The Detrimental Cost of Scaling Too Fast
September 13, 2026 – 4:42 pm
Photo by Greg Keith
Credit: Greg Keith
TL;DR
- 44% of organizations are scaling AI enterprise-wide (McKinsey).
- Yet, 95% of generative AI pilots show no measurable bottom-line impact (MIT).
- Greg Keith introduces the Scaling Instability Curve framework.
- It identifies the point where delivery velocity surpasses architectural maturity.
- Organizations become unstable when governance doesn’t scale with technology.
- Keith emphasizes questioning: "Why are you doing this?" before implementing change.
AI has significantly penetrated enterprises, as evidenced by McKinsey’s 2026 global survey revealing that 44% of organizations now scale AI across the enterprise—a rise from 38% the previous year, with 54% among companies generating over $1 billion in revenue.
However, a concerning gap exists between AI adoption and organizational change. McKinsey’s findings highlight that workflow redesign has the strongest association with AI’s EBIT impact, contrasting with a 2025 MIT study showing that most generative AI pilot projects lack measurable bottom-line effects. This disparity suggests technology might be outpacing organizations’ ability to adapt.
Cloud economics adds another layer of complexity. The FinOps Foundation’s 2025 research on organizations spending over $69 billion in public cloud found that workload optimization and waste reduction remain top priorities, with AI spending managed by 63% of respondents. This scale could transform a seemingly efficient technology decision into a substantial financial commitment.
The issue often lies not in the technology itself but in inadequate governance and unchanged decision-making processes. Organizations might introduce new platforms, increase headcount, automate processes, or restructure teams without addressing the underlying drivers, leading to recurring failures, longer recovery times, rising costs, and growing friction among responsible teams.
Greg Keith, founder of MGKgroup, drawing from over 25 years of experience in engineering, data, cloud, architecture, and technology leadership, has developed the Scaling Instability Curve. This framework, based on real-world observations, identifies the point where delivery velocity surpasses architectural maturity and operational control. It explains how rapid system growth can lead to a cycle of high costs, blurred team ownership, and slower deployments.
According to Keith, organizations become unstable when governance and decision-making fail to keep pace with their growth. He emphasizes questioning the intent behind proposed changes: "Why are you doing this?" to ensure decisions remain grounded and emotionally detached.