Guides
Compute, explained
Short, careful explainers on the processors and technologies behind modern computing, with the companies building them.
8 guides
- Comparison CPU vs GPU vs TPU vs QPU: What Each Does Best A plain-English comparison of CPUs, GPUs, TPUs and QPUs: how each processor works, which jobs suit it, and why none of them replaces the others.
- Explainer What Is a QPU? Quantum Processors Explained What a quantum processing unit is, how qubits and error correction work, the five main hardware types, and why qubit counts alone tell you little.
- Explainer What Is Photonic Computing? Light for Logic and Links Photonic computing covers two ideas: using light to calculate, and using light to move data between chips. Learn how each works and which is shipping.
- Explainer What Is Neuromorphic Computing? Neuromorphic chips borrow ideas from the brain: spiking neurons, event-driven processing and memory next to compute. See how they work and where they are used.
- Explainer What Is RISC-V? The Open Instruction Set Explained RISC-V is an open instruction set that anyone can build processors on without paying an ISA licence fee. See how it works, who uses it and why it matters.
- Explainer The Memory Wall: HBM, CXL and In-Memory Computing Why memory, not arithmetic, often limits AI chips, and how HBM, CXL, large on-chip SRAM and in-memory computing each try to move the bottleneck.
- Explainer AI Accelerators Explained: GPUs, TPUs, NPUs and More An AI accelerator is a chip made to run neural networks. Compare GPUs, TPUs, NPUs, inference ASICs and wafer-scale chips, and training versus inference.
- Guide How to Evaluate a Compute Company A checklist for judging chip and compute start-ups: process node, software stack, memory bandwidth, customers, funding and tape-out status, plus red flags.