AIoT, the convergence of intelligent computing and IoT, is seeing an uplift in intelligence in IoT devices which is enabling greater autonomy and automation as IoT devices contain capabilities to analyze data and make decisions on the device. This is particularly useful for constrained devices which have limited battery power and seek to minimize the need to transmit data to centralized processing capabilities. To unlock the full benefits of AIoT, a new generation of more intelligent, computing-capable devices is needed. However, this isn’t a case of simply bolting-on intelligent computing technology to existing devices.

There is a sophisticated set of variables to consider that are at the heart of efficient design and manufacturing of intelligent computing-enabled IoT devices. That sophistication demands a highly specific set of skills that many enterprises lack. To optimize AIoT, developers, engineers and designers that understand both intelligent computing and IoT are needed, and for many, those skills are in short supply or prohibitively expensive to recruit.

With AIoT connection volumes set to radically increase at a CAGR of 20% in the decade to 2033, according to Transforma Insights, it’s clear that upskilling is needed to meet demand, which the firm says will reach 9.1 billion devices by the end of 2033. The challenge is for IoT organizations that need to act now. How will they bring intelligent computing-enabled IoT devices to market to support current use cases such as predictive maintenance, IT automation, supply and logistics applications that are already being used to maximize productivity and competitiveness?

Complexity against the clock

An obvious answer is to streamline the device design and development process by working with an original design manufacturer (ODM) which can combine the hardware, software and manufacturing expertise needed to bring AIoT devices to market at scale. Rather than learning new capabilities, hiring technical experts, building up internal teams, and bearing the cost of doing so, enterprises can turn to an ODM which, in addition to those attributes, has the experience gained across multiple projects to bring to the enterprise’s product.

Intelligent computing presents a more demanding landscape than traditional IoT device engineering. AIoT devices must be able to balance power efficiency and inference performance while also considering the connectivity and real-time data processing requirements of the use case. Designers need to be aware of new developments and create devices that make use of the latest advances in intelligent computing chips and software frameworks. These are evolving rapidly so developing with future capabilities in mind is essential.

Optimized AIoT devices combine energy efficiency, bandwidth optimization, reliability, and inference accuracy. However, achieving this is complex. Few IoT organizations have the time or resources internally to execute across all those disciplines, so the ODM approach increasingly is seen as a means to accelerate time-to-market and reduce development and production friction.

The benefits of adopting an ODM approach have been detailed in a recent Quectel Masterclass titled ‘How partnering with an ODM optimizes intelligent computing device development’. The Masterclass presented by Vladimir Rakic, the Director of R&D for Europe at ODM Ikotek, and Vijay Prakash, the product manager for smart modules, India, at Quecel, details an ODM’s role in AIoT product development.

The Masterclass, which provides an introduction to AIoT, explores the impacts of edge computing, real-time decision-making and enhanced security in IoT. In addition, it explains the attributes of innovative AIoT solutions and how ODMs can rapidly accelerate successful device introduction. The Masterclass can be viewed here.