Ngokombiko we-Jiemian News wanamuhla, i-Institute of Artificial Intelligence Research ye-China Academy of Information and Communications Technology, ngokubambisana ne-Humanoid Robot (Shanghai) Co., Ltd. kanye ne-Embodied Intelligence Testing Laboratory, ikhiphe ngokuhlanganyela i-《Embodied Intelligence Training Ground Research Report (2026)》. Umbiko uveza ukuthi uchungechunge lwezimboni lwezinkundla zokuqeqesha seluqalile ukwakheka, futhi ngokujwayelekile lunophawu olufana nonxantathu obheke phansi: ukuhlinzekwa kobuchwepheshe ezansi komfula kuyasebenza, kanti izicelo ezingenhla komfula zisalindele ukwandiswa.
Izinkundla zokuqeqesha zidinga ukutshalwa kwezimali okukhulu ezimpahleni ezingaguquki. Izindleko zokwakha zihlanganisa ukwakhiwa kwezimo, ukuqoqwa kwedatha nemishini eyisisekelo, amandla okucubungula, inethiwekhi, isitoreji kanye nezinhlelo zokuphatha, phakathi kokunye. Izindleko zokusebenza kwansuku zonke nazo ziphezulu.
Njengamanje, izinsizakalo zebhizinisi zigxile kakhulu ekuthengiseni imikhiqizo yedatha, kanti kuhlolwa nezinsizakalo zokuqeqesha nokuqinisekisa amamodeli nemishini eyisisekelo, ucwaningo nentuthuko ngokuhlanganyela, kanye nokuqinisekisa ukuhlolwa. Ukuthengisa imikhiqizo yedatha kuyindlela eshesha kakhulu yokuthola imali, kodwa ukuthembela ekuthengiseni idatha kuphela akukwazi ukumboza ukutshalwa kwezimali okukhulu ezimpahleni, ngakho isikhathi sokubuyisa imali side. Iningi lezinkundla zokuqeqesha lisanda kwakhiwa, ngakho kudingeka ngokuphuthumayo kwakhiwe isakhiwo semali engenayo esihlukahlukene ngezindlela ezifana nokuhwebelana okusekelwe ezinkundleni, ukwandiswa kwezinsizakalo ezengeziwe, kanye “nokuqeqeshwa njengensizakalo”.