Every hype wave feels like a revolution. Only a handful actually rewire the world. AI is one of them — but we are in an AI bubble right now, and it will correct.
After decades of watching technology cycles — ERP euphoria, the dot-com reckoning, cloud’s slow-burn dominance — I’ve learned one uncomfortable truth: every hype wave feels like a revolution. Only a handful actually rewire the world.
AI is one of them. I have no doubt about that. But let me be direct: we are in an AI bubble right now. And it will correct. The only real question is what the other side looks like — and whether your organisation is building for that world, or this one.
Why I say we’re in a bubble
The signs are textbook. Organisations are overgeneralising capabilities. ROI narratives remain murky across large enterprise investments. Vendor proliferation has outpaced differentiation. And executive FOMO is driving adoption faster than governance structures can absorb. We’re compressing a 10–15 year transformation into 2–3 year expectations. That never ends cleanly.
What the correction looks like (2026–2028)
This won’t be a dot-com-style crash. Think of it as a slow, necessary exhale — a multi-phase deflation that actually clears the path for real value creation.
- Phase 1ROI reality check
The question shifts from “where can we use AI?” to “where does AI create measurable value?” Experimental pilots consolidate. Budgets follow outcomes, not momentum.
- Phase 2Vendor shakeout
Hundreds of startups — most of them thin wrappers around foundation models — disappear. Survivors will have proprietary data advantages, embedded workflows, and measurable outcomes.
- Phase 3Talent recalibration
“Hire AI experts everywhere” fades. Value consolidates around domain expertise paired with AI fluency — and builders who think in products, not projects.
AI won’t be your advantage by 2030. How you use it will be.
The stable state by 2030
By 2030, AI looks less like a revolution and more like electricity — invisible, embedded, expected. Core capabilities become commoditised. The differentiation shifts entirely to your proprietary data, decision models, and execution speed.
The real productivity gain will be 30–40% efficiency improvement across functions — steady, structural gains. Not mass job elimination. Consulting still needs consultants. Finance still needs judgment. Marketing still needs genuine strategy. But the fabric of how work flows will look fundamentally different.
What leaders should focus on now
Not theory — operating reality. Build augmented teams: small groups with embedded AI capabilities delivering the output of much larger ones. Design human-in-the-loop as a structural principle, not an afterthought. AI does 70–80% of the work; humans validate, steer, and own the judgment calls.
Move from project-mode pilots to AI products embedded in your workflow ecosystems. Invest in your internal data as infrastructure. And stop treating the AI model as the strategic asset — your proprietary data, context, and feedback loops are where sustainable advantage will live.
The bottom line
The winners of the AI era will not be the first adopters or the biggest spenders. They will be the leaders who operationalise AI with clarity, restraint, and intent — while everyone else is still chasing the noise.
First published on LinkedIn on 21 May 2026. Read the original and join the discussion
