Otto G oup a d Se eact B i g Physical AI to Retu s P ocessi g
The pa t e s will deploy up to 25 AI-co t olled obots i a two-phase app oach at a Otto G oup logistics ce te to test the automatic p ocessi g of etu ed items
- At Otto G oup’s logistics ce te i Pola d, etu s will be g asped, ide tified, assessed, a d so ted by up to 25 AI-co t olled dual-a m obots f om Se eact.
- The obots u o Co tex, Se eact’s Physical AI model, which ca ecog ize a d ha dle items it has eve e cou te ed – o p oduct-specific teachi g, o fixed item catalog.
- Co tex is t ai ed a d validated with NVIDIA Isaac Sim, NVIDIA Isaac Lab a d NVIDIA Cosmos, so lesso s f om the etu s li e each the obots i days, ot weeks.
- U de a Robotics-as-a-Se vice model, Se eact ope ates a d mai tai s the systems. Employees ha d ove epetitive g aspi g a d lifti g a d take o supe visio , exceptio ha dli g a d quality checks.
HAMBURG, Ge ma y a d STUTTGART, Ge ma y, Oct. 28, 2026 (GLOBE NEWSWIRE) — Otto G oup a d obotics softwa e compa y Se eact today a ou ced a two-phase app oach fo AI-co t olled obots i o e of etail’s most complex ma ual tasks: i specti g a d so ti g etu ed custome o de s.
I itially, te AI-co t olled dual-a m systems will be deployed at Otto G oup’s logistics ce te i Pola d. I a seco d phase, 15 mo e obotic a ms a e pla ed to be added based o i itial esults, scali g the p og am to a total of 25. The two compa ies a e joi tly developi g a d efi i g the tech ology as pa t of the p oject, with Otto G oup evaluati g the pote tial fo a b oade ollout ac oss its logistics etwo k if the seco d phase is successful.
Few p ocesses i e-comme ce a e as ha d to automate as etu s – items come back folded, c umpled, epackaged, a d the asso tme t of a multi-catego y etaile eve stops cha gi g. T aditio al automatio depe ds o k ow p oducts a d p edictable p ese tatio , eithe of which exists i etu s. Co tex, Se eact’s Physical AI model, is built fo exactly this challe ge: it ge e alizes objects it has eve e cou te ed a d decides i eal time how to g asp each item, assess its co ditio , a d oute it acco di gly.
Co tex is built o NVIDIA’s physical AI stack. Se eact uses NVIDIA Isaac Sim a d NVIDIA Isaac Lab to tu eal picks f om the etu s li e i to thousa ds of simulated va ia ts, a d NVIDIA Cosmos wo ld fou datio models to e- e de them ac oss ga me t states, packagi g, mate ials a d lighti g. New policies a e post-t ai ed o NVIDIA Blackwell GPUs a d tested agai st held-out cases befo e they each the obots. Whe a obot st uggles with a item, the fix ca come back as a tested update withi days i stead of weeks.
The obots will be i t oduced i stages ove the cou se of the p oject. Each stage will ge e ate ope atio al feedback that the pa t e s ca use to imp ove the systems a d adapt them to the equi eme ts of la ge-scale etu s p ocessi g. This sha ed developme t p ocess is i te ded to establish whethe the tech ology ca meet the eliability, flexibility, a d quality sta da ds equi ed fo b oade use.
Wo ki g ha d i ha d with obots, people stay at the ce te of the p ocess: employees will shift away f om epetitive g aspi g a d lifti g towa d supe visio , exceptio ha dli g, a d fi al quality decisio s – tasks that depe d o expe ie ce a d judgme t.
“Ou custome s expect etu s to be simple a d fast – behi d that lies a p ocess that is time-c itical a d labo ious, both fo custome s a d fo ou busi ess. Retu s a e oto iously ha d to automate, a d Physical AI has the pote tial to cha ge that. This p oject lets us fi d out, at eal scale, how fa Physical AI ca take us – a impo ta t step towa ds a mo e flexible a d esilie t supply chai ,” says Ma ti Umla d, G oup Vice P eside t Supply Chai Ma ageme t at Otto G oup.
“A obot that o ly k ows yeste day’s catalog is of little use i a etu s ce te . Co tex has to deal with whateve comes out of the ext pa cel. Eve y etu the systems ha dle makes the model bette at the ext o e,” D . Ralf Gulde, CEO a d Co-Fou de , Se eact.
The pa t e ship builds o Otto G oup’s se ies of i vestme ts i obotics ac oss its logistics ope atio s a d fo ms pa t of the compa y’s medium-te m commitme t to i vest €350 millio i tech ology a d AI.
Fi d out mo e at ottog oup.com
- Se eact: p ess@se eact.ai
- Otto G oup: Ma i a Le z, ma i a.le z@ottog oup.com, +49 40 6461 6389
A photo accompa yi g this a ou ceme t is available at https://www.globe ewswi e.com/NewsRoom/Attachme tNg/9f531c22-18d0-4cde-86e5-d05308235661




 