IN Brief:
- AI systems often lack timely, item-level information about products moving through physical supply chains.
- Flexible integrated circuits can bring identification and connectivity to low-cost, high-volume everyday items.
- Combining physical item data with conventional computing could support more responsive inventory and operational decisions.
In recent years, artificial intelligence has revolutionised almost every aspect of business, from writing reports and optimising workflows to analysing data and automating decision making.
Despite this rapid progress, most AI systems are still fed by digital sources – yet much of the world’s information remains trapped in the physical world. Products move through supply chains with limited visibility. Packaging provides little insight into its journey. Assets disappear into warehouses, distribution networks and retail environments with only periodic updates on their status and location.
If AI is to fulfil its potential for driving understanding and efficiency, that gap must close.
Enter physical AI.
What is physical AI?
Physical AI refers to systems that perceive, reason about, and interact with the physical world.
Unlike traditional AI, which might suggest a movie or write an email, physical AI bridges digital intelligence with real-world action by enabling machines to perceive their surroundings, reason, make decisions, and execute tasks autonomously.
This is where the concept of the Internet of Everything (IoE) becomes a reality – where everyday items identify themselves, communicate their status and contribute data throughout their lifecycle.
But the ability to operate in and interact with the physical world, rather than software alone, isn’t just dependent on more powerful AI models. AI needs to ‘see’ and ‘feel’ the physical world at a granular level. And for that, it needs large amounts of real-time, item-level data.
Item-level intelligence: the next frontier
Most current AI models suffer from a ‘data gap’ in the physical world. Much of today’s supply chain and inventory data remains aggregated, delayed or incomplete. We may know a pallet has entered a warehouse, but we often lose visibility of the individual items on that pallet.
Item-level intelligence changes that equation.
Connected items create new streams of real-time information that can be analysed to identify emerging risks, detect bottlenecks and optimise operations. Inventory can become more visible. Product flows can become more transparent. Exceptions can be identified sooner because patterns that were hidden become visible. Decisions can be based on real-world conditions rather than historical assumptions.
The result is not simply greater efficiency, but also greater resilience.
Flexible intelligence at scale
The challenge with connecting billions of everyday items is one of economics and scalability.
Traditional silicon technology excels at delivering high-performance computing, but it isn’t optimal for embedding intelligence into low-cost, disposable or high-volume items such as packaging, labels and consumer goods.
This is where flexible semiconductor technology – such as FlexICs – comes into its own.
These flexible chips can go from design to high-volume production in weeks. A simplified production process drives down the carbon footprint of production, and their physically flexible form factor means they can also conform to curved surfaces.
This, in combination with their low cost, means they can be integrated into applications that have historically been impractical or uneconomic – financially or environmentally.
Rather than limiting intelligence to smartphones, computers and industrial equipment, flexible electronics make it possible to connect everyday products at unprecedented scale.
This dramatically increases the number of physical touchpoints capable of generating quality data for digital systems and AI platforms.
The next frontier of digitalisation
The future of AI is unlikely to be defined by a single technology. Advanced silicon processors will continue to perform the intensive computational workloads required for training and running sophisticated AI models. Data centres and cloud infrastructure will remain essential.
At the same time, flexible semiconductors have the potential to play a complementary role by extending intelligence to the edge of the physical world.
In this hybrid model, traditional silicon provides the processing power while flexible electronics provide the visibility. One enables AI to think. The other helps AI see.
Together, they create a foundation for a new generation of intelligent systems that are not limited to digital environments, but connected directly to everyday assets in the real world.
As AI becomes increasingly central to business decision-making, the organisations with the greatest advantage will be those able to generate the richest understanding of the real world. That requires more than smarter algorithms. It requires better data.
Flexible semiconductors offer a path towards that future – bringing intelligence to everyday items to create new streams of real-world information.
By bridging the gap between the digital and the physical, we’re moving toward a world where objects aren’t just smart – they’re active participants in a global, AI-driven ecosystem. This transition will redefine efficiency in every industry, driven by the tiny, flexible chips that give the physical world a digital voice.
