Cruq AI is now in AWS Activate, the Amazon Web Services program for startups. We joined on the Portfolio tier, which is for startups backed by an Activate provider, with $12,000 in AWS credits.
What the program provides
Activate gives startups AWS credits, plus training and technical resources. The credits apply to most AWS services, including EC2, S3, SageMaker and Amazon Bedrock, and generally last up to two years.
Where the credits go
Model inference. Cruq AI is model-agnostic: a workspace connects its own provider keys and picks the model for each agent. Alongside that, we run a few small models of our own, like Jev, which answers the yes-or-no questions behind guardrails and routing. The credits pay for GPU time to serve those models and for testing agents against models on Amazon Bedrock.
Model training. We train and fine-tune those models ourselves, including speech models for Indian languages. Training runs need GPUs for hours at a time, and that is where most of the credits will go.
Testing on AWS. Some customers want Cruq AI in their own AWS account, under a self-managed licence. The credits let us run it on AWS the way they would.
It fits a bring-your-own-keys model
Customers use their own provider keys and their own data. The credits cover our side of the work: the models we build and the AWS paths we test. A customer's agents still run on that customer's own model accounts, so their spend and their data stay theirs.
