AI everywhere, value somewhere: McKinsey's State of AI 2025 for investors
McKinsey's State of AI 2025 survey found 88% of organisations use AI, but only about a third are scaling it. Here is what the adoption gap means for investors and founders building AI products.
Key points
- McKinsey's 2025 survey found 88% of organisations use AI in at least one function, up from 78% a year earlier.
- 62% are experimenting with AI agents and 23% are scaling agentic AI in at least one function.
- Only about one-third have begun scaling AI across the enterprise, and significant earnings impact remains rare.
- Investors should distinguish pilots from scaled use and focus on retention and workflow change.
Artificial intelligence has become the dominant theme in venture capital. For investors and founders, the key question is not whether organisations are experimenting with AI but whether they are getting enough value to keep paying for it. McKinsey’s latest global survey, The State of AI in 2025, published in November 2025, offers useful evidence.
Adoption is now mainstream
McKinsey surveyed 1,993 participants across 105 countries in mid-2025. It found that 88% of respondents say their organisations regularly use AI in at least one business function, up from 78% a year earlier. More than two-thirds said their organisations use AI in more than one function.
Interest in AI agents, systems that can plan and carry out multi-step tasks, is high. Sixty-two per cent of respondents said their organisations are at least experimenting with AI agents, and 23% said they are scaling agentic AI in at least one function.
Value is not yet widespread
The more important finding for investors is the gap between use and impact. McKinsey reported that only about one-third of organisations have begun to scale AI across the enterprise. Most remain in experimentation or piloting. Meaningful enterprise-wide impact on earnings remains rare.
McKinsey identified a small group of high performers, about 6% of respondents, who report significant value from AI. These organisations tend to set growth and innovation goals for AI, not just cost-cutting targets, and redesign their workflows rather than layering AI on top of existing processes.
What this means for investors
For investors in AI startups, the survey suggests a few lenses:
- Distinguish experimentation from commitment. Many enterprise AI deals begin as pilots. Revenue that depends on pilots converting to full deployments carries more risk than revenue from scaled use.
- Look for workflow change. Products that change how work is done, rather than adding a feature, are more likely to be the source of lasting value.
- Watch retention closely. In a market full of experiments, customer retention and expansion are better signals than initial sales.
- Consider the buyer’s capacity. Many organisations lack the data, skills and change management to scale AI. Startups that help solve those problems may find durable demand.
What it means for founders
For founders building AI products, the adoption gap is both a challenge and an opportunity. Customers are willing to try new tools, but they need help turning experiments into results. Founders who invest in implementation, measurement and change management, not just the model, can stand out.
Regional differences
McKinsey’s survey is global, but adoption varies. The United States leads in AI investment and has the deepest pool of AI startups. Europe has strong research but has historically attracted a smaller share of AI funding. In Asia, China has a large AI ecosystem, and several Southeast Asian economies are investing heavily in AI infrastructure. Australian organisations are active adopters, though local surveys suggest many are also still moving from pilots to scale.
The cost question
Another consideration for investors is the cost of delivering AI products. Running large models can be expensive, and pricing pressure is intense as model providers compete. Startups whose gross margins depend heavily on the price of third-party models face a risk that their costs, or their customers’ expectations, change quickly. Investors increasingly ask how a company’s margins will evolve as usage grows, whether it can switch between model providers and whether its value lies in proprietary data, workflow integration or distribution rather than access to a particular model.
Frequently asked questions
Does high AI adoption mean AI startups will succeed?
Not on its own. Adoption creates demand, but competition is intense and many experiments do not convert into large, lasting contracts.
What are AI agents?
AI agents are systems that can plan and carry out multi-step tasks with some autonomy, such as researching, drafting and acting within software tools, rather than responding to single prompts.
What distinguishes AI high performers in the survey?
According to McKinsey, they set ambitious goals for AI beyond cost reduction, redesign workflows and invest a larger share of their digital budgets in AI.
References
- McKinsey: The state of AI in 2025 – Agents, innovation, and transformation
- McKinsey: AI at work but not at scale (chart)
Related reading
Source: McKinsey: The state of AI in 2025 – Agents, innovation, and transformation
General information only. This article reflects the author’s views at the time of writing. It does not take into account your objectives, financial situation or needs, and it is not financial, legal or tax advice or a recommendation to invest. Figures are from the sources named and may change. See our terms of use.