- Half of companies think that the advantages of AI prevail over its environmental impacts
- Many still have trouble seeing tangible results or a high return on investment
- Energy consumption continues to increase in the middle of the adoption of AI
Many companies are apparently happy to invest in AI without necessarily considering negative implications, including environmental effects, according to new research.
A Capgemini report says that the three -fifths (61%) of world organizations plan to increase the expenses of the Genai over the next year, and this occurs after a year when 88% have already increased the expenses of the Genai.
To this air, more than half (51%) think that the advantages of generating AI will prevail over its environmental impact – a more pronounced feeling in the United Kingdom (56%).
Do companies invest blindly in AI?
With the use of the arrow in recent years, the emphasis has been largely put on companies behind the AI tools that we use, not the companies that use them.
For example, Google recently discovered a 27% increase in its energy consumption in the data center in the last half of 2024 and the first half of 2025. Microsoft saw a 168% increase in total energy consumption between 2020 and 2024.
Companies do not only fail on sustainability – one in five (21%) are not satisfied with their results so far, and many note “invoice shocks”, where the efforts of the AI scale go beyond initial projections.
“Rapid adoption does not necessarily result in a large -scale deployment with a tangible nurse return,” said CTO Franck Greverie.
For the future, more and more companies are turning to models of small languages to increase profitability, with 12% of IT budgets on average dedicated to a generative AI alone.
However, investing freely in AI is not the most effective solution. Greerie explained that “companies must set up a solid data foundation, in a confidence of confidence confidence, secure and ensures the necessary confidentiality” to see the biggest return on investment.
For the future, Capgemini suggests a more targeted approach to maintain a mastery of the environmental and economic impacts of the AI, and this started with the establishment of governance policies (which only 46% have done) and to focus on high impact areas such as customer operations, marketing and risk management.
