For an entrepreneur or an investor, the decisive question is not only where AI creates value, but who captures it.
5.1The macroeconomic order of magnitude
Three estimates are the reference. They are projections, to be handled with care; they nevertheless converge on one point: the impact is counted in trillions of dollars.
- McKinsey (2023) puts the annual contribution of generative AI between $2.6 and $4.4 trillion, up to $7.9 trillion when its integration into existing software is taken into account. Notably, ~75% of this value is concentrated in four functions (customer operations, marketing and sales, software engineering, R&D).
- Goldman Sachs (2023) estimates that generative AI could raise global GDP by 7% (~$7 trillion) over ten years, while exposing the equivalent of 300 million full-time jobs to automation.
- PwC (Sizing the Prize) puts forward a contribution of $15.7 trillion to the global economy by 2030.
$4.4tn
Annual contribution (McKinsey, high estimate)
+7%
Global GDP over 10 years (Goldman Sachs)
$15.7tn
Contribution by 2030 (PwC)
5.2Opportunities by sector
- Health: diagnostic support, drug discovery (AlphaFold), administrative automation, transcription of consultations.
- Knowledge work & productivity: writing, analysis, programming, the most immediate source of value.
- Finance & insurance: fraud detection, risk analysis, compliance, customer relations.
- Legal: contract review, case-law research, drafting of documents.
- Education: personalised tutoring, generation of course materials, accessibility.
- Industry & robotics: predictive maintenance, quality control, physical automation.
- Climate & energy: grid optimisation, modelling, materials.
- Agriculture, media, marketing, logistics: dozens of vertical use cases.
5.3Who captures the value? A map of the players
- Infrastructure providers ("picks and shovels"). Nvidia and compute providers capture an enormous share of the value, whoever wins the model race.
- Foundation model developers. A prestigious position, but extremely capital-intensive and competitive.
- Vertical SaaS. Beware of the difference between a simple wrapper (a fragile layer) and a genuinely defensible value proposition (proprietary data, deep integration, distribution).
- Consulting and integration. A considerable market, since adoption remains the main bottleneck in companies.
- Sector-specific niche start-ups. Where fine domain expertise plus rare data create a lasting advantage.
Key takeaway for entrepreneurs
A defensible advantage rarely comes from the model itself (available to anyone through an API) but from what surrounds it: proprietary data, integration into a real workflow, distribution and trust. Vertical niches poorly served by the giants are often the best opportunities.
Sources
McKinsey, The Economic Potential of Generative AI (2023, mckinsey.com); Goldman Sachs, Generative AI Could Raise Global GDP by 7% (2023); PwC, Sizing the Prize (2017).
Finished this module?
Your progress is saved on this device.