Did all early computers use binary?

No. ENIAC (1946), one of the first general-purpose electronic computers, actually used decimal counters built from vacuum tubes. Binary became dominant over the following decade because it was cheaper and more reliable to build with the switching components (relays, then transistors) that followed.

Could a computer be built to use decimal directly?

Yes, and some historical machines did — but detecting ten distinct, reliable voltage levels in a circuit is far harder and more error-prone than detecting two (on/off), especially as components shrink and signals get noisier. Binary's simplicity is what let transistor counts scale into the billions.

Why do programmers still deal with binary if code is written in decimal?

Because the hardware underneath — memory addresses, bitwise flags, network protocol fields, floating-point representations — is fundamentally binary, and some of that structure leaks through to what a program can and can't represent exactly (see why 0.1 + 0.2 doesn't equal 0.3 in most languages).

Is binary the most 'efficient' base for computers?

Not by every measure — base-3 (ternary) is slightly more information-efficient in theory, and a few experimental ternary computers were built in the 1950s-60s. Binary won because two-state components are dramatically easier and cheaper to build reliably at scale, not because 2 is mathematically optimal.

Guide

Why computers use binary instead of decimal numbers

A transistor is much better at reliably being 'on' or 'off' than at reliably being one of ten distinct voltage levels.

By Buğra SözeriPublished September 25, 2026

It isn’t tradition, and it isn’t that binary is mathematically special — it’s that a two-state electrical signal is far easier to build and detect reliably than a ten-state one. A transistor switching between “on” and “off” only has to clear one threshold to be unambiguous; a component that needed to distinguish ten different voltage levels for a decimal digit would need ten times the precision, and any noise in the circuit (heat, interference, tiny manufacturing variance) becomes far more likely to flip a reading to the wrong digit.

Decimal computers actually existed first

ENIAC, one of the first general-purpose electronic computers (1946), used decimal counters built from vacuum tubes — it represented numbers in base 10, matching how humans already count. But building reliable decimal circuitry required far more components and more careful engineering than binary would have, and as computing moved from vacuum tubes to transistors over the following decade, binary’s simplicity won out almost universally. By the 1960s, decimal computing was effectively obsolete outside of niche financial hardware.

Two states vs. ten states

The core argument is about signal margin. A binary circuit only needs to distinguish “clearly high voltage” from “clearly low voltage” — there’s a wide buffer zone in between that noise can occupy without flipping the reading. A decimal circuit trying to distinguish ten discrete voltage bands has to squeeze those bands into the same total voltage range, leaving each one a fraction of the margin binary gets — any electrical noise is proportionally far more likely to push a reading into the wrong band. That reliability gap is why binary components can be packed by the billions onto a modern chip while staying accurate, something that would be far harder with multi-level decimal circuitry.

What binary buys beyond reliability

Property	Why it matters
Simple logic gates (AND, OR, NOT)	Directly implement Boolean algebra with two-state signals
Easy error detection	A corrupted bit is either 0-that-should-be-1 or the reverse — simple to check with parity
Uniform storage	Every value — numbers, text, images, instructions — reduces to the same 0/1 units

Why binary still leaks into everyday computing

Programmers rarely write binary literals directly, but the underlying binary structure shows up constantly: memory sizes in powers of two, IPv4 addresses built from four binary octets (see the IPv4 subnet mask guide for a concrete example), bitwise flags packed into a single integer, and the reason floating-point arithmetic sometimes produces surprising results — many decimal fractions, like 0.1, have no exact binary representation, the same way 1/3 has no exact decimal representation. Understanding how binary converts to decimal makes those quirks make sense instead of feeling like bugs.

Frequently asked questions

Did all early computers use binary?: No. ENIAC (1946), one of the first general-purpose electronic computers, actually used decimal counters built from vacuum tubes. Binary became dominant over the following decade because it was cheaper and more reliable to build with the switching components (relays, then transistors) that followed.
Could a computer be built to use decimal directly?: Yes, and some historical machines did — but detecting ten distinct, reliable voltage levels in a circuit is far harder and more error-prone than detecting two (on/off), especially as components shrink and signals get noisier. Binary's simplicity is what let transistor counts scale into the billions.
Why do programmers still deal with binary if code is written in decimal?: Because the hardware underneath — memory addresses, bitwise flags, network protocol fields, floating-point representations — is fundamentally binary, and some of that structure leaks through to what a program can and can't represent exactly (see why 0.1 + 0.2 doesn't equal 0.3 in most languages).
Is binary the most 'efficient' base for computers?: Not by every measure — base-3 (ternary) is slightly more information-efficient in theory, and a few experimental ternary computers were built in the 1950s-60s. Binary won because two-state components are dramatically easier and cheaper to build reliably at scale, not because 2 is mathematically optimal.

Sources & references

Authoritative references cited by this piece. Verified by Buğra Sözeri on the dates shown and re-checked at every deploy.

IEEE — Milestones: ENIAC (1946) — Context on early decimal-based computing (ENIAC used decimal counters) before binary became the near-universal standard(as of 2026-09-25)
IEEE 754-2019 — Standard for Floating-Point Arithmetic — The modern standard for how computers represent numbers in binary at the hardware level(as of 2026-09-25)