PwC surveyed 1,217 executives across 25 industries and concluded that 74% of AI's economic value is captured by the top 20% of companies.
That number sounds harsh, but the reasons behind it are more nuanced than you might think — it's not because those 20% bought more expensive models or have more data scientists.
Where Leaders and Laggards Truly Differ
PwC developed an "AI Fitness Index" comprising 60 practices in two broad categories:
- AI Use: product integration, personalized experiences, strategic decision-making
- AI Foundations: governance, data quality, talent, infrastructure, ROI measurement
Leading companies (top 20%) achieved these numbers:
| Metric | Value |
|---|---|
| Value creation multiple vs. competitors | 7.2x |
| Profit margin advantage over competitors | 4 percentage points |
| Likelihood of reinventing business models with AI | 2.6x |
| Likelihood of identifying new growth opportunities with AI | 2-3x |
| Likelihood of redesigning workflows (rather than layering tools) | 2x |
The single most critical gap: Leaders use AI to create growth, not just to cut costs.
Why Most Companies Are Stuck
Many enterprises today deploy AI like this: find repetitive processes that AI can replace, install tools, save some labor costs, and then report "AI empowerment" in their presentations.
This approach works in the short term, but it has a low ceiling.
PwC's data shows that leaders redesign workflows to fit AI capabilities, rather than layering AI tools on top of existing processes. That sounds like a small difference, but it actually represents a completely different depth of transformation.
Another problem: 42% of companies have no idea whether their AI investments are generating returns. Without a clear value measurement framework, they don't know where to double down or where to pull the plug. PwC specifically released an "Agentic Business Value Maximization Framework" focused on strategy, use case prioritization, value mapping, and continuous optimization.
The Gap Will Widen
The report includes a direct statement:
"If they don't change their approach, the performance gap between AI leaders and laggards is likely to widen further, as leading companies continue to accelerate learning, scale validated use cases, and safely automate decisions."
This is a positive feedback loop: leaders get better data, more accurate models, and broader automation coverage. Laggards are still running POCs, building slide decks, and searching for ROI.
The gap isn't linear — it's exponential.
Is There Still a Chance for Laggards?
The report outlines three actionable paths, primarily for small and medium enterprises:
- Join consortia to share AI infrastructure and split costs
- Embrace open-source models (at the level of Gemma 4, Llama 4) rather than feeling compelled to use flagship closed-source models
- Engage specialized service providers for production deployment instead of building from scratch
Frankly, the window to catch up isn't as long as many imagine. That 20% is still accelerating, not waiting to be caught.
A Question Worth Asking
If you ask today's executives "What is your AI strategy?", most will answer: efficiency, cost reduction, automation.
That's not wrong, but according to PwC's data, that path only gets a company to the middle of the pack.
Using AI to develop new products, open new markets, and rebuild business models — that's what separates the top 20%. The problem is that this growth-oriented AI strategy requires bigger bets at the organizational level, and most companies' risk appetite can't support that move.
So the 74% vs. 20% divide isn't really just about AI.
Sources: Three-quarters of AI's economic gains are being captured by just 20% of companies (PwC Press Release); CocoLoop, PwC: 74% of AI Gains Go to 20% of Companies. The Rest Are Stuck (pasqualepillitteri.it)