Demand for AI follows 'almost unlimited' even with companies demanding more return on investment

The appetite for artificial intelligence shows no signs of deceleration. Sector executives heard by CNBC state that the demand for AI infrastructure is "almost unlimited", even in the midst of an increasing movement of companies to measure more accurately the return on their investments. The expansion of data centers and the development of more efficient models continue at an accelerated pace, contrary to fears of a possible excess capacity.

Why is the demand for AI still so strong even with pressure for return?

The race to adopt artificial intelligence has become a strategic priority for companies in all sectors, from health to manufacturing. According to chip and data center executives, the need for processing to train and execute AI models grows exponentially, and the current offer still does not follow this pace. Even with the recent volatility in the shares of companies such as NVIDIA and AMD, the volume of orders to build new processing units did not decrease.

How do companies balance costs and ambition in the demand for AI?

The speech of "IA at any cost" is giving way to a more pragmatic approach. The companies are closely analyzing the operational costs and the financial return of each deployment. This has led to an optimization movement: instead of simply buying more capacity, many are looking for smaller and more efficient models for specific tasks. This trend was recently highlighted by Hugging Face CEO, who defended the exchange of expensive APIs for open source solutions — a movement that gave strength between startups and large companies.

Volatility of chip actions: warning signal for AI demand?

The oscillation in the stock price of semiconductor manufacturers has generated uncertainty in the financial market, but does not seem to reflect a real slowdown in demand. For executives, it is a natural adjustment after a period of euphoria. Building new data centers remains a strong indicator: companies like Google, Microsoft and Amazon continue to invest billions in infrastructure dedicated to AI. Meanwhile, a new generation of models — smarter and cheaper — is changing the way companies view technology, as they show initiatives such as the AI race swaps giant models for smart and cheap systems.

Frequently Asked Questions about AI demand

Can the demand for AI really be considered "unlimited"?

Industry executives use the term to describe the current mismatch between the installed capacity and the computational needs of companies, which continue to grow at an accelerated pace.

What sectors are driving this demand?

All sectors, but with emphasis on technology, health, finance and manufacturing, are investing heavily in AI-based predictive automation and analysis.

Can the focus on return on investment stop innovation?

Not necessarily. It is making companies seek more efficient solutions, such as smaller models and open source, which can democratize access to AI rather than inhibit it.