For decades, scale was the ultimate moat: more people, more processes, more markets, more capital. Today, that same scale is an anchor. AI is exposing just how brittle most large, monolithic organisations really are.
Legacy you can’t map = value you can’t unlock
AI doesn’t run on vibes and town halls. It runs on clean, connected, well-described knowledge.
Most big corporates don’t have that. They have:
- 20+ years of undocumented workarounds
- Systems bolted together with middleware and hope
- Critical knowledge trapped in email, PDFs, and the heads of people about to retire
To make AI genuinely useful, you have to surface, structure, and govern that knowledge. That’s not a workshop. It’s a ground-up rewiring. And the larger the organisation, the less politically and culturally possible that becomes.
Even AI will eat its own dinosaurs
The irony? The same pattern of disruption is already emerging within AI itself. Just as large corporations are weighed down by their legacy systems, today’s massive AI models are becoming their own kind of monolith—expensive, slow, and increasingly vulnerable.
Distillation, compression and abstraction will create smaller, cheaper, more agile models that outpace today’s giant, slow, expensive incumbents. The same way monolithic organisations struggle to adapt, monolithic models will be leapfrogged by lighter, specialised engines plugged directly into clean, well-structured knowledge.
The end of the monolith
This is why I believe AI will be the death of the classic monolithic corporation.
The future belongs to:
- Modular organisations that treat knowledge as an asset, not an accident
- Networks and ecosystems, not empires
- Companies that assume everything about how they work today is up for renegotiation
Yes, some giants will transform—those rare few with both the courage to cannibalize their own business models and leaders willing to dismantle their empires. But they’re the exception that proves the rule.
AI rewards speed, clarity, and courage. Unfortunately for big corporates — and for bloated AI models — none of those are legacy strengths.
AI won’t kill jobs first. It’ll kill the companies too big to change.