For 48 hours, Euro­pe’s brigh­test robo­tics talent came tog­e­ther to build — and to push the boun­da­ries of what phy­si­cal AI can do.

We’­re proud to have part­ne­red with Robo­TUM, START Munich, and ESRA — three stu­dent asso­cia­ti­ons — to bring this to life as part of our HMI2Market initia­ti­ve, which seeks to acti­va­te and ele­va­te the entre­pre­neu­ri­al poten­ti­al of Euro­pean hig­her edu­ca­ti­on insti­tu­ti­ons (HEIs) in the field of Human-Machi­ne Inter­ac­tion (HMI).

The sca­le of inte­rest said it all: 350+ appli­cants vied for a spot, 108 were sel­ec­ted, and in the end, 77+ buil­ders wal­ked through TUMor­row Hub and Lab Cam­pus to spend two days tur­ning bold ide­as into working sys­tems.

Buil­ders from dif­fe­rent count­ries fil­led the space, test­ing and ite­ra­ting around the clock. Sin­ce each track posed a fun­da­men­tal­ly dif­fe­rent chall­enge, teams first batt­led it out within their own track befo­re the three track win­ners met for one final show­down to crown an abso­lu­te cham­pi­on. Pri­zes included Anthro­pic cre­dits and cash.

Three Tracks, One Mis­si­on

Lead-spon­so­red by Igni­te Next, the hacka­thon unfold­ed across three distinct chall­enge tracks, each shaped by an indus­try or rese­arch part­ner — giving teams the free­dom to spe­cia­li­ze while pushing toward a com­mon goal: robots that can sen­se, move, and act more intel­li­gent­ly in the real world.

Track 1: Intel’s Chall­enge — Indus­tri­al Robo­tic Arms

Sup­port­ed by hard­ware from TUM MIRMI, this track chal­len­ged teams to expand what Visi­on-Lan­guage Model (VLM)-based robots can per­cei­ve. Team RoBoost rose to the occa­si­on, engi­nee­ring a sys­tem that lets VLM-dri­ven robots take an objec­t’s phy­si­cal pro­per­ties — like weight — as an input, giving them a way to sen­se the world bey­ond sight alo­ne. Their breakth­rough ear­ned them first place over­all.

Track 2: Wen­dy’s Chall­enge — Auto­no­mous Indus­tri­al Inspec­tion (Qua­dru­ped)

With hard­ware sup­port from TNG Tech­no­lo­gy Con­sul­ting, Mybot­shop, and ETHRC, this track asked teams to build auto­no­mous inspec­tion sys­tems fit for indus­tri­al envi­ron­ments. Team Arvit took the win with an agen­tic inspec­tion sys­tem, secu­ring third place over­all.

Track 3: MIR­MI’s Chall­enge — Dro­ne Racing

Backed by ITQ, with hard­ware sup­port from TUM’s Chair of Safe­ty, Relia­bi­li­ty and Per­for­mance of Lear­ning Sys­tems, this track tas­ked teams with app­ly­ing rein­force­ment lear­ning to real-world flight. Team Honey Bad­gers clai­med the track by trai­ning a dro­ne to clear a full gate cour­se in just 8 seconds — three seconds fas­ter than the run­ner-up — landing them second place over­all.

A Shared Mis­si­on Across Euro­pe

At TUMin­ter­na­tio­nal, we’­re proud to part­ner with such ambi­tious stu­dent orga­niza­ti­ons, and our invol­vement in HMI2Market — an EU-fun­ded initia­ti­ve under the EIT HEI Initia­ti­ve — is what makes it pos­si­ble.

None of this would have been pos­si­ble wit­hout all the insti­tu­ti­ons, part­ners, and spon­sors men­tio­ned abo­ve. A huge thank you to all of you.

Inte­res­ted in lear­ning more about HMI2Market?

HMI2Market is a pro­ject that aims to trans­form uni­ver­si­ty-based inno­va­ti­on eco­sys­tems to acce­le­ra­te the mar­ket rea­di­ness and adop­ti­on of Human-Machi­ne Inter­ac­tion (HMI) solu­ti­ons. By bridging the gap bet­ween aca­de­mic rese­arch and indus­try appli­ca­ti­on, the pro­ject fos­ters entre­pre­neur­ship, inclu­si­ve inno­va­ti­on, and digi­tal trans­for­ma­ti­on across Euro­pe’s manu­fac­tu­ring and labor-inten­si­ve sec­tors.

For more infor­ma­ti­on check out the HMI2Market Web­site