Dubai-based AI infrastructure startup Enclave42 has launched a commercial AI inference service designed to give SMEs and mid-market businesses more control over how their AI workloads are executed, verified and deployed.
The DIFC-based company runs supported models inside Trusted Execution Environments (TEEs), which are designed to prevent infrastructure administrators, operating systems and hypervisors from accessing the data and code being processed. Customers can independently verify the protection by generating a code, downloading attestation evidence and checking that the confidential environment was active.
Enclave42 currently offers 13 open-source models through a single API, covering language, reasoning, coding, vision, embeddings, speech, reranking and multimodal workloads. Subscription pricing starts at $99 for access to eight models, while the $399 Business tier includes frontier models. Inference usage is billed separately through a pay-as-you-go wallet, and a free explore option is available.
The company said it has access to around 100 cloud GPUs sourced for price and performance and is now deploying its own GPU infrastructure in the UAE for sovereign workloads. Following a pilot phase, self-service access is available. Its next phase includes private model deployments, dedicated GPUs, private inference endpoints and custom model hosting.
Source: Middle East AI News
