make leaving expensive.
Bottom line
The Google for Startups Cloud Program is one of the highest-value perks a funded startup can claim in 2026: up to 200,000 dollars for equity-funded companies and up to 350,000 dollars for AI-first ones, covering GCP, Firebase, and Gemini usage, with no equity taken. Bootstrapped founders should grab the 2,000 dollar Start tier now and re-apply after raising. The two rules that catch people: do not accept a small Google credit that disqualifies you from the big tier later, and remember year two only covers 20 percent of your bill.
We track this program, its current terms, and 200+ other hand-verified credits and perks worth over 5.5 million dollars in the Perkstack catalog, with each offer re-checked against the official page. See the Google for Startups Cloud Program AI tier entry, browse the catalog, or create an account to start claiming.
Related reading: How to Get $100k+ in Startup Cloud Credits, The Startup Credits Checklist, AWS vs Google vs Microsoft for Startups