What Google Cloud offers startups
Google for Startups Cloud is the most generous credit program among major hyperscalers, offering up to $200,000 in Google Cloud credits for AI-focused startups (and $100,000 for general startups). Credits cover the full GCP stack — Compute Engine, GKE, BigQuery, Vertex AI, Gemini API, and TPU access for ML training.
Google Cloud's AI tier is its differentiator. Vertex AI gives you direct access to Gemini Pro, Imagen, and TPUs — a serious advantage for startups training or fine-tuning their own models. The two-year window is also longer than most competing programs.
Everything in the program
- Up to $200,000 in Google Cloud credits over 2 years (AI-first tier)
- $100,000 standard tier for non-AI startups
- Full access to Vertex AI, Gemini API, and TPU instances for model training
- BigQuery, Cloud Run, GKE, and Firebase included
- Technical mentorship from Google Cloud solutions architects
- Access to the Google for Startups community and accelerator events
Who qualifies
- Pre-seed to Series A funded startup
- Less than 5 years old
- For the $200k AI tier: building an AI-first product or using Vertex AI / Gemini in production
- Backed by an approved investor, accelerator, or VC partner
- Not currently a Google Cloud paying customer beyond a small spend
The application path
- Visit the Google for Startups Cloud Program page
- Submit the application with company details, AI use case, and funding info
- For AI tier: include details on the models you plan to train, fine-tune, or call
- Google reviews applications within 2–4 weeks
- Once approved, credits are applied to your billing account — start with a small workload to validate setup
Maximize your approval odds
- Apply for the AI tier ($200k) if you can credibly show AI is core to your product — the bar is reasonable
- Use Cloud Run for stateless services to minimize idle spend
- BigQuery is metered by data scanned — partition tables aggressively to extend credits
- Reserve TPU quota in advance — TPU availability for startups is limited and credits are useless without quota
Founders ask before applying
How do I qualify for the $200,000 Google Cloud tier?
You need to be building an AI-first product — typically using Vertex AI, Gemini API, or training custom models on Google Cloud.
How long do Google Cloud startup credits last?
Credits are valid for 2 years from the date of approval.
Can I use Google Cloud credits for Gemini API calls?
Yes. Gemini API usage via Vertex AI is fully covered by your credit balance.
Is TPU access included?
Yes, but TPU quota must be requested separately and is subject to availability.