If 2025 sent imaginations running wild with AI’s potential, 2026 is the year organisations must take stock, adding the foundational building blocks to make those dreams reality. From infrastructure overhaul to talent nurturing, quantum internet and trust issues – Cisco CTOs look back to look ahead to understand what the next year in tech might look like…

1. An Infrastructure Inflection Point

A quiet but critical conflict hides beneath the glittering promise of the AI revolution: the collision of legacy “technical debt” with emerging “AI infrastructure debt.” In the race to deploy AI, many organisations are stacking quick fixes and scattered data on top of aging infrastructure. The result is a growing liability: smart systems running on foundations that were never built for today’s speed, scale or security demands. The result? Organisations risk stunting innovation and growth.

But this stark warning is also an opportunity. The year ahead will be defined by those who modernise their fundamental network infrastructure, and we may see those the furthest behind, with the most to upgrade, leapfrog their competition. Prioritising a “secure-by-design” overhaul today, will do more than pay off the debts of yesterday. It will build the resilient, AI-ready backbone to power a safer, faster, transformative future.

2. AI Moves to the Edge

Data fuels AI, but we’ve barely started to access what truly exists. With 22.4 billion IoT devices generating more than 90 zettabytes a year, 2026 will see organisations finally tap into the vast well of telemetry, machine, IoT and IIoT data. AI can analyse and combine these streams in ways humans can’t, by training domain-specific models that could reshape industries as dramatically as generative AI did. To enable these models, 2026 will bring a shift toward AI at the Edge – preserving privacy, critical in industrial environments.

Manufacturing, energy and logistics teams already use IIoT data to cut downtime and improve efficiency. This adoption will accelerate in 2026, marking the second phase of AI’s evolution. This shift is powered by advances in specialised AI chips, TinyML, for ultra-efficient on-device inference, while federated learning trains models across distributed edge devices without centralising sensitive data. Embedding security into the infrastructure will be essential to protect these workloads as they scale.

3. Digital Sovereignty in Practice

Digital sovereignty won’t slow innovation. It’ll redefine where and how it happens – shifting from theory to execution in 2026, as tighter data-localisation laws take hold. Nations and blocs will assert control over their infrastructure, data and technology stacks, reshaping the digital landscape. The push for sovereignty will extend to AI, accelerating investment in domestic capacity, sovereign compute and AI factories building models trained on sensitive regional data.

Demand for sovereign cloud solutions will rise, along with greater reliance on regional providers and renewed interest in on-premises or air-gapped data centers. A full overhaul of global infrastructure is unlikely, but selective migrations and diversified cloud strategies will become the norm, creating demand for local talent and skills.

4. Identity as the New Perimeter

Deepfakes, transparency gaps, bias and accountability issues have made trust a prerequisite for AI adoption. Critical systems – physical, digital and everything in between – need protections that scale with distributed workloads and a blended human–digital workforce. One path forward is embedding security and observability directly into the network, creating a safety layer that continuously monitors AI models and agents.

With 82% of EMEA organisations planning to deploy AI agents (Cisco AI Readiness Index 2025), identity management will become a defining trend. And with AI agents shifting roles instantly, traditional identity systems won’t cut it, raising the need for purpose-built identity frameworks to authenticate and trust an AI agent

While identity is integral, trust goes even further. As the line between humans and AI agents blurs, organisations must govern the human–agent pair: who’s in charge, what they can access, how is their behaviour monitored, and what happens when things go wrong?

Recommended reading

• Will AI Agents Be the Enterprise Disruptor of 2026?
• What Will CFOs Prioritise in 2026?
• Comment| Five Predictions That Will Shape the Digital Assets Sector in 2026

5. The Human Systems Upgrade

On any single day, AI can be framed as both a job taker and an economic boom bringer. That binary focus hides a deeper problem: the workforce isn’t ready to realise AI’s potential. Companies race to deploy agents, yet their human systems – hiring, career paths, skills development – remain rooted in a pre-AI era. This isn’t a failure of people; it’s a failure of workforce infrastructure. To fuel innovation, organisations must reskill, retool and redesign the pathways and support systems their teams rely on.

2026 will demand more than basic AI literacy. Companies will need a full-stack curriculum spanning the whole career ladder – covering networking and cybersecurity fundamentals, to data science, vibe coding and advanced AI capabilities. Forward-looking organisations will focus on building skills internally, nurturing talent they already have. The leaders of the next decade will evolve beyond the “tech vs. non-tech” binary, creating hybrid experts who orchestrate intelligence rather than simply generate inputs. Democratising technical depth ensures that augmented employees become architects of the AI era.

6. Quantum in the Real World

Quantum computing is shifting from “can we really do this?” to “what can it unlock?”, with complex problems in medicine and physics among the primary targets. The race to quantum-safe infrastructure will intensify, with organisations investing in post-quantum cryptography and regional quantum innovation hubs.

In 2026, Cisco engineers will continue working on a network built on the unique behaviour of quantum particles, to connect quantum computers and share information securely. This year, Cisco introduced an entanglement-source chip that generates millions of entangled photon pairs per second, enabling quantum communication over existing fibre without specialised infrastructure.

A distributed, scalable quantum network could unlock a vast new computational space and support entire classes of emerging technologies. By the late 2030s, this work may culminate in a quantum internet connecting quantum computers, sensors, and other devices – opening possibilities such as ultra-secure communication and precise monitoring of climate, weather and seismic activity.