Nvidia is sending GPUs to the moon
THE SO WHAT
GPUs heading to the moon is more than a stunt—it points to AI compute becoming part of off‑planet infrastructure, not just ground‑side processing. Space and remote‑sensing operators should assume future payloads will include on‑site model execution, changing latency, autonomy, and bandwidth assumptions.
READ THE SOURCE
MORE FROM THE WIRE
Deep & Emerging TechNew White House Science Strategy Points to a Quantum-Style Innovation Model
The White House floating a “quantum-style” innovation model for broader science is a signal that public–private consortia and long-horizon roadmaps are becoming the default for deep tech. If you’re building in quantum or adjacent fields, align your grant, partnership, and hiring plans to this two-track world: national labs plus commercial scale-up.
Deep & Emerging TechFord and Geely will build EVs together at Ford’s plant in Spain. Ford owns 66%, Geely 34%.
Ford handing 34% of a Valencia EV JV to Geely is a reminder that EV economics are pushing even incumbents into shared manufacturing and platform deals. Suppliers and software vendors should treat this as one customer with two decision centers — procurement and integration choices will be negotiated across Detroit, Gothenburg, and Hangzhou.
Deep & Emerging TechIBM acquires HRL Laboratories from Boeing and General Motors, seeking to integrate HRL's electron spin quantum circuits into its quantum computers
IBM buying HRL to fold electron spin qubits into its stack is a bet on a dual-track quantum roadmap — superconducting today, spin-based tomorrow. If you’re building quantum software or partnerships, assume hardware heterogeneity is here to stay and design for portability across qubit types, not just vendors.
Deep & Emerging TechMusk thanks Micron for chips, and builds a $55bn fab to replace it
A $55B “Terafab” alongside public praise for Micron underlines how memory has become a first-class constraint in AI — not just compute. If you’re planning large models or edge deployments, you need a memory strategy as explicit as your GPU plan: supplier diversity, form factors, and how much you’re willing to vertically integrate.