Abstract: The deployment of artificial intelligence workloads has crystallised around two principal paradigms: public cloud hosting and private on-premise infrastructure. This paper provides a comprehensive, evidence-based comparative analysis of these paradigms across four critical dimensions: total cost of ownership (TCO), performance and scalability, data privacy and regulatory compliance, and security risk. Drawing on peer-reviewed literature.......
[1].
Knoop, S., & Holtmann, J. (2026). Benchmarking Consumer Blackwell GPU Inference: Token Economics, NVFP4 Quantisation, And Break-Even Analysis For Self-Hosted Llms. Arxiv:2601.09527.
[2].
Armbrust, M., Fox, A., Griffith, R., Joseph, A. D., Katz, R., Konwinski, A., ... & Zaharia, M. (2010). A View Of Cloud Computing. Communications Of The ACM, 53(4), 50–58. Https://Doi.Org/10.1145/1721654.1721672
[3].
Bhavsagar, A. (2025). Cloud Security Risks In AI Deployments: Breach Statistics, Financial Impact, And Zero-Trust Mitigation Strategies. International Journal Of Science And Advanced Technology, 16(3). Https://Doi.Org/10.71097/Ijsat.V16.I3.7823
[4].
Brown, I., Hooper, M., & Sherwood, J. (2012). Multi-Tenancy Security Risks And Taxonomy In Cloud Computing Environments. Proceedings Of The 15th International Conference On Network-Based Information Systems (NBIS).
Https://Doi.Org/10.1109/NBIS.2012.142
[5].
Park, J., & Han, D. (2025). Specedge: Speculative Decoding For Edge-Cloud Hybrid LLM Inference With 1.91× Cost Efficiency And 2.22× Server Throughput. Arxiv:2505.17052