Compute
Logic, alt paradigms, edge inference, quantum. How instructions become physics — and the substrates competing to do it.
Conventional logic
CPU/SoC, AI accelerators, RISC-V — von Neumann logic on silicon. Mature, mainstream, still the substrate under most production AI workloads.
Non-conventional computing
Computing that isn't von Neumann — analog, neuromorphic, probabilistic, Ising machines, thermodynamic, reversible, nanomechanical, memristors. The substrates that compute by other physics, mostly pre-revenue.
Compute architecture
Chiplets, UCIe, CFET, GAA nanosheet, FD-SOI, backside power delivery, cryo-CMOS. How transistors are arranged once you can't shrink them any further.
AI at the edge
Edge AI infrastructure, on-device LLM inference, TinyML, IoT compute. AI that runs where the data is created — milliwatts, not megawatts.
Quantum computing
All qubit platforms — silicon spin, trapped ion, superconducting, neutral atom, photonic, topological, bosonic cat. Plus quantum software, compilers, error correction, cryogenic control.