Zigor Gaubeca, CIO of Grupo Aire, offers a retrospective look at the trends of 2025 and anticipates the phenomena that will mark 2026, a year in which cybersecurity will become a critical element and the ethical use of AI tools will have to advance in regulatory frameworks.
The consolidation of autonomous AI agents
If we anticipated a transformation of automation in telecommunications through autonomous agents by 2025, the passage of time has confirmed its explosion. AI agent technology, once a niche market, has become "much more accessible to the general public," with companies like OpenAI and Anthropic launching their own orchestrators. Examples like Perplexity's "Comet" browser, which automates Amazon purchases, have already generated controversy, as these agents, now available to everyone, will force a redefinition of traditional business models, especially those based on advertising and digital consumption, by 2026.
Advances in post-quantum cryptography
The concern about the arrival of quantum computers, which in 2025 was a future threat, has transformed into an urgent need for action. Google, Microsoft, and IBM are accelerating their developments, and the ability to decrypt current data in the near future is no longer a hypothesis, but a tangible risk. "Data stolen today, even if encrypted, could be decrypted in X years, when quantum technology is available," warns Gaubeca. Grupo Aire is already conducting tests on photonic networks with quantum cryptography to get ahead of this scenario. By 2026, investment in PQC technologies will be critical, primarily in the areas of cybersecurity, encryption, and espionage, where the major battles of the future will be fought.
Advanced processors for AI and the geopolitics of the chip
NVIDIA has maintained its undisputed leadership, but 2025 has witnessed intense competition with Google, AMD, and Intel developing their own chips. The major trend is the shift from processing in large superdata centers to a distributed model across local devices (mobile, PC, IoT), thanks to the optimization of AI models to be smaller and more specific. Processors like Apple Silicon (M3, M4, M5) already demonstrate the ability to run large models locally. However, this hardware revolution is intimately linked to geopolitics: the manufacturing of next-generation chips, dominated by ASML (Netherlands) in lithography machinery and TSMC (Taiwan) in production, is a battleground. Control of this production has become a "war for supremacy" for the coming decades.
Responsible AI: the need for ethical regulation and transparency
Despite the intentions of the European AI Act to require transparency in the training of AI models, similar to ingredient labeling on food, 2025 has revealed significant lack of oversight in practice. Cases such as Elon Musk's "Grok" AI creating its own Wikipedia without respecting Creative Commons licenses, or the numerous lawsuits against Perplexity for unauthorized use of content, demonstrate the lack of effective regulation. Large companies do not disclose their training methods. By 2026, the effective implementation of the AI Act and jurisprudence surrounding intellectual property and the use of data in AI will be crucial to overcoming this unregulated situation.
NIS-2 Directive and the common European cybersecurity framework
Faced with the obligation to comply with the NIS-2 regulation, European businesses will raise their cybersecurity standards, reinforcing the importance of working within a common European framework on this matter. It is expected that the national transposition into Spanish law will be approved in the coming weeks, which will undoubtedly provide a significant boost to the protection of companies against potential cyberattacks.
Talent: continuous training and a focus on cybersecurity
The adoption of new technologies will be gradual, not immediate, and has spurred upskilling and reskilling programs, especially for AI developers. However, prioritizing productivity and speed is creating a real cybersecurity risk. By 2026, skills development must be intrinsically linked to unwavering cybersecurity robustness, or technological advances could generate even greater risks than they are intended to solve.
2026 is shaping up to be a year of consolidation for autonomous AI, a year of urgent adaptation to new cryptography, and a year of increasing tensions in the chip supply chain. For Grupo Aire, the key will be navigating these complexities with a strategic vision, prioritizing security, ethics, and human adaptability in a technological ecosystem that continues to surprise and challenge.
