Dr Fazal Ali

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 two days to two 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$60 bil­lion and cap­i­tal ex­pen­di­ture was US$400 bil­lion.

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 fore­cast over $600 bil­lion in AI in­fra­struc­ture by 2028, and Mi­crosoft is mov­ing to­wards $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 per cent) and South Ko­rea (8 per­cent). 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 per cent of the glob­al mar­ket share. Each of them has just be­come more ex­pen­sive and more un­cer­tain due to the Strait of Hor­muz block­ade.

Dr Fazal Ali com­plet­ed his Mas­ter's in Phi­los­o­phy at the Uni­ver­si­ty of the West In­dies. He was a Com­mon­wealth Schol­ar who at­tend­ed the Uni­ver­si­ty of Cam­bridge, Hugh­es Hall, the provost of the Uni­ver­si­ty of Trinidad and To­ba­go and the act­ing pres­i­dent, and chair­man of the Teach­ing Ser­vice Com­mis­sion. He is present­ly a con­sul­tant with the IDB.