The shifted operation model and MSPs in the AI Age

2026-04-09 · Source: Thoughtworks Insights · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Cloud Computing & IT Infrastructure · Depth: Intermediate, short

Summary

The article, published April 09, 2026, asserts that intelligence, specifically AI, has transitioned from a tool to critical infrastructure, necessitating a fundamental shift in operating models and managed services. This transition is driven by AI becoming economically viable across business scenarios, fast enough for real-time workflows, and risky enough to cause systemic business impact. Since 2022, the cost of equivalent intelligence (MMLU) has fallen by roughly 10x per year, with a "good enough" ~75% MMLU model (GPT-5 Nano) projected to cost about \$0.00067 per MMLU point by 2026, a ~2,000x reduction from GPT-3-level capability in 2021. This evolution requires enterprises and Managed Services Providers (MSPs) to rethink their offerings, moving from traditional IT asset management to operating probabilistic, decision-making intelligent systems.

Key takeaway

For Directors of AI/ML or AI Architects evaluating their operational strategy, recognize that AI's infrastructure status demands a complete overhaul of traditional IT operating models. You must prioritize providers capable of managing probabilistic intelligent assets, focusing on metrics like decision quality, risk exposure, and AI compliance, rather than just uptime. Shift your governance to include model evaluations, policy-as-code, and human-in-the-loop processes to ensure safe and traceable AI operations.

Key insights

AI's shift to infrastructure demands new operating models and managed services focused on probabilistic intelligent assets.

Principles

• Capabilities become infrastructure when viable, fast, and systemically risky.
• AI systems are probabilistic, not deterministic, requiring new validation.
• Reliability now includes decision quality and behavioral stability.

In practice

• Evaluate providers on AI safety and model optimization capabilities.
• Redesign MSP services for AI adoption and decision quality metrics.
• Implement policy-as-code and human-in-the-loop for AI governance.

Topics

• AI Infrastructure
• Managed Services Providers
• Operating Models
• Probabilistic AI Systems
• AI Governance
• MMLU

Best for: CTO, VP of Engineering/Data, Executive, Director of AI/ML, AI Architect, MLOps Engineer

Related on AIssential

• From AI projects to an operational capability — AI becomes an operational capability when it impacts P&L, business KPIs, and is supported by modern architecture and governance.
• Best Practices for AI Infrastructure Planning — Robust AI infrastructure planning is crucial for scalable, cost-effective, and secure AI model training and deployment.
• AI Models as a Service: Powering Agentic AI, Privacy, & RAG — Models as a Service provides a sovereign, internal API gateway for managing and deploying multiple AI models.
• AI agents are entering their rebuild era as enterprises confront the reliability problem — Enterprise AI agent reliability in production hinges on robust workflow orchestration, state management, and recovery, not just LLM performance.
• Dell Technologies: Is Your Infrastructure AI-Ready? — Purpose-built infrastructure is critical for unlocking AI's full potential and overcoming limitations of legacy systems.
• 8 AI and data trends shaping financial services in 2026 — Successful AI adoption in financial services hinges on systemic execution and robust data infrastructure, not just model efficacy.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Thoughtworks Insights.