Who Invented Gunpowder?

Nobody invented gunpowder as a weapon. It emerged from Chinese alchemy during the Tang dynasty (618–907 CE), among practitioners who were trying to produce medicines and an elixir of immortality. The Chinese name for it, huoyao, translates as “fire medicine,” and it remains the standard word for gunpowder in Chinese today.

The earliest surviving complete formula appears in a military manual of 1044, roughly two centuries after the first written descriptions of the mixture. Those earlier texts are not recipes. They are safety warnings.

What the alchemists were doing

Daoist alchemy in China was a serious technical practice with centuries of accumulated method behind it. Practitioners heated, combined, and purified minerals in pursuit of substances that would extend life, cure disease, or transform base materials. Saltpeter and sulfur were both standard reagents in that work, valued for what they did under heat.

By the fifth century, alchemists had noticed that saltpeter burns with a distinctive purple flame, which gave them a way to test purity and work with more consistent material. That detail matters more than it looks. A mixture that behaves unpredictably because one ingredient is contaminated is a curiosity. A mixture that behaves the same way every time is a technology.

The earliest texts are warnings

The Taishang Shengzu Jindan Mijue, compiled around 808 CE, records that combining saltpeter, sulfur, and carbon caused flames that burned hands and faces and set houses alight.

A second Daoist text from roughly 850 CE, the Zhenyuan Miaodao Yaolüe, explicitly warns readers against mixing these substances because of the fire risk.

Read those two documents together and the implication is clear. A text does not warn against a procedure nobody is performing. By the mid-ninth century, alchemists were combining these three ingredients often enough, and injuring themselves often enough, that the hazard needed writing down.

This is the part most accounts flatten into a single dramatic explosion. There is a later story, in a book called Gui Dong published in 1185, about a Tang alchemist whose furnace exploded, but it is not established that gunpowder caused it. No contemporary account records a eureka moment, and the documentary evidence points to accumulated accidents rather than one.

From hazard to weapon

The transition took time. Early applications in China were largely pyrotechnic: fire arrows, signal flares, and rudimentary incendiary devices appear in records from the late Tang and early Song periods.

The Wujing Zongyao, a military compendium completed around 1044, contains the oldest surviving complete gunpowder formulas. It is a substantial document, cataloguing incendiary bombs and fire lances alongside weapon names including the “flying incendiary club for subjugating demons.” The formulas it gives are lower in saltpeter than later black powder, which makes them better incendiaries than explosives.

The fire lance, a tube packed with powder and directed at an opponent, is generally treated as the direct ancestor of the firearm. Metal-barrelled guns were in use in China by the thirteenth century, and cannon followed.

Gunpowder is counted among the Four Great Inventions of ancient China, alongside papermaking, printing, and the compass.

Why the framing matters

Calling gunpowder an accident is close to right but slightly lazy. Accident implies a single unlucky moment. What the record shows is a research programme with a defined goal, running for centuries, in which a repeatable and dangerous reaction was identified, documented, refined, and eventually redirected toward an application nobody set out to find.

The alchemists were not careless. They were systematic, and their systematic work produced a result orthogonal to their objective. That pattern recurs throughout the history of invention, and gunpowder is its oldest well-documented instance.