Beyond the AI Hype

Artificial intelligence is entering a new phase. While attention often focuses on the latest model launches, the bigger story is how organisations adopt AI and turn its capabilities into real productivity gains.

Recent Deutsche Bank Research Institute research has examined AI from several angles: the lessons of past AI cycles, the rise of agentic AI and increasing competition from open-source models. Together, these themes help explain where AI may be heading next.

What is the impact of the falling cost of intelligence?

Capabilities once confined to expensive frontier models are increasingly available through more affordable alternatives. This is lowering barriers to adoption and enabling organisations to apply AI across a growing range of activities.

As explored in AI at 70: 14 lessons from a lifetime of boom and bust, falling costs have been a recurring feature of technological revolutions. As AI becomes cheaper and more accessible, adoption often accelerates as new use cases emerge.

How is open-source AI reshaping the competitive landscape?

The rise of open-weight and open-source models is reshaping competition. Many organisations are expected to adopt a mix of proprietary and open models, balancing performance, flexibility and cost.

“ Open AI models do not need to beat every proprietary model on every benchmark. They need to be good enough, inexpensive and ubiquitous.”

Adrian Cox, Senior Strategist, Deutsche Bank

For many business applications, organisations may prioritise cost, flexibility and ease of deployment over marginal improvements in performance. Models that are sufficiently capable and significantly cheaper could therefore accelerate adoption. This theme is explored further in Open-source AI 101: the battle for the future of AI.

What does history tell us?

AI's history offers an important reminder: transformative technologies rarely follow a straight path. While advances can happen faster than expected, their full economic impact often takes longer to emerge as businesses adapt processes, build infrastructure and develop practical applications.

Another lesson is the Jevons Paradox: making a resource cheaper and more efficient can increase overall consumption. As AI becomes cheaper and more capable, organisations are deploying it across a growing range of use cases.

“ When capable AI agents become cheaper, people do not use less AI, they find more ways to use it.”

Yi Xiong, Chief Economist, Deutsche Bank

Agentic AI, capable of planning, reasoning and carrying out multi-step tasks, could further expand the range of activities that can be automated.  This theme is explored in The Expanding Frontier of Agentic AI: A User's View.

Looking beyond the headlines

The bigger story is not who builds the most powerful model. It is how falling costs, intensifying competition and accelerating adoption shape the economy.

For a deeper exploration of these themes, read our latest research on AI's history, agentic systems and the growing role of open-source models on the Deutsche Bank Research Institute website.