AI infrastructure Computational Infrastructure; MLOps
noun phrase
Definition: The hardware, software, platforms, and operational resources that support the training, deployment, serving, scaling, and maintenance of AI systems, including compute, storage, networking, accelerators, cloud services, and supporting toolchains [Yu et al. 2025].
Example in context: “Vertex AI, and in providing an adequate infrastructure to train deep-learning models cost-effectively using high-performance cloud GPUs and TPUs with the availability of its AI Infrastructure.” [Oliveira et al. 2024]
Synonyms: AI stack infrastructure; AI platform infrastructure (context-dependent)
Related terms: MLOps; cloud AI; training infrastructure; model serving infrastructure; GPU/TPU infrastructure