AI trends for 2026: From hype to productive application

CANCOM Slovakia: You work in the Competence Center Cloud and Datacenter at CANCOM. What exactly is that?

Eva: The Competence Center Cloud and Datacenter is a specialized unit within CANCOM that is centrally organized throughout Germany. My focus here is on AI infrastructures. The tasks of our Competence Center are very diverse: we take care of innovation development, portfolio design, and strategic cooperation. Specifically, this means that we determine which focus and growth topics should be driven forward by our business units. To do this, we carry out analyses based on customer, manufacturer, and market trends, develop suitable solutions and services, and also ensure that these are adequately promoted. We support colleagues from the various business units in complex customer projects and tenders and create synergies between experts and strategic partners.

CANCOM Slovakia: So you have your finger on the pulse. Looking ahead to the coming year, what AI trends do you see for 2026? How will the coming year differ from the current one?

Eva: Artificial intelligence has developed rapidly in recent years. A striking example was the release of ChatGPT at the end of 2022, which suddenly brought large language models into the spotlight. Since then, innovation cycles have continued to shorten, and it feels like there is something new every week. Gartner’s “AI Hype Cycle” provides a good guide to these innovation phases: a new technology starts with a trigger, reaches the “peak of inflated expectations,” i.e., the hype peak, then falls when the high expectations are not immediately met, and finally stabilizes at a productivity plateau. This is exactly what we are seeing with ChatGPT. After the hype in 2022, 2023/2024 saw a phase in which companies worked harder to make generative AI applications truly productive – moving away from pure fascination toward concrete implementation.

The big trend in 2025 will be AI agents, i.e., systems that perform tasks independently. In 2026, this hype will level off somewhat. Achieving widespread application requires time, resources, and learning processes. 2026 will therefore be the year in which many technologies, such as AI agents, will be further developed internally within companies. The focus will be on refining existing applications, using them productively, and learning from experience – through trial and error. This is normal, as AI is still a young and complex technology.

CANCOM Slovakia: What are your hopes for AI in the coming year?

Eva: I hope that 2026 will bring noticeable efficiency gains, especially in the area of enterprise AI. Until now, many companies have been using AI for relatively simple tasks such as text summarization. I hope that we will see more real productivity gains in the coming year – in other words, that employees will really be relieved of some of their workload. This is not yet the case across the board, but I believe we will reach the next level in 2026. So-called super AI or artificial general intelligence (AGI) will not yet exist; that will certainly take another five to ten years. But hopefully we will already be able to observe smaller efficiency gains, targeted applications, and productive uses in 2026.

CANCOM Slovakia: Where do you see the biggest challenges and risks in the coming year?

Eva: In my area—AI infrastructure—I see two major priorities: providing suitable infrastructure and ensuring its sovereignty. Added to this are limited resources. Many companies are currently faced with the question of where to operate their AI systems:

• In a large public cloud with a hyperscaler, where resources can be rented flexibly, but costs can quickly rise.
• In a private cloud with a German provider.
• Or completely locally – with their own “mini supercomputer” on site, which, however, can pose physical security risks.

The key is not just to look at the costs, but to have a reliable service provider or a stable internal platform that guarantees security, data protection, and performance.

Another issue concerns European resources: cloud and computing capacities in Europe are limited. Some large manufacturers hardly offer any AI computing resources or GPUs anymore – a veritable race for computing power is emerging. That is why providers such as CANCOM are relying on private AI cloud infrastructures. We have our own private cloud infrastructure supported by Hewlett Packard Enterprise (HPE), which is operated in our data center in Hamburg – a secure environment for AI workloads that is hosted entirely by CANCOM. This “private cloud AI” is based on powerful rack servers with GPUs that are specially designed for AI applications. This enables us to offer secure, flexible, and independent AI workloads – an offering that is currently attracting a great deal of interest.

CANCOM Slovakia: What would you recommend to companies that want to get started with AI?

Eva: I recommend seriously considering setting up a sovereign AI infrastructure in a data center. In my opinion, this option is currently the most secure and practical solution. A dedicated server in a data center is part of a protected network and can be specifically shielded from external risks. This allows companies to develop, test, and operate AI applications without being dependent on public cloud providers. Physical security risks – such as theft or unauthorized access – are also largely eliminated.

However, the costs should be estimated realistically: depending on the equipment, especially the number of modern GPU cards, an AI server can quickly become expensive. However, it also depends on the customer’s use case and the right sizing. There are many variables to consider here to ensure that the customer gets the right solution, which is neither too small nor too large and therefore cost-intensive. The investment costs can be a barrier to entry at first, but they pay off in the long term through security, control, and independence.

CANCOM Slovakia: Which industries will be particularly strongly influenced by AI in 2026?

Eva: The transformation will be particularly evident in healthcare and manufacturing.

An incredible number of new use cases are currently emerging in the healthcare sector. AI supports doctors in diagnostics, for example, by analyzing medical images or detecting abnormalities that are barely visible to the human eye. It also plays a central role in personalized medicine – for example, in the development of individualized therapies. AI also helps to automate administrative processes and reduce the workload for medical staff.

Developments in the manufacturing sector are equally exciting, especially in industry. Machines generate huge amounts of data that can only be meaningfully evaluated with AI. AI enables predictive maintenance, quality control, and efficiency gains in production. IoT devices form the basis for this – they connect the physical and digital worlds.

CANCOM Slovakia: How would you assess the current willingness of German companies to invest?

Eva: That depends heavily on the respective strategy and maturity in dealing with AI. Companies that already employ data scientists or AI engineers show a high willingness to invest. They know that technological leadership is only possible through early investment. When we talk to such companies and show them that CANCOM, as an Nvidia Elite Partner, has the latest NVIDIA Blackwell GPUs that are ready for use within a few days, price often plays a secondary role – it’s all about speed and competitiveness. The situation is different for companies that do not yet have a dedicated AI department. They are often cautious, especially in economically challenging times. Many want to observe before they invest.

However, I recommend that every company include AI in its budget and resource planning today. Artificial intelligence is not a fad, but a key technology. If we hesitate too long in Germany, we risk falling behind.

CANCOM Slovakia: If you could choose a headline about AI for 2026, what would it be?

Eva: “AI makes it into small and medium-sized businesses – innovation becomes adaptation.” For me, this headline sums up two things: On the one hand, I would like to see AI really take off in small and medium-sized businesses – i.e., in companies that are often still hesitant. On the other hand, it stands for the shift from theory to practice: away from pure innovation and toward real application. For me, that would be the best signal for 2026.

CANCOM Slovakia: Thank you very much for the interview and your assessment.