Stay current: AI trends and early signals for 2026

Three forces shaping what comes next in AI

Here are three major AI focus areas for 2025: agentic AI, physical AI, and sovereign AI. This overview explains what they are, why they matter, and what experts expect to see in the near future.

Are you keeping up with what’s changing in AI? The pace is moving fast. New tools and new approaches are showing up constantly, and they are starting to reshape work, markets, and everyday life. Three trends stand out right now because they could drive the biggest changes: agentic AI, physical AI, and sovereign AI. Each one brings real upside, but each one also comes with risks that leaders need to plan for.

To understand where things may be heading, we asked AI leaders and decision-makers in several industries for their views. We also gathered input from a wider group of people on LinkedIn. The questions were not exactly the same, since one group could speak from inside organizations and the other offered a broader public view. What follows is a clear summary of these trends, why they matter, what people are saying, and early predictions for 2026.

What is agentic AI, and why does it matter?

Agentic AI is a type of AI that can act on its own to reach a goal. It can adjust when conditions change, make decisions that involve more than one step, and work alongside people or other AI systems. Instead of only handling simple, repeated tasks, agentic AI can manage full workflows that require planning and follow-through.

This matters because it could help organizations run faster and more smoothly. It may also open the door to new services and new ways of doing business. Just as important, it can take some of the heavy workload off teams so people can focus on higher-value work that needs judgment, creativity, or strong relationships.

How can agentic AI be used?

Agentic AI can be applied in many areas, including:

• Customer service: AI agents can sort support requests, solve common problems, and send only the hardest cases to a human rep.
• Supply chains: Agents can track demand, adjust inventory levels, and coordinate purchasing and shipping as conditions shift.
• Finance: AI agents can support portfolio oversight, spot unusual activity, and help monitor compliance tasks that require constant attention.

What are people saying about agentic AI?

Most organizations are still early in adoption. Many leaders report they are testing agents in small pilots, or they have not launched them at all. Large-scale rollouts are still uncommon. When teams do report broader use, it tends to be in larger companies, often in industries that already invest heavily in advanced technology.

Public expectations, however, are more aggressive. In LinkedIn responses, close to half of participants said they expect autonomous agents to change their organizations in the next two to three years. Only a small share believed agents will make no difference in that time. Most others expect a limited or moderate impact, rather than a complete overhaul right away.

Three early predictions for 2026: Agentic AI

1. Pilots become everyday tools.
   More companies will move agentic AI from small tests into real workflows. Larger organizations will likely lead, since they have more funding and more technical staff. At the same time, more “ready-to-use” agent solutions will hit the market, making it easier for other industries to adopt them.
2. Rules and oversight become a priority.
   As AI agents take on more responsibility, organizations will tighten their guardrails. Expect clearer policies, stronger approval steps, and better tracking of what agents do and why they did it.
3. New roles and training take off.
   Companies will invest more in training so employees can work well alongside agents. Some teams will specialize in running and supervising agents, including monitoring performance, updating agent behavior, and making sure usage stays safe and compliant.

What is physical AI, and why does it matter?

Physical AI brings AI into the real world. It is when machines can sense what is happening around them, understand it, and take action. This often blends AI with robotics, self-driving systems, smart sensors (IoT), and digital twins.

It matters because it can improve safety and efficiency in places where automation used to be too expensive or too risky. Physical AI can also help in settings where conditions change quickly and humans cannot watch everything at once.

Common uses for physical AI

Physical AI already shows up in several areas:

Manufacturing: Smarter robots and automated quality checks can reduce defects and slowdowns.

Logistics: Autonomous vehicles, robots, and drones can speed up moving and sorting goods.

Health care: Wearables and sensors can track patients in real time and support more responsive care.

What are people saying about physical AI?

Many leaders expect physical AI use to grow, but not explode overnight. Adoption tends to move slower because hardware is costly, safety requirements are strict, and maintenance is ongoing. On top of that, companies still have to deal with regulations, older infrastructure, workforce training, and whether the public trusts these systems.

Public opinion is more mixed. A sizable group expects only a small impact in the next few years, but more than half expect moderate to major changes. A smaller group expects no impact at all. The most realistic outlook is that physical AI will hit some industries faster than others, especially where the ROI is clear and the environment is easier to control.

Three early predictions for 2026: Physical AI

1. Adoption will be uneven, but fast where the economics work.
   Physical AI is likely to take hold first in asset-heavy, task-intensive sectors like manufacturing, logistics, health care, and agriculture. These environments often have repeatable workflows, measurable productivity gains, and clearer ROI. In contrast, many knowledge-based and service-led industries may move more slowly when work is primarily digital, highly interpersonal, or tightly constrained by privacy, security, and brand-experience requirements.
2. Safety and security will become the gating factors.
   Organizations that deploy physical AI will need to treat safety and security as non-negotiable requirements, not add-ons. Expect a blend of straightforward physical safeguards (for example, emergency-stop controls, proximity sensors, and restricted zones) paired with fail-safe software behaviors, strong cybersecurity, and clear logging so systems can be audited after incidents. The standard will shift toward “prove it is safe” before scaling.
3. Human-machine teaming will become a practical operating model.
   Workforces will start learning how to operate alongside physical AI as integrated crews. The goal will be to move humans toward higher-value judgment, exception handling, and relationship-based work while machines handle repeatable execution. Companies that invest in intuitive interfaces, targeted reskilling, and change management will transition faster and with less friction.

What is sovereign AI, and why does it matter?

Sovereign AI is the idea that an organization’s data, model assets (including weights), and compute can be kept within specific national or regional boundaries. It matters because privacy rules are tightening, geopolitical concerns are rising, and organizations want more control over where sensitive information lives and how it is processed. Done well, sovereign AI can reduce regulatory exposure, strengthen customer and partner trust, and lower dependency on external providers that may be subject to foreign laws or cross-border disruptions.

Where sovereign AI can have the most impact

Sovereign AI tends to matter most anywhere data is sensitive or regulated, including:

• Health care: Keeping patient data processed and stored locally to meet privacy and compliance requirements.
• Finance: Ensuring transaction records and model operations remain in-country to satisfy regulators and risk controls.
• Public sector: Designing systems that support transparency, local accountability, and national control over critical digital capabilities.

What people are saying about sovereign AI

Sovereign AI is increasingly viewed as a strategic planning issue, not just a technical preference. Importance varies by industry, but urgency tends to rise sharply in highly regulated or high-stakes sectors such as banking and insurance, telecom, energy and industrials, and life sciences and health care. In these environments, strict compliance expectations, critical infrastructure concerns, intellectual property protection, and national-security sensitivities push organizations toward local control of data and models.

More broadly, sovereign AI is also becoming a board-level conversation because cross-border data disputes, cyber risk, and evolving policy expectations are making “where data and compute live” a material business decision.

Three early predictions for 2026: Sovereign AI

1. Regulatory pressure will intensify, with a complicated picture in the US.
   More jurisdictions will introduce or tighten rules around data privacy, security, and AI governance. The US outlook may remain less predictable as proposals develop across states and federal channels. Organizations that move early on compliance foundations, such as transparency practices, monitoring, and documentation, will be better positioned to avoid last-minute scrambles.
2. Early demand will rise for sovereignty-ready solutions.
   Organizations will increasingly look for solutions that can satisfy local data and compute constraints. This will push adoption of multi-cloud designs, in-region deployments, and edge strategies where appropriate. Providers that can meet sovereignty requirements without sacrificing performance will gain an advantage in regulated markets.
3. Regional AI ecosystems will accelerate.
   More regions will invest in local AI capacity, including infrastructure, talent pipelines, and partnerships, to reduce reliance on external capabilities and keep economic value closer to home. These hubs will compete to attract capital and talent while building domestic resilience.

‍

‍