What's Trending?
AIoT
EDITORS TAKE, JANUARY 2026
COMMENT
Has AIoT sparked your imagination?
This month at IoT Now, we’re focusing on AI – and who isn’t?
We’re not getting stuck in an endless doom loop of generative AI-created hallucinated responses to questions and we’re not wondering if any of AI is affordable or sustainable from data centre capacity, power and cooling perspectives. Instead, we’re focusing on AIoT which has the advantage of being already here as well as being able to pay for itself in the business benefits it delivers.
AIoT isn’t simply the force-fitting of one acronym to another, although it is worth pointing out that AI frameworks and architectures are quite different to IoT platforms and systems, so significant integration efforts are required to get to the prizes. And yes, there are prizes to be won in the form of increased productivity, reduced downtime, enhanced capacity, faster and richer responses and, perhaps counter-intuitively given the cost of AI, reduced operational expenses.
Use cases such as predictive maintenance, IT automation and supply and logistics dominate currently but we’re at the start of the next wave of deployments which will see further automation, enabling some offerings to become viable for the first time. It’s a truly exciting period and IoT looks set to reap the benefits of AI with precision and focus. That’s because in IoT, AI is not a technology looking for an application, it’s a critical enabler that the IoT industry already recognises as the bringer of value.
All that’s needed are more ideas that benefit from combining AI with IoT and, with 9 billion AIoT connections expected by 2033, it’s clear these are already in the works.
If you’d like to interact with a comment, question, argument or additional info on this topic, leave a Voice Note at [LINK] and we’ll discuss it further in a forthcoming audio presentation.
By the numbers
The potential for AIoT is significant
0bn
AIoT connections forecast globally by 2033
0%+
Annual growth rate for AIoT connections to 2033
0bn
Total IoT devices forecast worldwide by 2033
0%
IoT devices expected to include onboard AI by 2033
0%
Organisations using AIoT for predictive maintenance
0%
Enterprises planning AIoT adoption
Market report
AIoT combines the best of both worlds to re-ignite use case innovation
AIoT, formed by combining AI and IoT, is a relatively new term but not a new activity. Intelligence, in various forms including edge and artificial, has been utilised by AI use cases in IoT since IoT itself was conceived. The application of intelligence is essential to aid greater automation, control costs and improve precision, writes George Malim
Amid all the AI hype, a good place to start is by addressing what AIoT actually is. It isn’t simply edge intelligence or the application of generative AI to IoT processes mashed together into another acronym. It’s the integration of AI and IoT that allows devices to not only collect data but analyse it and make autonomous decisions in real-time.
“Let’s be clear, AIoT is the real deal,” says Nick Earle, the executive chairman of Eseye. “It’s far more than just a marketing wrapper for edge intelligence. We define it as the catalyst for sentient AIoT, a fundamental shift that’s rewriting the connectivity playbook.”
“When we say ‘sentient’, we mean the device is no longer just a passive data collector,” he adds. “It is self-aware, it knows when its connection is failing and has the autonomy to fix it without asking a human or the cloud for permission. This establishes intelligence across three critical areas: the device, the application and the network. It moves us away from dumb pipes and reactive systems to proactive agents that verify their own data in real-time.”
Truly intelligent IoT devices
Far from a marketing wrap, Syed Zaeem Hosain, the co-founder and chief evangelist at Aeris, sees AIoT as transformative. “We see AIoT as a genuine evolution of IoT rather than a marketing construct,” he confirms. “While the term itself comes from combining AI and IoT, its real significance lies in deploying AI models directly on IoT devices, allowing them to analyse data and take action locally instead of relying entirely on cloud processing.”
“This approach is increasingly necessary as IoT systems generate billions of messages every day, making it impractical and costly to send all data to centralised servers,” he adds. “By processing data at the edge, AIoT enables faster real-time action, lower operating costs, improved privacy and security, and greater resilience. These benefits are difficult to achieve with cloud-only architectures. In short, AIoT makes IoT devices not just connected, but truly intelligent.”
Frank Jones, the chief executive of IMS Evolve, sees numerous use cases coming to fruition across the IoT landscape. “AIoT goes far beyond edge intelligence,” he explains. “AI is a powerful addition to the capabilities that IoT provides, creating IoT systems that don’t just capture data but that interpret it to recognise anomalies, make decisions and trigger automated responses instantly. AIoT will result in IoT devices running more effectively, with data-driven decisions being made continuously and intelligently.”
IoT users turn to AIoT
The findings of a recent IDC InfoBrief, sponsored by data and AI specialist SAS, have uncovered that AIoT usage is pervasive across IoT. Predictive maintenance is the most widely adopted AIoT use case, selected by 71% of respondents, while IT automation and supply and logistics were the next most cited AIoT use cases, chosen by 53% and 47% of respondents respectively.
AIoT benefits are yet to fully crystallise with the survey uncovering that 63% of respondents believe AI will boost productivity and competitiveness, 54% anticipating significant cost savings, 52% expecting smarter and faster innovation and 49% predicting streamlined operations thanks to their AIoT investments. Perhaps the clearest signal is that 62% of respondents say they have already adopted AIoT with 31% planning to do so. The survey also uncovered that 43% have achieved widespread or fully integrated deployments.
“Our research found that heavy users of AIoT were almost twice as likely to report benefits that significantly exceeded expectations, while less than 3% of industrial executives surveyed said AIoT’s value did not meet expectations,” said Kathy Lange, IDC research director for AI Software. “The takeaway is clear: AIoT is fueling innovation, streamlining operations and driving smarter, faster decisions.”
Real world use cases proliferate
The IDC research was based on a global survey of more than 300 industrial executives in the manufacturing and energy industries, which are among the early adopter segments for AIoT. Jason Mann, the vice president of IoT at SAS, says the survey results reflect what he is seeing among customers. “This IDC InfoBrief confirms what manufacturing and energy customers are telling us worldwide: AIoT has evolved from a buzzword to a potent technology and business imperative,” he says. “Whether enhancing the predictive maintenance of critical equipment or improving operations across factories and electric grids, AIoT drives major cost savings, quality improvements and efficiency gains.”
The gestation from experimental to practical deployments has already completed for many. “The value is no longer theoretical, it is happening right now in the field,” confirms Earle. “We are seeing devices essentially healing their own connections. We are also seeing visual anomaly detection. AI is now being trained to spot operational irregularities instantly, like a defect on a fast-moving production line or a safety hazard in a public space.”
“The early adopters are the sectors where ‘good enough’ simply isn’t acceptable, where responsiveness and zero-latency are critical,” he adds. “Healthcare, manufacturing, smart cities and autonomous vehicles are the clear leaders here. They cannot wait for a round-trip to the cloud to make a decision. In terms of use cases, intelligent monitoring is at the forefront. Whether it is a security camera or an industrial sensor, the ability to identify and alert on an unusual behaviour instantly, rather than waiting for post-event analysis, is a game changer.”
AIoT opens up new opportunities
Maturing technology means that AIoT can be deployed at scale with confidence and that is enabling new approaches to traditional businesses and processes. “AIoT is already delivering practical value across a wide range of use cases, including connected vehicles, alarm systems, agriculture, smart homes and predictive applications,” says Hosain at Aeris. “In these scenarios, local intelligence enables real-time decision-making and automation that would be difficult or inefficient using cloud-only approaches.”
“In agriculture, for example, AIoT-enabled drones can autonomously identify and treat specific areas of crops, reducing chemical usage, labour requirements and overall operational overhead,” he adds. “In infrastructure and monitoring scenarios, intelligent cameras can analyse video locally and transmit only relevant insights rather than raw footage, significantly reducing bandwidth usage, deployment complexity and transmission costs. These examples show how AIoT can both lower operating costs and enable solutions that would otherwise be impractical at scale.”
Colin Grealish, the director of product line management for IoT & Data Intelligence at Motive, sees a large and exciting landscape of opportunities. “Applying human approaches at scale using artificial intelligence fundamentally changes what is possible in IoT,” he says. “Troubleshooting is a prime example. By its nature, troubleshooting is rarely straightforward. In complex IoT ecosystems, such as smart metering, failures can occur at many points in the service chain. These may include the meter itself, the modem, the base station, the cellular network, communication protocols, the device management platform or the application server. When performed manually, diagnosing these issues depends heavily on individual experience, intuition and the selective use of diagnostic tools.”
“AI changes this dynamic,” he explains. “By analysing vast volumes of records, logs and protocol traces, it enables rapid, accurate, root cause identification. Enriched with additional contextual data such as network status, end-to-end service views and component health, AI-driven troubleshooting also provides holistic visibility across the entire IoT ecosystem. This capability is a key to navigating complex IoT systems and proves that AIoT is far more than a marketing term; it is a critical enabler transforming processes across IoT systems.”
Multi-billion AIoT device connections
For Jones at IMS Evolve, the number of use cases unlocked by the convergence between IoT and AI is growing constantly. “Industries and businesses with large IoT estates are currently leading the charge as they can see the greatest benefits via economies of scale,” he says. “Use cases such as predictive intelligence, real-time response and autonomy are all ultimately leading to greater business value for industries such as retail and facilities management.”
Although there’s still some difficulty in nailing down the size of the AIoT market, it’s clear the growth curve is poised to accelerate. Transforma Insights has forecast that total AIoT connections will grow from 1.4 billion at the end of 2023 to 9.1 billion at the end of 2033. This is a more than six-fold growth in a decade, resulting in a CAGR of more than 20%. Overall AIoT represents a significant market with net additions growing from less than half a billion in 2023 to just over 900 million in 2033.
For context, the firm forecasts a total of 39 billion IoT devices at the end of 2033, up from 16 billion at the end of 2023. As illustrated in the chart, AIoT devices 2023-33, the firm has reported that 9% of IoT devices will have onboard AI in 2023, rising to 23% in 2033. The rate of growth of AIoT penetration of IoT devices slows towards the end of the forecast period, primarily due to the AIoT penetration of key IoT applications reaching saturation, the firm says.
With that level of growth in prospect and an already mature set of real world deployments that are generating value, AIoT is a combined technology that transforms IoT capabilities provided that the business case can sustain the cost of AI. Excitingly, this is less of a problem in IoT, where processes are typically tied to valuable outcomes, than it is in the consumer world. For many in IoT, AI opens the door to the previously impossible and the future is only really limited by what can be imagined.
INTERVIEW
AIoT enables IoT deployments to be deployed, optimised and operated at hyperscale, which wouldn’t have been possible before.
Erik Brenneis
Chief executive of Vodafone IoT
IoT Now: How do you define AIoT? Is it the real deal with AI tech integrating with IoT, or is it just a marketing wrap for edge intelligence?
Erik Brenneis: AIoT is the convergence of AI and IoT, and it represents a big step forward in how IoT is deployed and operated. It’s certainly more than a marketing wrap – it opens up a whole new set of use cases for IoT.
What makes AIoT so powerful is its ability to deploy AI through the whole IoT value chain: in the device and sensor, in the network, in the application, in analytics and in the processes used to drive IoT services. IoT is also inherently data-centric and highly automated, so it’s a natural fit for AI from the outset. When put together, AIoT enables IoT deployments to be deployed, optimised and operated at hyperscale, which wouldn’t have been possible before.
IoT Now: AIoT, in contrast to other AI deployments, appears to be closely tied to business benefits with tangible value already being created. What examples have you seen of this?
EB: At Vodafone IoT, we’ve been using AI within IoT for quite a while before the term AIoT was coined. In that time, we’ve seen – and continue to see – significant benefits in how it allows us to monitor, configure and optimise our network.
A good example is our advanced, next-generation network platform that uses AI to continuously monitor our network. This allows us to detect, alert and respond to issues in near real-time, ensuring our devices always stay connected and services remain reliable.
At the edge, we’ve seen AI deployed in what we call ‘Sight as a Sensor’. By combining high-resolution video, IoT connectivity and AI, a single sensor can detect many things all at the same time – whilst remaining secure and observing privacy. This is especially valuable in more chaotic IoT environments, like sorting waste, food products or pallet inspections – the sort of systems that never know what the next object will be. AIoT brings order to what would otherwise be very chaotic processes.
Finally, as IoT generates so much data, AI can be used to process and, more importantly, model insights from the data pool. This finds even more value from IoT as it allows organisations to optimise their fleet of connected devices and provide advanced services such as adaptive maintenance and building controls. As we move to a contestable IoT world where connectivity providers can be selected over the air using new standards, this will be even more important.
IoT Now: Are AIoT benefits truly quantifiable in financial terms, or are softer benefits like enhanced productivity or improved competitiveness factors that are driving adoption?
EB: The short answer is yes – the benefits of AI are very much measurable. In many cases, it enables entirely new use cases that weren’t possible before, while also delivering significant improvements to internal processes. Our use of AIoT has already delivered tangible and measurable benefits in terms of continuous process improvement, like streamlining customer engagement and lead generation, as well as supporting performance and network and platform operations.
As a managed service provider operating at scale, Vodafone IoT is in a strong position to demonstrate these benefits early. We can capture value across the entire IoT supply chain and at a scale where automation is essential, so we can quantify the impact clearly, rather than relying solely on softer productivity or competitiveness metrics.
IoT Now: Which industries and use cases are leading the charge in AIoT adoption?
EB: Early adopters tend to be organisations with very high data-processing requirements or those operating complex, distributed networks – the likes of energy, water, transport and road infrastructure businesses. In data-intensive environments like these, AI is already being deployed at the edge and centrally to pre-process, filter and optimise traffic. This includes medical applications as well as video and CCTV use cases, where data can be pre-processed in numerous stages in the process.
Another major group includes industries managing large networks of physical assets, like fleets of distribution vehicles and energy, road and rail networks. In these use cases, AIoT provides both the sensing layer and the intelligence needed to optimise operations across the entire network operation.
IoT Now: How do you see AIoT developing in the future?
EB: In my view, the next phase of AIoT will be a shift from IoT as a system to IoT as an ecosystem. Today, devices within a given use case typically connect back to a central head-end platform. But with AIoT, devices will become more autonomous and capable of communicating directly with one another through peer-to-peer interactions.
This evolution will take IoT into a new dimension, where devices can discover others within their ecosystem, authenticate securely, and transact independently. Ultimately, the result will be a fundamental change in how IoT solutions are designed, deployed and operated.
EDITORS TAKE
When Intelligence Leaves the Cloud
In this first episode of Hot Takes with George Malim, we put the microphone directly in front of the industry. Short, unfiltered opinions on what is really happening in IoT, what delivers real value, and what is still more promise than progress. Expect candid takes, disagreement, and perspectives you do not always hear on stage or in press releases
