Cloud-based LLMs risk enterprise stability

"Enterprises are embracing cloud-hosted large language models (LLMs) at unprecedented rates. Lured by the promise of rapid deployment, scalability, and transformative capabilities, organizations are becoming increasingly entwined with these outsourced intelligence engines. However, a dangerous underlying pattern is emerging, one too often overlooked until catastrophe strikes."

"Recent events, especially the major outages of 2025 that shut down production for hours and cost billions for global companies, highlight the need for serious reconsideration. We must understand that LLM outages are not rare anomalies; they are becoming more likely and can have serious, companywide impacts."

"LLMs, whether from Anthropic, OpenAI, or others, are mostly accessed through a small number of large cloud providers. This shift marks a major departure from the traditional shop model of earlier internet days, where each company managed its own system, and failures were contained. Today, when an LLM or its cloud host encounters issues, the impact spreads quickly across dozens and sometimes hundreds of dependent businesses in real time."

Organizations are rapidly adopting cloud-hosted large language models for their scalability and transformative capabilities, but this trend is creating dangerous architectural vulnerabilities. Recent major outages in 2025 demonstrated that LLM failures are not rare anomalies but increasingly likely events with severe business impacts. The centralization of LLM services through a small number of large cloud providers creates a single point of failure affecting hundreds of dependent businesses simultaneously. Unlike earlier distributed internet architectures where failures remained contained, today's LLM-dependent systems experience cascading outages across entire enterprise ecosystems. Enterprise architects and CTOs must recognize this pattern and implement resilience strategies to protect against these emerging infrastructure risks.

Read at InfoWorld