NinjaTech launches fixed-price AI platform for enterprises

NinjaTech AI has launched SuperNinja Enterprise, an AI workforce platform that companies can run in their own cloud environments. It is sold at a fixed annual price that includes computing capacity.

The Palo Alto-based company says the platform lets businesses deploy AI agents inside their own cloud or data centre rather than relying on an external service. It uses open-weight models and also supports models from Anthropic and OpenAI.

SuperNinja Enterprise combines software, model-serving infrastructure and GPU-backed inference capacity under one contract. Customers can deploy the system on single-tenant capacity through Microsoft with Fireworks AI or AWS, while organisations with stricter security requirements can choose a licensed deployment in an air-gapped environment using their own hardware.

The launch comes as companies look to move beyond small AI trials while keeping tighter control over spending and data handling. Usage-based pricing has made it difficult for many businesses to forecast costs as AI agents take on longer-running tasks and consume more computing resources than simple chatbot interactions.

Cost model

NinjaTech says organisations running open-weight models on reserved capacity can reduce end-to-end costs to about one-tenth of comparable deployments based on frontier lab models. The company is positioning the offer as an alternative to per-token billing, with annual pricing designed to stay stable as usage rises.

That pricing model addresses a central concern for large companies testing AI tools at scale. Businesses often have to secure GPU capacity, deploy model-serving systems and source applications through separate suppliers, adding cost uncertainty and procurement complexity.

Packages are sized for 100, 500 or 1,000 AI employees. The platform also integrates with Slack and Microsoft Teams, allowing workers to use the system through tools already common in the workplace.

Data control

NinjaTech is also emphasising data residency and separation. Enterprise data stays within the customer's own environment and is not pooled to train external models, a point likely to matter in sectors handling regulated or commercially sensitive information.

"With SuperNinja Enterprise, your CISO can rest assured data never goes outside your own walls," said Babak Pahlavan, Chief Executive Officer of NinjaTech AI. "Your CFO gets an AI bill that doesn't move. And everyone sleeps well, knowing you can scale AI without waking up to skyrocketing costs."

For the first time, NinjaTech is supplying the GPU and inference capacity for the platform rather than leaving customers to assemble those components themselves. In practice, that means buyers get the AI software and the underlying computing arrangement under one commercial agreement.

Partners involved

Healthcare customers will work with Optimum HealthcareIT for integration, while other enterprise customers will work with Infosys. Infosys is also described by NinjaTech as an investor and implementation partner.

The product keeps model choice open, allowing customers to move workflows between open-weight systems and proprietary models. That approach reflects a broader enterprise effort to avoid dependence on a single model provider as pricing, performance and policy terms continue to shift across the AI market.

NinjaTech says many enterprise AI projects remain stuck at the pilot stage because finance teams struggle to predict usage-based charges and security teams resist sending sensitive workloads to outside providers. It argues that placing the system inside a customer-controlled environment with fixed annual pricing removes both barriers.

"Most vendors keep AI on a usage meter," said Pahlavan. "We give enterprises predictable capacity and costs, so adoption can spread across the organisation instead of stopping at a pilot."