GPU compute credits for Series A+ startups
Dedicated GPU capacity for training and inference, credited instead of billed on demand. The larger, usually partner-gated tiers that open once you have a priced round. 4 programs for Series A+ startups, up to $150K+ combined published ceiling.
One of the more generous open-model inference programs, with the allocation scaling by which partner refers you. Useful if you are serving Llama, Qwen, DeepSeek or other open weights and want to move off per-token frontier pricing without standing up your own GPUs.
Free to join, no fees and no equity. The direct value is preferred pricing on select NVIDIA hardware and software, free cloud credits from NVIDIA and its partners, and free self-paced technical training. The indirect value is often bigger: Inception Capital Connect makes investor introductions, and membership is a recognized signal that unlocks better tiers in other programs.
Serverless GPU credits as a one-time grant, with direct access to Modal's engineering team and go-to-market help around launches and fundraising. Modal does not publish amounts. The Seed to Series A track needs either a partner-network VC or more than $1M raised from any fund; the Series B+ track needs $30M+ raised and a partner-network investor.
Aimed at funded companies migrating real workloads rather than at first-time builders: up to $100,000 in credits for the migration itself, then long-term discounts of up to 35%, executive sponsorship, architecture reviews and a dedicated account manager. Relevant specifically because GPU access is carved out of several larger cloud programs — if you need sustained GPU time rather than API calls, a GPU-friendly provider often beats a bigger credit you cannot spend on accelerators.
Browse by category
Find out what you qualify for today
Tell us four facts and we will come back with an ordered list — or look through the database yourself. No account, no charge.