From Ai agents to spatial computing, tech trends for 2025 and beyond

And if 'agentic' AI has the potential to increase productivity within the organisation and promise a virtual workforce that can relieve and augment human labour, the next step will be the rise of the 'super agent', i.e. an orchestrator of multiple AI systems that will optimise their interactions. Some numbers.

According to Capgemini Research Institute, 32% of managers on a global scale identify agents as the main technology trend in data & AI for 2025.

Gartner, for its part, predicts that by 2028, at least 15 per cent of day-to-day business decisions will be made autonomously by 'agent-based' AI systems, and confirms how the goal-oriented capabilities of this technology will provide more adaptable software systems capable of completing a wide range of tasks. Finally, 60 per cent of executives and venture capitalists surveyed by the Capgemini Research Institute believe that general artificial intelligence (AGI) will reach maturity and become commercially viable by 2030.

AI is radically transforming the cybersecurity industry, opening the front for more sophisticated cyber attacks but also for more advanced protection systems. 44% of senior executives, according to a forthcoming Capgemini report, consider the impact of generative artificial intelligence as the top cybersecurity technology theme for 2025.

In the new year, Gen AI-enhanced attacks will become increasingly sophisticated and widespread, and to address these risks, efforts will be intensified to automate threat detection, in particular through threat intelligence systems, and cryptographic algorithms will be strengthened, with a growing focus on post-quantum cryptography, whose task is to protect data decryption risks of next-generation computing.

According to Gartner's predictions, by 2029, advances in quantum computing will make most conventional asymmetric encryption solutions insecure to use.

Another important emerging trend is security against disinformation, a category of technologies that aims to provide methodological systems to ensure integrity, assess authenticity and track the dissemination of malicious information.

The wide availability and advanced status of AI and machine learning tools used for fraudulent purposes will increase the number of such incidents targeting businesses. Gartner predicts that, by 2028, 50 per cent of companies on a global scale (today it is less than 5 per cent) will start adopting products, services or features specifically designed to address security use cases against misinformation.

The Capgemini Research Institute survey speaks for itself: 55% of executives and 44% of venture capitalists expect quantum computing to be one of the top three technologies in computing & networking, with a significant impact as early as 2025. Specifically, 41% of the executives surveyed expect to experiment with proof of concept quantum computing with limited use cases, while 27% expect the technology to be partially brought to scale in certain business sectors within the next twelve months.

When, then, is the real leap into the quantum era? While waiting for answers that are not yet there, other paradigms of computing technologies, from edge computing to neuromorphic computing, are emerging in parallel. Hybrid computing, analysts explain, combines different mechanisms of computational processing, storage and networking and will be used to solve complex computational problems and to create highly efficient transformative innovation environments compared to conventional environments, enabling technologies such as AI to overcome current technological limitations.

A further frontier of computing leads to spatial computing that will digitally enhance the physical world with technologies such as augmented reality and virtual reality. This is, according to experts, the next level of interaction between physical and virtual experiences and many organisations will benefit over the next five to seven years through simplified workflows and smoother collaboration.

By 2033, in particular, Gartner predicts that the turnover of spatial computing will grow to $1.7 trillion worldwide, up from $110 billion in 2023, and will also be fuelled by the increasing demand for sophisticated visualisation tools to support decision-making and the search for efficiency in sectors such as healthcare, retail and manufacturing.

The growing focus on combating climate change and energy consumption is progressively accelerating innovation in areas such as renewables, biofuels and low-carbon hydrogen. Nuclear power, in this scenario, remains a central theme for 2025 and will return to the centre of the corporate agenda, driven by the urgent need for clean, controllable and (above all) capable energy to support the needs of high-consumption technologies such as AI.

The development of small modular reactors, in particular, is expected to accelerate by the end of next year, while other priorities reported by experts include projects and investments for advanced modular reactors and waste management systems.

What Gartner calls invisible ambient intelligence, however, is a model enabled by ultra-cheap, small-scale smart tags and sensors that provides affordable, real-time tracking and sensing solutions at scale.

In the long term, this 'technology' unobtrusively integrated into the environment will enable deeper integration of perception and intelligence into everyday life, while until 2027 its first applications will focus on solving problems such as retail inventory control or perishable goods logistics, enabling low-cost tracking of objects and improving visibility and efficiency.

Advances in the field of artificial intelligence have already accelerated the development of next-generation robots, exploiting the innovations brought about by mechatronics and expanding their uses beyond traditional industrial applications. Looking ahead, the large-scale development of Gen AI will further foster the creation of humanoid robots and cobots (collaborative robots), i.e. machines that can adapt to different scenarios and continuously learn from their environment.

24% of senior executives and 43% of venture capitalists consider AI-powered automation and robotics to be one of the top three technology trends of 2025, according to Capgemini Research Institute. Moreover, by the end of next year, advances in natural language processing and machine vision will further enhance their capabilities, enabling robots to take on increasingly complex roles in industries such as manufacturing, logistics and agriculture.

With artificial intelligence taking on complex decision-making roles, the future of work could consequently see a change in the traditional leadership structure within organisations: the rise of AI-powered machines capable of imitating human behaviour, in other words, will prompt a rethinking of the role of humans in the company.

The other front of next-generation robotics calls for multifunctional robots, i.e. machines designed to work alongside humans by providing advanced capabilities to perform more than one task and to switch seamlessly from one task to another (and for this reason destined to replace robots designed to repeatedly perform only a single task).

According to analysts at Gartner, 80 per cent of people (inside and outside companies) will interact with intelligent robots on a daily basis by 2030, compared to 10 per cent today.