Screen Economies and AI-powered Price Tags

AI-powered digital price tags, or Electronic Shelf Labels (ESLs), are being adopted by major retail chains. These tiny screens can be updated wirelessly via a central system, delivering instant price updates for thousands of products and enabling price calibration, like ride-sharing apps, based on demand, time of day, competitor prices, and inventory levels. These frontier AI tools are already gradually replacing static paper price tags pinned to the plastic shelf channels along gondolas. These screens use AI algorithms to analyse data, such as food expiry dates and foot traffic, and can draw on fluctuations in information that ripple through the screen economy via the butterfly effect of transient events, such as the blockade of the Strait of Hormuz.

The global market for these devices is projected to reach $7.54 billion by 2033. These miniature screens allow workers to update shelf prices via an app, reducing the time required to make a price change from 2 days to 2 minutes. On supermarket gondolas, they provide operational data and insights from the shop floor to pinpoint opportunities to lower prices on perishable, seasonally specific, or otherwise time-sensitive items, including key staples. Industry-driven academic research at the University of California, working with industry advisory boards, shows that these screens reduce food waste by up to 21 per cent.

These screens improve operational efficiency and reduce waste, but they raise concerns about “surge” pricing, where costs could rise instantly due to shipping disruptions and rising petrol prices. Currently, there is no evidence that any retailer is using AI-powered price tags to engage in surge pricing, although weeks of war-driven pressure have driven up petrol prices and disrupted supply chains.

One large merchant intends to roll out digital tags to 2,300 stores by the end of 2026, and another has already deployed them in hundreds of locations. One popular UK retailer is also rolling out the new technology. Retailers claim that these reduce operational costs, such as those associated with replacing paper tags. However, parliamentarians and shoppers fear the risks of “personalised pricing” that depends on the shop’s location.

These screens are a glimpse of the coming wave of cultural artefacts of the infosphere revolution. However, this buildout is itself hampered by the Strait of Hormuz blockage in two ways: (1) investor confidence in returns in AI architectures, and (2) cost increases along supply chains in chip manufacturing. In the energy-hungry AI industry, where the business model is not yet clearly and firmly established and where investments in AI architectures are financed by heavy debt, the challenges are particularly acute. A note published by Quinn Emanuel in March 2026 shows that sector revenues in 2025 were about US$60bn and capital expenditure was US$400bn.

The opacity of the burgeoning private credit sector has attracted regulators’ attention, especially as data centre operators create off-balance-sheet special-purpose vehicles that own the data centres and the future rental income and borrow against them. Behemoths are revisiting their procurement plans and cash flow projections. The US$1.5 trillion in committed AI infrastructure is built on an assumption of frictionless global supply chains, which the Strait of Hormuz conflict has fundamentally broken. Apple plans to spend about $500 billion over four years. Amazon is estimating $200 billion in data centre spending in 2026. Google’s investment is somewhere between $175- $185 billion. Meta has forecasted over $600 billion in AI infrastructure by 2028, and Microsoft is moving toward $105 billion for the year.

The Bank of England’s Financial Policy Committee has highlighted the link between the share prices of AI companies and energy costs. Before the events in the Strait of Hormuz, selling pressure was mounting as investors worried whether AI-related investments would materialise. Today, the conflict could worsen these concerns, given the energy-intensive nature of the supply chains for key components and of datacentres. Investors understand that a prolonged period of high energy prices could crimp investment in AI architectures.

The production of advanced AI chips involves thousands of discrete manufacturing steps. Shin-Etsu and SUMCO in Japan control nearly 60 per cent of the global market. Their facilities, located in places like Niigata and Saga, specialise in producing the highest quality 300mm wafers required for advanced chip production. Siltronic AG, in Bavaria and Saxony, is one of the world’s top silicon wafer producers. Fabrication of the most advanced chips that power AI workloads is done almost entirely in Taiwan (92%) and South Korea (8%). Assembly and testing happen in Malaysia, Vietnam, and the Philippines. There are more than 50 points across this chain where a single country controls more than 65% of the global market share. Each of them has just become more expensive and more uncertain due to the Strait of Hormuz blockade.