The AI Talent Shortage in 2026: Why Organisations Struggle to Hire Data Scientists

Boards in Europe, the United States and Asia-Pacific keep reporting the same problem: they cannot find the AI talent they need. Investment in AI has reached record levels, and the supply of experienced data scientists, machine-learning engineers and AI researchers has not kept pace.

Key takeaways

• 78% of organisations surveyed reported using AI in 2024, up from 55% in 2023, and 71% used generative AI in at least one business function (Stanford HAI, AI Index 2025).
• 63% of employers name skills gaps as the main barrier to transforming their business, and the World Economic Forum projects that 59 in every 100 workers will need reskilling or upskilling by 2030 (World Economic Forum, January 2025).
• The shortage is most acute at senior level, because it takes years of applied work to become someone who can lead production AI systems.
• Recruitment processes built for software engineering tend to screen out strong AI candidates and miss the ones who are not looking.
• Organisations that hire an experienced AI leader first, and invest in data infrastructure and career paths, find the rest of the team easier to recruit and keep.

This is a lasting imbalance rather than a short recruitment cycle, and it affects how competitive and innovative organisations can be. The Stanford HAI AI Index 2025 found that 78% of organisations surveyed used AI in 2024, up from 55% the year before, while use of generative AI in at least one business function more than doubled, from 33% to 71% (Stanford HAI, AI Index 2025). The World Economic Forum’s Future of Jobs Report 2025 projects 170 million new roles and 92 million displaced by 2030, a net gain of 78 million, and 63% of employers already name skills gaps as the main barrier to transformation (World Economic Forum, January 2025).

For organisations hiring in this market, it helps to understand what drives the shortage and why conventional recruitment struggles with it.

What the numbers show

Demand for AI skills has been growing faster than the supply of experienced people for more than a decade. AI-related job titles are among the fastest-growing on the major professional networks, while the number of people who combine deep technical expertise, domain knowledge and experience of production systems remains small.

Adoption is broad, but most organisations have yet to move far beyond pilots, and in our experience the limiting factor is usually people rather than technology, particularly people who can lead AI work across an organisation as well as build it.

Pay reflects the imbalance. PwC’s 2025 Global AI Jobs Barometer found that jobs requiring AI skills paid an average premium of 56% over similar roles, up from 25% a year earlier (PwC, June 2025). At senior level, packages for Chief AI Officers and heads of data science at leading firms now compare with those for other C-suite roles.

Why conventional recruitment struggles with AI roles

Most corporate talent acquisition teams, and most generalist agencies, are set up for volume hiring against clearly defined specifications. Senior AI hiring works differently in three ways.

The roles cross disciplines. A senior data scientist working on computer vision for autonomous vehicles needs deep-learning architecture, signal processing, real-time systems engineering and knowledge of the automotive domain. A machine-learning engineer building fraud detection for a bank needs distributed computing, statistical modelling, regulatory compliance and an understanding of financial markets. Matching keywords on a CV to a specification will not find these people; it takes an understanding of how the technical skills apply to the business problem, which many recruitment teams do not have.

The best candidates are not looking. The strongest AI professionals rarely browse job boards or answer mass InMail campaigns. Reaching them depends on networks in the research community, universities and the small groups of practitioners who move between large technology companies, research laboratories and fast-growing start-ups.

Assessment is a specialist skill. Competency interviews and technical tests designed for software engineers do not show whether a candidate can turn an ambiguous business problem into a machine-learning task, choose a suitable modelling approach, manage production ML systems or lead a cross-functional team through uncertain work. Panels without AI expertise make expensive mistakes in both directions: they turn down strong candidates they cannot evaluate, and they hire people whose academic record does not translate into applied results.

Why the shortage is worst at senior level

The shortage affects every level, but it matters most at the top. Universities are producing more data-science and AI graduates than ever. The path from graduate to a senior data scientist, principal ML engineer or AI lead who can design production systems, develop junior colleagues and connect technical work to business strategy still takes, in Fergal Nolan’s experience, at least five to seven years of applied work.

Training programmes and boot camps cannot close that gap on their own. The World Economic Forum expects 59 in every 100 workers to need reskilling or upskilling by 2030 (World Economic Forum, January 2025), but the leaders who set AI strategy, build and manage teams and handle the ethical and regulatory questions of enterprise AI have to come from a small global pool of experienced practitioners.

Gartner’s survey of chief data and analytics officers illustrates the stakes. 70% of CDAOs now have primary responsibility for their organisation’s AI strategy and operating model, and 36% report to the chief executive, up from 21% the previous year. Gartner also predicts that by 2027, 75% of CDAOs not seen as essential to their organisation’s AI success will lose their C-level position (Gartner, May 2025). The cost of a poor senior appointment is correspondingly high.

What higher pay can and cannot do

Many organisations have responded by paying more, and up to a point that works. In financial services and technology, total packages for senior AI leaders can exceed seven figures.

Pay alone does not hold these people for long, though. The AI professionals most in demand weigh several other things: the problems they would work on, the quality of the data infrastructure, the calibre of colleagues, whether the organisation’s commitment to AI is real, opportunities to publish and attend conferences, and how much autonomy they would have.

Organisations that pay well but fall short on these factors often lose the people they hired at great expense. In our experience, AI professionals who find a gap between what they were told about AI investment and the reality of the data infrastructure and executive support tend to leave early.

Common mistakes

Across a career that has included hundreds of data and AI placements, Fergal Nolan has seen the same failures recur.

Hiring AI specialists like software engineers. Their skills, motivations, careers and evaluation criteria differ. Processes that send AI candidates through standard engineering assessment, including algorithmic coding tests and system-design interviews, filter out people whose strength is statistical thinking, research method and framing problems rather than optimising code.

Unrealistic specifications. A job description that asks for deep learning, natural language processing, computer vision, reinforcement learning, MLOps, cloud architecture and business strategy, with ten years’ experience in fields that barely existed ten years ago, tells experienced candidates the organisation does not know what it needs. The strongest of them do not apply.

Underestimating time to hire. Senior AI roles take much longer to fill than comparable technology positions. Organisations expecting to go from advert to signed offer in four to six weeks are routinely disappointed; effective processes are built on relationships, and a candidate may need months before they are ready to consider a move.

Ignoring reputation in AI communities. AI practitioners form a small, well-connected global community, and word travels quickly. Organisations known for poor data infrastructure, weak executive sponsorship or a gap between what they promise at interview and what candidates find on arrival will struggle to hire whatever they pay.

What successful organisations do

Organisations that attract and keep senior AI talent tend to share four practices.

They hire the AI leader before the team. Appointing a credible, experienced AI leader, whether titled Chief AI Officer, VP of Data Science or Head of Machine Learning, before scaling the team is the most effective single decision. That person sets the technical direction and engineering culture, designs the interview process and attracts other senior practitioners. The rising share of CDAOs reporting to the chief executive reflects the same logic (Gartner, May 2025).

They invest in the conditions AI teams need. That means data infrastructure before data scientists, career paths specific to AI roles, support for publication and open-source work, conference budgets, research partnerships with universities and direct access to senior decision-makers. For experienced AI practitioners these are part of the reason to join.

They use specialist search rather than generalist recruitment. Because the candidate pool is small and passive and assessment needs specialist knowledge, generalist approaches, internal or external, tend to fall short. Specialist AI executive search brings the networks, technical understanding and market knowledge needed to find, approach and evaluate senior candidates.

They look beyond the main hubs. AI talent is unevenly spread. The San Francisco Bay Area, London and Beijing remain large concentrations, but Toronto, Tel Aviv, Singapore, Zurich and several Central and Eastern European cities also have strong talent pools that are often less contested and less expensive. Organisations open to remote or hybrid arrangements for senior AI roles widen their options considerably.

The longer view

The AI talent shortage will not resolve soon. Universities are expanding AI and data-science programmes, and new routes such as online courses and industry-sponsored apprenticeships are widening the pipeline, but it takes years to turn a graduate into an experienced professional who can lead enterprise AI.

With the World Economic Forum projecting a net 78 million new jobs by 2030 and skills gaps already the main barrier to transformation (World Economic Forum, January 2025), demand for experienced AI leaders is likely to outstrip supply for the rest of the decade. For organisations that depend on AI, and in 2026 that includes most of them, a long-term approach to hiring AI talent belongs on the board agenda.

Organisations that treat AI hiring as a board-level issue, invest in the conditions that attract strong people and work with specialists who know this market will be better placed than those that treat it as routine recruitment. Our guides to hiring a Chief Data Officer and building a healthcare AI leadership team cover the senior appointments in more detail, and you can talk to us about a specific search.

Fergal Nolan is the founder of Banba, a specialist executive search firm for Healthcare AI and Life Sciences AI, with offices in New York, London and Berlin. He has spent more than 25 years in talent acquisition, including as a founding team member at SThree / Real Staffing and as founder of Upstream, where he has built Data Science & AI teams and AI leaders for start-ups, scale-ups and global enterprises.