Learn

AI semiconductors & photonics, explained

Deep, plain-language guides to the hardware powering the AI buildout — the substrates, lasers, optics, memory, packaging and power that sit upstream of NVIDIA. Each guide is standalone; companies from @aleabitoreddit's tracked universe are linked where relevant.

Silicon photonics

Silicon photonics builds optical components - waveguides, modulators, detectors - on a silicon chip so data travels as light instead of electrical signals over copper. It uses standard chip factories

Co-packaged optics (CPO)

Co-packaged optics (CPO) integrates the optical engine that converts electrical signals to light into the same package as a switch or accelerator chip, replacing pluggable transceivers. By shortening

Indium phosphide (InP)

Indium phosphide (InP) is a III-V compound semiconductor that, unlike silicon, can efficiently generate and detect infrared laser light. It is the core material for the 1310/1550 nm lasers, photodiode

Gallium arsenide (GaAs)

Gallium arsenide (GaAs) is a III-V compound semiconductor made from gallium and arsenic. Unlike silicon, it efficiently converts electricity into light and handles high frequencies, making it the back

Silicon-on-insulator (SOI)

Silicon-on-insulator (SOI) is a semiconductor wafer that adds a thin buried oxide insulating layer beneath the active silicon. This isolation reduces leakage and parasitic capacitance, making transist

Epiwafer / epitaxy

An epiwafer is a wafer with ultra-thin, atom-by-atom "epitaxial" layers of compound semiconductor (GaAs, InP, GaN) grown on a substrate. Epitaxy gives these layers the precise crystal structure needed

MBE (molecular-beam epitaxy)

Molecular-beam epitaxy (MBE) is a way to grow crystalline semiconductor films one atomic layer at a time. Beams of pure atoms are fired at a heated substrate inside an ultra-high vacuum, giving near-p

CW DFB laser

A CW DFB laser is a continuous-wave distributed-feedback laser: a semiconductor chip that emits a steady, single-wavelength beam of light. In AI data centers, it supplies the "carrier" light for silic

Optical transceiver

An optical transceiver is a small plug-in module that converts electrical data into laser light to send over fiber, and converts incoming light back into electrical signals. It's the workhorse interco

800G / 1.6T

"800G" and "1.6T" are the data rates of optical transceivers - the laser-based plugs that move 800 gigabits or 1.6 terabits per second of network traffic between switches and AI servers. They are core

Fiber array unit (FAU)

A fiber array unit (FAU) is a precision component that holds and aligns multiple optical fibers in a fixed array so they couple light efficiently into a silicon photonics chip. FAUs are a critical bui

OSAT

OSAT stands for Outsourced Semiconductor Assembly and Test. OSAT firms take finished silicon wafers from chip factories, cut them into individual dies, package those dies into usable chips, and test t

Advanced packaging

Advanced packaging is a set of techniques for combining several chips (chiplets), memory stacks, and even optics into a single high-performance package, instead of relying on one large monolithic chip

Glass-core substrate

A glass-core substrate is an advanced chip-packaging base that swaps the usual plastic-like organic core for a thin pane of specialty glass. Because glass stays exceptionally flat and electrically cle

HBM (high-bandwidth memory)

HBM (high-bandwidth memory) is DRAM stacked vertically into towers and wired directly beside a processor, delivering terabytes per second of memory bandwidth. It is the memory inside modern AI acceler

DRAM

DRAM (dynamic random-access memory) is the fast, volatile working memory that holds the data a computer is actively using. Each bit lives in a tiny transistor-and-capacitor cell that must be refreshed

NAND

NAND flash is non-volatile memory that stores data as trapped electrical charge, retaining it without power. It is the storage inside SSDs, phones, and AI data-center servers. Four firms—Samsung, SK H

Memory supercycle

A "memory supercycle" is a prolonged, demand-driven boom in memory chips (DRAM, NAND and especially HBM) where prices and revenue surge for years rather than a few quarters. The 2025-2026 cycle is fue

Metrology

Metrology is the science of measurement. In semiconductors it means precisely measuring and inspecting wafers during manufacturing to catch defects and keep yields high. With 400-600 process steps per

Neocloud

A neocloud is a cloud provider that rents out GPU computing power specialized for AI training and inference, rather than the general-purpose services of hyperscalers like AWS or Azure. Neoclouds such

GaN / SiC power

GaN (gallium nitride) and SiC (silicon carbide) are "wide-bandgap" semiconductors that switch electricity faster and waste less of it as heat than ordinary silicon. They are central to AI data centers

Harmonic drive / humanoid robotics

A harmonic drive, or strain wave gear, is a compact, near-zero-backlash gear reducer that turns a motor's fast, weak spin into slow, strong, precise motion. It's the gearbox of choice for the rotary j

Foundry

A semiconductor foundry is a factory that manufactures microchips designed by other companies. "Pure-play" foundries like TSMC build but never design competing products, so fabless firms (Nvidia, Appl

Fabless

A fabless company designs and sells semiconductor chips but owns no factory ("fab"). It outsources manufacturing to a contract foundry such as TSMC, letting it focus capital on design and IP. Nvidia,

ADR (American Depositary Receipt): What It Is and How It Works

An American Depositary Receipt (ADR) is a negotiable certificate issued by a US bank that represents shares of a non-US company, letting Americans buy and sell that stock in US dollars on US markets.

The AI photonics supply chain

The AI photonics supply chain is the chain of companies that turn light into a data-mover for AI data centers. It runs from compound-semiconductor materials (indium phosphide, gallium arsenide) and la