The hype around AI has been building for years, but if there’s anything we’ve learned, it’s that the term “AI” no longer refers to a single technology. Banks are preparing to launch assistants with emotional intelligence capable of detecting frustration or hesitation in the customer’s voice. Ambulances are being connected to AI-controlled traffic systems and equipped with real-time sensors that allow doctors to diagnose and intervene remotely. And in factories, agentic AI is no longer a mere assistant to human operators. Now, you are able to manage entire production chains independently, making critical decisions throughout the process. What unites these revolutionary developments are not just smarter algorithms, but the ability to move and process large volumes of data quickly – securely, reliably and wherever it is needed.
And herein lies a challenge. According to the Gartner’s 2025 Hype Cycletechnologies such as agentic AI, multimodal generative AI, and open source LLMs have passed the “innovation trigger” phase and are now at “peak inflated expectations.” Before the technology fully matures, it will need to cross the “valley of disillusionment” and the “ramp to enlightenment.”
The real story of 2026 will be infrastructure, not software. The success of the next generation of AI solutions will depend entirely on networks capable of offering ultra-low latency and global reach, meaning poor connectivity becomes a limiting factor. Without a foundation of high-performance, low-latency, densely interconnected networks, applications like emotionally adaptive AI, connected ambulances, and autonomous microfactories risk becoming stagnant at the peak of inflated expectations rather than becoming scalable, real-world services.
Here are some trends we can expect for next year:
The regulatory paradox of “emotional AI”
Financial services could become the unlikely testing ground for emotionally adaptive AI in 2026. I’m referring to a new generation of digital assistants that promise to capture the hesitation, stress or frustration in the customer’s voice and respond with greater empathy and consistency than current systems. Instead of a neutral bot, these algorithms can adapt the approach based on the customer’s mood and significantly improve the user’s digital experience.
According to analysts, the Emotional Artificial Intelligence (AI) market will to grow to $9 billion by 2030. But this will be an evolution that will face tensions. The EU’s AI Act classifies Emotional AI as a “high risk” technology, subjecting it to strict standards and oversight that will create a paradox: the volume of sensitive data generated is exploding but the freedom to use it appears to be diminishing. The processing of financial, biometric and emotional data will therefore require a sovereign, regulatory-compliant infrastructure that guarantees privacy, security and control at all stages.
Remote medical assistance and “live” ambulances
Healthcare has been a pillar of AI innovation for years, and in 2026 this trend will continue with life-saving applications that work in real time across entire cities. Ambulances equipped with advanced cameras and sensors will transmit high-resolution video and patient data directly to hospital specialists, allowing doctors to begin collecting patient data, performing diagnoses and guiding paramedics through complex procedures even before the patient arrives. There are also ongoing tests in cities such as Chennai, where AI-controlled traffic lights can give priority to emergency vehicles, ensuring that ambulances reach hospitals faster and with clearer routes.
However, turning pilot projects like these into large-scale programs will require a radical upgrade of digital infrastructure. Governments and hospitals are already looking to expand the use of technology beyond testing, driving market growth. 5G healthcare from $95 billion in 2024 to more than $362 billion in 2030. And to realize this use, not only robust 5G and Beyond-5G (B5G) networks will be required, but also the integration of satellites in low Earth orbit (LEO) to expand coverage and ensure their availability wherever an emergency vehicle is. Lives will depend on the instantaneous performance of this connectivity infrastructure.
Agentic AI and Autonomous Microfactories
Next year, manufacturing will become the new stage for demonstrating how agentic AI can manage entire industrial ecosystems. Instead of simply automating routine tasks, AI agents are now able to analyze, predict and act on large-scale production chains, from predictive maintenance and quality assurance to logistics and planning. One of the most disruptive applications will be AI-enabled 3D printing, which allows companies to mass produce personalized goods with unprecedented speed and precision. This comes in parallel with the trend toward offshoring production, as companies seek to reduce dependence on fragile global supply chains and rising transportation costs.
However, the implications of agentic AI go far beyond our planet. With the Project OlympusNASA is already experimenting with autonomous 3D printing facilities capable of building landing pads and habitats from lunar regolith (loose debris, dust, soil and broken rocks that cover the Moon’s solid rock). The reader may think this is a futuristic example, but it is revealing of how agentic AI can completely decouple production from traditional supply chains. But this future depends on the processing and instantaneous exchange of data across large geographic areas, and until we achieve this on our planet, the Moon will have to wait. Microfactories, for example, only work if they are tightly integrated into cloud services, computing edge and interconnection platforms capable of scaling alongside the AI-driven supply chain.
And what digital infrastructure is needed to support AI-driven industrial revolutions?
Use cases like the ones mentioned above require low-latency connectivity to AI models and cloud services, the integration of different network technologies, and the exchange of sensitive data between partners along industry-specific or business model-based value chains. Whether it is healthcare, manufacturing or financial services, existing network architectures and infrastructures will not be sufficient to enable future AI-based products and services to perform optimally regardless of users’ location.
As companies move their capabilities closer to their data sources, micro data centers will proliferate across urban and industrial landscapes. The orchestration of fiber optic, mobile and satellite networks into an interconnected mesh will be essential to ensure reach everywhere. Innovative Internet Exchange Point (IXP) interconnection solutions — from multi-network peering to exclusive private environments from select partners — offer businesses sovereignty, control, and high-performance connectivity for all AI use cases.
But there is beginning to be a new solution in the implementation of Internet Exchange Points for AI (AI-IXs). Secure access to AI and cloud workloads for training and inference, through the AI-IX concept recently launched by DE-CIX that can be combined with the ability to create closed and personalized user groups, separate from the public internet. Innovative interconnection solutions give companies across all sectors control over their data flows and high-performance, standards-compliant data transfer for every use case.
Emotional adaptive AI, connected ambulances and autonomous microfactories will continue to grow in 2026 and the demand for AI infrastructures and high-performance interconnections will only grow. AI workloads deeply integrated into the infrastructure and low-latency interconnections will drive the connectivity needed to transform these proofs of concept into complete applications that can benefit everyone, everywhere, at all times. This is why connectivity is the new economy and latency is its most important currency.