The interesting part of the CERBERUZ design is the step it deletes. A sorting stage is a mechanism: it carries mass, it draws power, it needs somebody or something to decide where a laminated pouch goes, and it jams. Grinding the whole stream to a fine powder and using the Zotek foam as reinforcement instead of pulling it out [3] converts a classification problem into a materials problem. That is a trade rather than a free win, and NASA's announcement does not report the properties of the resulting parts, the throughput, or the power draw for any prototype [17].
The money is the clearest signal available. Add up the Phase 2 prizes NASA listed and the total is $1,350,000: $775,000 to the MIT team [2], $225,000 to Terasynth for second in the prototype track and $125,000 to Moon Made for second in the digital twin track, three $50,000 prototype category prizes, two $25,000 digital twin category prizes, and a $25,000 people's choice award that also went to Terasynth, bringing that team to $250,000 [10][11][1][4]. MIT's share of the listed Phase 2 awards is about 57 percent [2], and roughly a quarter of the $3 million the two-phase competition was built around [6][3].
What the awards actually rank is five days of behavior on a bench in Tuscaloosa in August [8]. That run happens at one g, in air, with terrestrial trash, and none of those conditions exercise what usually decides whether a recycler survives a surface mission: abrasive dust ingress, vacuum, thermal cycling, duty cycles measured in months. The digital twin track is NASA's hedge against that gap, since Phase 2 asked for a virtual model of the recycling system alongside the hardware [4][7]. It is a sensible way to buy confidence about conditions you cannot reproduce in Alabama, provided the model gets validated against hardware that has run long enough to break.
One structural detail deserves attention. Phase 1 was open worldwide and drew more than 1,200 registrations, a record across 20 years of NASA Centennial Challenges [13], from which 50 judges picked 17 winning teams out of nearly 200 submissions [14], roughly one submission for every six registrations [5]. Phase 2 was restricted to U.S. teams and required entries unrelated to Phase 1 [7][15]. That rule buys variety of approach and gives up two years of iteration on the Phase 1 ideas, which is a defensible choice for a prize whose product is options rather than a qualified system. Jennifer Edmunson, who manages Centennial Challenges at NASA's Marshall Space Flight Center, described the finale as the culmination of two years of concept-to-prototype work [12].
My read is that NASA has bought a design principle here, not a machine. The number that would move it from principle to plan is a tensile or flexural figure for a molded part made from the mixed powder, set next to the same figure for a part made from sorted single-polymer feedstock. Until that comparison is published, contamination tolerance is a promising hypothesis with a $775,000 vote behind it [2].
